WO2024036580A1 - 运动轨迹规划的方法、装置以及智能驾驶设备 - Google Patents

运动轨迹规划的方法、装置以及智能驾驶设备 Download PDF

Info

Publication number
WO2024036580A1
WO2024036580A1 PCT/CN2022/113426 CN2022113426W WO2024036580A1 WO 2024036580 A1 WO2024036580 A1 WO 2024036580A1 CN 2022113426 W CN2022113426 W CN 2022113426W WO 2024036580 A1 WO2024036580 A1 WO 2024036580A1
Authority
WO
WIPO (PCT)
Prior art keywords
trajectory
driving device
intelligent driving
derivation
motion
Prior art date
Application number
PCT/CN2022/113426
Other languages
English (en)
French (fr)
Inventor
申子豪
王超
朱良凡
赵彬
戴正晨
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to PCT/CN2022/113426 priority Critical patent/WO2024036580A1/zh
Publication of WO2024036580A1 publication Critical patent/WO2024036580A1/zh

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units
    • B60W30/08—Active safety systems predicting or avoiding probable or impending collision or attempting to minimise its consequences
    • B60W30/09—Taking automatic action to avoid collision, e.g. braking and steering
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/09—Arrangements for giving variable traffic instructions
    • G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
    • G08G1/0967—Systems involving transmission of highway information, e.g. weather, speed limits
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems

Definitions

  • the present application relates to the field of intelligent driving, and more specifically, to a method and device for motion trajectory planning and intelligent driving equipment.
  • the vehicle's decision-making module labels obstacles and provides reasonable input to the motion planning module to ensure that the autonomous vehicle responds reasonably to surrounding obstacles.
  • Traditional decision planning based on finite state machine (FSM) is usually affected by the accuracy of target trajectory prediction. For example, when the target is a pedestrian or a non-motor vehicle, the direction of movement is unclear due to the large degree of freedom of target movement. As a result, the trajectory prediction results have large jumps, and the predicted movement trajectories of targets often frequently invade the driving roads of autonomous vehicles.
  • FSM finite state machine
  • This application provides a method and device for motion trajectory planning and intelligent driving equipment, which can reduce the problems of intelligent driving equipment such as unexpected and frequent mis-braking, mis-braking and stopping, continuous conservative yielding, and falling into simultaneous starts and stops. It is helpful to Improve the traffic efficiency of intelligent driving equipment and improve user driving experience.
  • the intelligent driving equipment involved in this application may include road vehicles, water vehicles, air vehicles, industrial equipment, agricultural equipment, or entertainment equipment, etc.
  • the intelligent driving device can be a vehicle, which is a vehicle in a broad sense, and can be a means of transportation (such as commercial vehicles, passenger cars, motorcycles, flying cars, trains, etc.), industrial vehicles (such as forklifts, trailers, tractors, etc.) vehicles, etc.), engineering vehicles (such as excavators, bulldozers, cranes, etc.), agricultural equipment (such as lawn mowers, harvesters, etc.), amusement equipment, toy vehicles, etc.
  • the embodiments of this application do not specifically limit the types of vehicles.
  • the intelligent driving device can be a means of transportation such as an airplane or a ship.
  • a method of motion trajectory planning applied to a first intelligent driving device that can move autonomously.
  • the method may include: obtaining a planned driving path and a first motion parameter of the first intelligent driving device, the first The motion parameters include the speed and/or acceleration of the first intelligent driving device; obtaining the predicted motion trajectory and the second motion parameter of the first object, the second motion parameter including the speed and/or acceleration of the first object; in the planning When the driving path intersects with the predicted motion trajectory, the planned motion trajectory of the first intelligent driving device is determined based on the first motion parameter and the second motion parameter.
  • the planned motion trajectory includes the first intelligent driving device using the first predetermined motion trajectory. Set the speed to rush or give way to the movement trajectory of the first object; control the first intelligent driving device to drive according to the planned movement trajectory.
  • the first preset speed may be a lower speed that makes the driver and passenger of the first intelligent driving device feel safe, and may be adjusted according to the aggressiveness of the driver and passenger.
  • the first preset speed is greater than or equal to 3 kilometers per hour and less than or equal to 15 kilometers per hour.
  • the first preset speed may be 5 km/h, or may be 10 km/h, or may be other speeds.
  • the first intelligent driving device rushes to the first object or yields to the first object is determined based on which object passes through the conflict area first. If the first intelligent driving device passes through the conflict area before the first object, the first intelligent driving device is considered to have passed the first object; if the first intelligent driving device passes through the conflict area after the first object passes through, the first intelligent driving device is considered to have passed through the conflict area.
  • the smart driving device gives way to the first object.
  • the conflict area is an area where the predicted planned path of the first intelligent driving device overlaps with the deduced trajectory of the first object.
  • the first intelligent driving device can actively maintain a lower speed to test the behavior of the first object, so that the planned movement trajectory of the first intelligent driving device is not susceptible to the influence of the second object.
  • the impact of an object’ s predicted trajectory accuracy.
  • the first intelligent driving device gives way to the first object, the first intelligent driving device can drive at the first preset speed without stopping to give way, which helps to improve the driving stability of the first intelligent driving device.
  • the intelligent driving device carries passengers, it can improve the comfort of the passengers and improve the driving experience of the passengers; in addition, since the first intelligent driving device does not need to brake when yielding to the first object, it can also improve the performance of the first intelligent driving device. traffic efficiency.
  • the first intelligent driving device rushes over the first object at the first preset speed, since the first intelligent driving device travels at a lower speed, the first object can adjust its own movement trajectory according to the movement trajectory of the first intelligent driving device. , can prevent the first intelligent driving device from continuously giving way to the first object, which helps to improve the traffic efficiency of the first intelligent driving device.
  • the first intelligent driving device will not cause unexpected frequent mis-brakes, mis-brakes, continued conservative yielding, and other scenarios. Getting stuck in problems such as starting and stopping at the same time will help improve the driving stability and traffic efficiency of the first intelligent driving equipment.
  • the first intelligent driving device may be a vehicle that can move autonomously, or it may also be another object that can move autonomously, such as an intelligent robot.
  • the first object may be another intelligent driving device, including an intelligent driving device that can move autonomously, and/or an intelligent driving device that is manually controlled.
  • the intelligent driving device may be a motor vehicle, or it may be a non-machine vehicle. Train.
  • the first object may also be a person, or may be another object that can move autonomously except the intelligent driving device.
  • the planned driving path of the first intelligent driving device may be a path generated by the regulation module of the first intelligent driving device and expected to be traveled by the first intelligent driving device within a certain period of time in the future. It should be understood that the path may only contain spatial location information.
  • the predicted movement trajectory of the first object may be the trajectory prediction module of the first intelligent driving device predicting the movement of the first object within a certain period of time in the future based on the obtained position, speed and/or acceleration of the first object.
  • the trajectory of movement may also be sent to the first intelligent driving device by the cloud or other facilities.
  • ture period of time can be 10 seconds, or it can be 20 seconds, or it can be other lengths of time.
  • the first intelligent driving device needs to be planned.
  • the driving trajectory of the driving equipment Then the first game strategy can be determined based on the first motion parameter of the first intelligent driving device and the motion parameter of the first object, and then the planned motion trajectory of the first intelligent driving device can be determined.
  • the acceleration of the first intelligent driving device can be inferred based on the speed of the first intelligent driving device; the second motion parameter only contains When the speed of the first object is included, the acceleration of the first object can be inferred based on the speed of the first object.
  • the speed of the first intelligent driving device can be inferred based on the position information in the planned driving path of the first intelligent driving device and the acceleration; in the second motion parameter
  • the speed of the first object can be inferred based on the position information in the predicted motion trajectory of the first object and combined with the acceleration.
  • determining the planned motion trajectory according to the first motion parameter and the second motion parameter includes: determining the third motion trajectory according to the first sampled acceleration and the first motion parameter.
  • the first derivation trajectory of an intelligent driving device, the first sampled acceleration is the possible acceleration of the first intelligent driving device determined in the sampling space; determine the first object's acceleration according to the second sampled acceleration and the second motion parameter.
  • the second derivation trajectory, the second sampled acceleration is the possible acceleration of the first object determined in the sampling space; the first derivation trajectory and the second derivation trajectory indicate the first intelligent driving device and the third An object will not collide, or when the first object collides with the side or tail of the first intelligent driving device in the direction of travel of the first object at the first moment, according to the first derivation trajectory and the third
  • the second derivation trajectory determines the planned movement trajectory, and the first moment is the moment after the current moment.
  • first sampled acceleration and second sampled acceleration pairs can be obtained in the sampling space, and further, multiple sets of first deduced trajectory and second deduced trajectory pairs can be determined. Further, the planned motion trajectory of the first intelligent driving device can be determined based on the pair with the smallest automatic driving strategy cost among the plurality of pairs of first derivation trajectories and second derivation trajectories.
  • the minimum value of the first sampling acceleration and the second sampling acceleration may be -4m/s 2 , and the maximum value may be 3m/s 2 .
  • the sampling interval between the two groups of first sampled accelerations may be 1m/s 2 , or may be other values; the two groups of first sampled accelerations may be The sampling interval between two sampling accelerations can be 1m/s 2 , or it can also be other values.
  • the second deduced trajectory is the deduced possible movement trajectory of the first object.
  • taking the first intelligent driving device and the first object as both vehicles as an example, when the first intelligent driving device and the first object are both in a driving state, "the first object is in the first object "Colliding with the side or rear of the first intelligent driving device in the direction of travel” can be understood to mean that the front of the first object collides with parts of the body of the first intelligent driving device other than the front; or, the first object is in In the reversing state, the rear of the first object collides with the body of the first intelligent driving device except for the front of the vehicle.
  • the planned movement trajectory may also be determined based on the first derivation trajectory and the second derivation trajectory.
  • the collision mode is that the first intelligent driving device collides with the first intelligent driving device.
  • the game strategy of the first intelligent driving device can be determined based on whether the first intelligent driving device collides with the first object and the type of collision, which helps to improve the driving safety of the first intelligent driving device.
  • the first motion parameter includes the speed and/or acceleration of the first intelligent driving device, and the first motion parameter is determined based on the first sampled acceleration and the first motion parameter.
  • the first derivation trajectory of an intelligent driving device includes: determining a first sub-deduction trajectory based on the first sampled acceleration and the speed and/or acceleration of the first intelligent driving device, and the end point of the first sub-deduction trajectory is the third
  • the speed of an intelligent driving device is the first preset speed
  • the second sub-deduction trajectory is determined according to the first preset speed
  • the end point of the first sub-deduction trajectory is the starting point of the second sub-deduction trajectory; according to the first sub-deduction trajectory
  • the derivation trajectory and the second sub-deduction trajectory determine the first derivation trajectory.
  • end point of the first sub-deduction trajectory is not the point where the first intelligent driving device stops moving. This "end point of the first sub-deduction trajectory” is only used to distinguish two sub-deductions with different characteristics in the first deduction trajectory. As a characteristic of the trajectory, the "end point of the first sub-deduction trajectory” can also be called the “starting point of the second sub-deduction trajectory”.
  • the first sub-derivation trajectory may include the acceleration of the first intelligent driving device from The current acceleration change is the trajectory of the first sampled acceleration.
  • the first sub-derivation trajectory may also include the first intelligent driving device with the first Sample acceleration driving to reduce the speed of the first intelligent driving device to a trajectory of the first preset speed.
  • the first sub-derivation trajectory may include the acceleration of the first intelligent driving device from The current acceleration change is the trajectory of the first sampled acceleration.
  • the first sub-derivation trajectory may also include the first intelligent driving device with the first Sample acceleration driving to increase the speed of the first intelligent driving device to a trajectory of the first preset speed.
  • the first sub-derivation trajectory may include the first The intelligent driving device drives with the current acceleration (ie, the first sampled acceleration) so that the first preset speed changes from the current speed to the trajectory of the first preset speed.
  • the first sub-derivation trajectory may be empty. That is, the second sub-derivation trajectory is determined directly based on the current speed (ie, the first preset speed).
  • the first sub-derivation trajectory can also be determined based on the first sampled acceleration and the speed and/or acceleration of the first intelligent driving device.
  • the end point of the first sub-derivation trajectory is the first intelligent driving device.
  • the speed of the driving equipment is a second speed, and the second speed is the maximum speed that the first intelligent driving equipment can travel or the maximum speed limit of the road section where the first intelligent driving equipment travels; the second sub-speed is determined based on the second speed.
  • Deduction trajectory; the first deduction trajectory includes the first sub-deduction trajectory and the second sub-deduction trajectory.
  • the first sub-derivation trajectory may include the acceleration of the first intelligent driving device changing from the current acceleration to The trajectory of the first sampled acceleration.
  • the first sub-derivation trajectory may also include the first intelligent driving device using the first sampled acceleration. Driving to increase the speed of the first intelligent driving device to the trajectory of the second speed.
  • the first deduced trajectory of the first intelligent driving device may also be determined based on the second sampled acceleration and the first motion parameter. Specifically, it may include: determining a third sub-derivation trajectory based on the second sampled acceleration and the speed and/or acceleration of the first intelligent driving device, and the speed of the first intelligent driving device at the end of the third sub-derivation trajectory is the third Three speeds or zero, the third speed is the maximum speed that the first intelligent driving equipment can travel or the maximum speed limit of the road section where the first intelligent driving equipment travels; at the end of the third sub-derivation trajectory, the first intelligent driving equipment When the speed of the driving equipment is the third speed, the fourth sub-deduction trajectory is determined according to the third speed; the first derivation trajectory includes the third sub-deduction trajectory; or, the first derivation trajectory includes the third sub-deduction trajectory and a fourth sub-deduction trajectory, and the end point of the third sub-deduction trajectory is the starting point of the fourth sub-deduction trajectory.
  • the "sub-derivation trajectory" involved in the above technical solution can be understood as a set of points where the position of the first intelligent driving device changes with time in the forward direction of the first intelligent driving device, and each point does not include vertical The position in the forward direction of the first intelligent driving device.
  • the "forward direction of the first intelligent driving device” may be a direction parallel to the longitudinal symmetry plane of the intelligent driving device in a plane parallel to the ground.
  • the speed of the first intelligent driving device is minimized to the first preset speed, so that the first intelligent driving device can be avoided. Reducing the speed of the driving equipment to zero can effectively avoid frequent braking of the first intelligent driving equipment and help improve the traffic efficiency of the first intelligent driving equipment; when the first intelligent driving equipment carries passengers, it can also improve the safety of the passengers. Driving experience.
  • determining the second derivation trajectory of the first object based on the third sampled acceleration and the second motion parameter includes: based on the third sampled acceleration, and the The speed and/or acceleration of the first object determines the fifth sub-deduction trajectory.
  • the speed of the first object at the end of the fifth sub-deduction trajectory is the fourth speed or zero, and the fourth speed is what the first object can travel.
  • determining the planned motion trajectory according to the first derivation trajectory and the second derivation trajectory includes: determining according to the first derivation trajectory and the second derivation trajectory A first game strategy, the first game strategy is used to instruct the first intelligent driving device to rush or give way to the first object at the first preset speed; when the first duration is greater than the first time threshold, according to the first A game strategy, the planned driving path and the second derivation trajectory determine the planned motion trajectory.
  • the first time threshold may be 0.1 seconds, or it may be 0.3 seconds, or the first time threshold may also take other values.
  • the game strategy may include: the first intelligent driving device grabs the first object at the first preset speed; the first intelligent driving device yields to the first object at the first preset speed; the first intelligent driving device gives way to the first object at the first preset speed; Grab or give way to the first object in a non-creeping manner.
  • control device of the first intelligent driving device determines the game strategy every fixed period of time.
  • the fixed duration may be 20 milliseconds, or may be 50 milliseconds.
  • the game strategy determined in the current frame may be different from the game strategy determined in the previous frame (or previous period).
  • this application proposes to determine the planned motion trajectory based on the first game strategy when it is determined that the duration of the first game strategy is greater than the first time threshold.
  • the first game strategy is that the first intelligent driving device grabs the first object at a first preset speed
  • the second game strategy is that the first intelligent driving device grabs or gives way to the first object in a non-creeping manner
  • the third The game strategy decision is that the first intelligent driving device yields to the first object at the first preset speed.
  • the game strategy is determined to be the first game strategy, If the determined game strategy is the second game strategy, then the game strategy needs to be determined as the first game strategy five times in a row before the planned movement trajectory can be further determined based on the first game strategy, the planned driving path and the second derivation trajectory. .
  • the value of the first time threshold may be determined based on the game strategy before the game strategy is determined to be the first game strategy. In some possible implementations, if the game strategy determined before the first game strategy is determined is the second game strategy, the first time threshold is the first threshold; if the game strategy determined before the first game strategy is determined is the third game strategy policy, the first time threshold is the second threshold. Wherein, the first threshold is smaller than the second threshold. For example, the first threshold may be 0.1 seconds, and the second threshold may be 0.3 seconds.
  • the planned motion trajectory is determined based on the second game strategy .
  • the third threshold is greater than the second threshold.
  • the first time threshold is the second threshold.
  • the first threshold is smaller than the second threshold.
  • the third threshold may be 0.5 seconds.
  • the duration of the game strategy after the change can be used to determine whether to re-plan the trajectory according to the game strategy.
  • the duration of the game strategy after the change does not meet the conditions, the movement trajectory will not be re-planned. , can avoid frequent jumps in game strategies and help improve the stability of planned motion trajectories.
  • determining the planned motion trajectory based on the first game strategy, the planned driving path and the second derivation trajectory includes: based on the planned driving path and the third derivation trajectory.
  • the second derivation trajectory determines the time period when the first position space of the first object occupies the second position space of the first intelligent driving device; based on the first game strategy, the plan is determined according to the second position space and the time period.
  • the planned movement trajectory includes the trajectory of the first intelligent driving device traveling at the first preset speed in the second position space, or the planned movement trajectory includes the first intelligent driving device traveling at the first preset speed within the time period. The trajectory of traveling at the first preset speed.
  • determining the planned motion trajectory according to the second position space and the time period includes: determining according to the first game strategy, the second position space and the distance threshold safedis In the third position space, the first time period is determined according to the time period and the time threshold TimeGap; the planned motion trajectory is determined according to the first time period, the first preset speed and the third position space.
  • the method before determining the first game strategy based on the first deduction trajectory and the second deduction trajectory, the method further includes: determining the first deduction trajectory and the second deduction trajectory.
  • the strategy cost of the pair of derivation trajectories composed of the second derivation trajectory is the smallest.
  • the strategy cost may include a safety strategy cost, and may also include at least one of a comfort strategy cost, a passability strategy cost, and a right-of-way strategy cost.
  • the safety policy cost is used to characterize the driving safety of the moving object. The lower the safety, the higher the policy cost; the comfort policy cost is used to represent the comfort of the user in the moving object, generally the acceleration change rate of the moving object. The larger the value, the worse the comfort and the higher the strategy cost; the passability strategy cost is used to characterize the passability of the moving object through the conflict point. For example, the time when the first derivation trajectory of the first intelligent driving device passes the conflict point The greater the difference with the calibration time, the worse the passability and the higher the strategy cost.
  • the calibration time is deduced when the lateral offset of the first intelligent driving device is 0 and the sampling acceleration is 0.
  • the time when the trajectory of the first intelligent driving device passes the conflict point; the right-of-way strategy cost is used to represent whether to change the motion state of an object with a higher right-of-way. If the pair of the first derivation trajectory and the second derivation trajectory makes the right-of-way higher If the motion state of the object changes, the strategy cost will be higher.
  • moving object may be the first intelligent driving device or the first object.
  • the strategy cost may only include the safety strategy cost; or, in addition to the safety strategy cost, the strategy cost may also include the comfort strategy cost and passability strategy of the first intelligent driving device and the first object. Cost and right-of-way strategy cost, in which the weight of each strategy cost can be different. In some possible implementations, the weight of each strategy cost can also change with the driving scenario.
  • the minimum value of the policy cost can be 0.
  • the first object as an intelligent driving device as an example, according to the first derivation trajectory and the second derivation trajectory, it is determined that the first object deviates from the first intelligent driving device in the traveling direction of the first object.
  • the overall strategy cost can be determined based on the speed of the first object at the collision location. It should be understood that if the strategy cost only includes the security strategy cost, then when multiple sets of first derivation trajectory and second derivation trajectory pairs are deduced, the first derivation trajectory and the second derivation trajectory pair with the lowest security strategy cost are selected to determine the third derivation trajectory.
  • a game strategy is a game strategy cost.
  • the strategy cost of the first deduction trajectory and the second deduction trajectory pair it is possible to evaluate the strategy cost of the first deduction trajectory and the second deduction trajectory pair, and determine which first deduction trajectory/second deduction trajectory pair to use for trajectory planning of the first intelligent driving device based on the strategy cost, as follows: Helps improve the rationality of the planned motion trajectory of the first intelligent driving equipment. Furthermore, the planned motion trajectory can be more accurately evaluated by combining driving scenarios and multiple strategy cost evaluation dimensions.
  • determining the first derivation trajectory of the first intelligent driving device according to the first sampled acceleration and the first motion parameter includes: according to the first intelligent driving device The lateral offset, the first sampled acceleration and the first motion parameter determine the first derivation trajectory, and the lateral offset is an offset perpendicular to the traveling direction of the first intelligent driving device.
  • a sub-derivative trajectory (or can also be called a longitudinal trajectory) of the first intelligent driving device is determined based on the first sampled acceleration and the first motion parameter, that is, the first intelligent driving device is in parallel The trajectory in the driving direction of the first intelligent driving device.
  • the first derivation trajectory of the first intelligent driving device can be determined.
  • the first derivation trajectory is the coordinate of the first intelligent driving device changing with time.
  • the set of points is the result of the fusion of the longitudinal trajectory of the first intelligent driving device and the lateral offset of the first intelligent driving device.
  • the second derivation trajectory may also be determined based on the lateral offset of the first object, the second sampled acceleration, and the second motion parameter.
  • the method before obtaining the predicted motion trajectory and the second motion parameter of the first object, the method further includes: determining the relationship between the first object and the first intelligent driving device. The distance between them is less than or equal to the first distance threshold and greater than or equal to the second distance threshold.
  • the first distance threshold may be 100 meters, or it may be 200 meters, or it may be other values.
  • the second distance threshold may be determined based on closest in-path vehicle (closest in-path vehicle, CIPV) screening. It should be understood that if the second distance threshold is determined through CIPV screening, the second distance threshold may be different when the first object is located at different directions around the first intelligent driving device.
  • the second distance threshold can also be other time distances and/or spatial distances. For example, if it is a time distance, it can be 2 seconds, or 3 seconds, or it can also be other values; if is the spatial distance, it can be 1.5 meters, or 2 meters, or other values.
  • the distance between an object and the first intelligent driving device is greater than or equal to the second distance threshold. That is to say, when the distance between any first object around the first intelligent driving device and the first intelligent driving device is less than the second distance threshold, the automatic driving decision and trajectory planning of the first intelligent driving device will not be performed.
  • a method of motion trajectory planning may include: acquiring the planned driving path and first motion parameters of the first intelligent driving device; acquiring the predicted motion trajectory and second motion parameters of the first object; When the planned driving path intersects with the predicted movement trajectory, the planned movement trajectory of the first intelligent driving device is determined according to the first movement parameter and the second movement parameter, and the planned movement trajectory is used to instruct the first intelligent driving device Pass through the conflict area at a first preset speed, or be used to instruct the first intelligent driving device to drive at the first preset speed when the first object passes through the conflict area, where the conflict area is the first intelligent driving equipment The area where the planned driving path overlaps with the deduced trajectory of the first object; and the first intelligent driving device is controlled to drive according to the planned movement trajectory.
  • a method of motion trajectory planning includes: acquiring the planned driving path and first motion parameters of the first intelligent driving device; acquiring the predicted motion trajectory and second motion parameters of the first object; When the planned driving path intersects with the predicted motion trajectory, a first game strategy is determined based on the first motion parameter and the second motion parameter. The first game strategy is used to instruct the first intelligent driving device to drive at a first preset speed. Steal or give way to the first object; determine the planned movement trajectory of the first intelligent driving device according to the first game strategy.
  • the first game strategy decision of the first intelligent driving device to rush or yield to the first object at the first preset speed is to expect the first intelligent driving device to pass through the conflict area at the first preset speed, or to first The intelligent driving device drives at a first preset speed when the first object passes through the conflict area.
  • the speed of the first intelligent driving device when passing through the conflict area may not necessarily reach the first preset speed, or may be greater than the first preset speed. The following is a detailed description of the scenarios that may occur during actual driving of the first intelligent driving device by combining the two major types of situations.
  • the situation where the first intelligent driving device passes the conflict area before the first object can be further refined as:
  • the first intelligent driving device passes through the conflict area while accelerating from a driving speed lower than the first preset speed to the first preset speed. Further, since the interaction between the first intelligent driving device and the first object ends after the first intelligent driving device passes through the conflict area, therefore, after the first intelligent driving device passes through the conflict area, a decision can be made to plan the first intelligent driving device to Continue driving at a driving speed higher than the first preset speed.
  • the speed has reached the first preset speed (it may be decelerating from a driving speed higher than the first preset speed to the first preset speed, or it may be decelerating from a speed lower than the first preset speed). (the driving speed of a preset speed accelerates to the first preset speed), then the first intelligent driving device passes through the conflict area at the first preset speed. Further, since the interaction between the first intelligent driving device and the first object ends after the first intelligent driving device passes through the conflict area, therefore, after the first intelligent driving device passes through the conflict area, a decision can be made to plan the first intelligent driving device to Continue driving at a driving speed higher than the first preset speed.
  • the first intelligent driving device passes through the conflict area during the process of decelerating from a driving speed higher than the first preset speed to the first preset speed. Further, since the interaction between the first intelligent driving device and the first object ends after the first intelligent driving device passes through the conflict area, therefore, after the first intelligent driving device passes through the conflict area, a decision can be made to plan the end of the first intelligent driving device In the deceleration state, it is controlled to continue driving at a speed higher than the first preset speed.
  • the first intelligent driving device robs the first object at a driving speed less than or equal to the first preset speed; in the third case, the first intelligent driving device drives at a speed less than or equal to the first preset speed; A driving speed higher than the first preset speed overtakes the first object.
  • the situation where the first intelligent driving device passes the first object after passing through the conflict area can be further refined as:
  • the first intelligent driving device When the first intelligent driving device is driving toward the conflict area, it is in a state of accelerating from a driving speed lower than the first preset speed to the first preset speed. At this time, the first object passes through the conflict area. Further, since the interaction between the first intelligent driving device and the first object ends after the first object passes through the conflict area, after it is determined that the first object passes through the conflict area, a decision can be made to plan the first intelligent driving device to be higher than the first object. Continue driving at a preset speed.
  • the first intelligent driving device When the first intelligent driving device is driving toward the conflict area at the first preset speed, the first object passes through the conflict area. Further, since the interaction between the first intelligent driving device and the first object ends after the first object passes through the conflict area, after it is determined that the first object passes through the conflict area, a decision can be made to plan the first intelligent driving device to be higher than the first object. Continue driving at a preset speed.
  • the first intelligent driving device When the first intelligent driving device is driving toward the conflict area, it is in a state of decelerating from a driving speed higher than the first preset speed to the first preset speed. At this time, the first object passes through the conflict area. Further, since the interaction between the first intelligent driving device and the first object ends after the first object passes through the conflict area, after it is determined that the first object passes through the conflict area, a decision can be made to plan the first intelligent driving device to be higher than the first object. Continue driving at a preset speed.
  • the first intelligent driving device yields to the first object at a driving speed less than or equal to the first preset speed; in the third case, the first intelligent driving device yields to the first object at a high speed Yield to the first object at a driving speed of the first preset speed.
  • the first game strategy determines that the first intelligent driving device rushes to the first object at the first preset speed or gives way to the first object
  • its speed may be A preset speed, or may be higher or lower than the first preset speed.
  • the movement behavior of the first object will no longer affect the decision-making and planning of the movement trajectory of the first intelligent driving device. It should be understood that after the interaction between the first intelligent driving device and the first object is completed, other objects that are too close to the first intelligent driving device may appear, and the first intelligent driving device may also stop in time after passing through the conflict area. , instead of continuing to drive at a speed higher than the first preset speed.
  • a device for motion trajectory planning may include: an acquisition unit for acquiring the planned driving path and first motion parameters of the first intelligent driving device; acquiring the predicted motion trajectory of the first object and a second motion parameter; a processing unit configured to determine the planned motion trajectory of the first intelligent driving device based on the first motion parameter and the second motion parameter when the planned driving path intersects with the predicted motion trajectory; the planning
  • the movement trajectory includes the movement trajectory in which the first intelligent driving device rushes or yields to the first object at a first preset speed; the first intelligent driving device is controlled to drive according to the planned movement trajectory.
  • the processing unit is specifically configured to: determine the first derivation trajectory of the first intelligent driving device according to the first sampled acceleration and the first motion parameter, the first The sampling acceleration is the possible acceleration of the first intelligent driving device determined in the sampling space; the second derivation trajectory of the first object is determined according to the second sampling acceleration and the second motion parameter, and the second sampling acceleration is the The possible acceleration of the first object determined in space; the first derivation trajectory and the second derivation trajectory indicate that the first intelligent driving device and the first object will not collide, or the first object will not collide with the first object at the first derivation trajectory.
  • the planned movement is determined based on the first derivation trajectory and the second derivation trajectory.
  • the first moment is the moment after the current moment.
  • the first motion parameter includes the speed and/or acceleration of the first intelligent driving device
  • the processing unit is specifically configured to: according to the first sampled acceleration, and The speed and/or acceleration of the first intelligent driving device determines the first sub-deduction trajectory, and the speed of the first intelligent driving device at the end of the first sub-deduction trajectory is the first preset speed; according to the first preset speed Determine a second sub-deduction trajectory; the first derivation trajectory includes the first sub-deduction trajectory and the second sub-deduction trajectory.
  • the processing unit is specifically configured to: determine a first game strategy based on the first derivation trajectory and the second derivation trajectory, and the first game strategy is used to indicate the The first intelligent driving device grabs or gives way to the first object at the first preset speed; when the first duration is greater than the first time threshold, according to the first game strategy, the planned driving path and the second deduced trajectory Determine the planned motion trajectory.
  • the processing unit is specifically configured to: determine the first position space of the first object occupying the first intelligent driving according to the planned driving path and the second derivation trajectory. The time period when the device is in the second position space; based on the first game strategy, the planned motion trajectory is determined according to the second position space and the time period, and the planned motion trajectory includes the first intelligent driving device in the second position The trajectory of the space traveling at the first preset speed, or the planned movement trajectory includes the trajectory of the first intelligent driving device traveling at the first preset speed within the time period.
  • the processing unit is specifically configured to: determine a strategy cost of a pair of derivation trajectories composed of the first derivation trajectory and the second derivation trajectory that minimizes the strategy cost.
  • the processing unit is specifically configured to: determine the first motion parameter based on the lateral offset of the first intelligent driving device, the first sampled acceleration, and the first motion parameter.
  • the trajectory is deduced, and the lateral offset is an offset perpendicular to the traveling direction of the first intelligent driving device.
  • the processing unit is further configured to: determine that the distance between the first object and the first intelligent driving device is less than or equal to the first distance threshold and greater than or equal to the first distance threshold. equal to the second distance threshold.
  • a device for motion trajectory planning may include: an acquisition unit for acquiring the planned driving path and the first motion parameter of the first intelligent driving device; and acquiring the predicted motion trajectory and the first motion parameter of the first object. Two motion parameters; a processing unit, configured to determine the planned motion trajectory of the first intelligent driving device according to the first motion parameter and the second motion parameter when the planned driving path and the predicted motion trajectory intersect, and the planned motion
  • the trajectory is used to instruct the first intelligent driving device to pass through the conflict area at a first preset speed, or to instruct the first intelligent driving device to drive at the first preset speed when the first object passes through the conflict area, wherein,
  • the conflict area is an area where the planned driving path of the first intelligent driving device overlaps with the deduced trajectory of the first object; the first intelligent driving device is controlled to drive according to the planned movement trajectory.
  • a device for motion trajectory planning includes: an acquisition module for acquiring the planned driving path and first motion parameters of the first intelligent driving device; and acquiring the predicted motion trajectory of the first object and the second motion parameter. Motion parameters; a processing module configured to determine a first game strategy based on the first motion parameter and the second motion parameter when the planned driving path intersects with the predicted motion trajectory, and the first game strategy is used to indicate the third game strategy.
  • An intelligent driving device rushes or yields to the first object at a first preset speed; and determines the planned motion trajectory of the first intelligent driving device according to the first game strategy.
  • a seventh aspect provides a device for motion trajectory planning.
  • the device includes: a memory for storing a program; a processor for executing the program stored in the memory.
  • the processor is configured to execute the above The method in any possible implementation manner from the first aspect to the third aspect.
  • An eighth aspect provides an intelligent driving device, which includes the device in any one of the implementations of the fourth aspect to the seventh aspect.
  • the intelligent driving device is a vehicle.
  • a computer program product includes: computer program code.
  • the computer program code When the computer program code is run on a computer, it enables the computer to execute any one of the first to third aspects. method within the method.
  • the above computer program code may be stored in whole or in part on the first storage medium, where the first storage medium may be packaged together with the processor, or may be packaged separately from the processor. This is not the case in the embodiments of this application. Specific limitations.
  • a computer-readable medium stores program codes.
  • the computer program codes When the computer program codes are run on a computer, the computer can perform any one of the first to third aspects. Methods in the implementation.
  • a chip in an eleventh aspect, includes a processor for calling a computer program or computer instructions stored in a memory, so that the processor executes any of the possible implementations of the first to third aspects above. method within the method.
  • the processor is coupled to the memory through an interface.
  • the chip system further includes a memory, and a computer program or computer instructions are stored in the memory.
  • Figure 1 is a functional block diagram of an intelligent driving device provided by an embodiment of the present application.
  • Figure 2 is a schematic diagram of the sensing ranges of various sensors provided by the embodiment of the present application.
  • Figure 3 is a schematic diagram of the system architecture required for the implementation of a motion trajectory planning method provided by an embodiment of the present application.
  • FIG. 4 is a schematic diagram of the system architecture required for the implementation of a motion trajectory planning method provided by an embodiment of the present application.
  • Figure 5 is a schematic flow chart of a motion trajectory planning method provided by an embodiment of the present application.
  • FIG. 6 is a schematic diagram of an application scenario of a motion trajectory planning method provided by an embodiment of the present application.
  • FIG. 7 is a schematic diagram of a CIPV provided by an embodiment of the present application.
  • Figure 8 is a schematic flow chart of a motion trajectory planning method provided by an embodiment of the present application.
  • Figure 9 is a schematic diagram of a sampling space provided by an embodiment of the present application.
  • Figure 10 is a demonstration diagram of lateral path estimation provided by an embodiment of the present application.
  • Figure 11 is a longitudinal trajectory push demonstration diagram provided by an embodiment of the present application.
  • Figure 12 is a schematic side view of a vehicle provided by an embodiment of the present application.
  • Figure 13 is a schematic diagram of the game strategy application scenario provided by the embodiment of the present application.
  • Figure 14 is a schematic flow chart of the game strategy stabilization processing provided by the embodiment of the present application.
  • Figure 15 is a schematic diagram of the movement trajectory of a self-vehicle and a gaming target provided by an embodiment of the present application.
  • Figure 16 is a schematic diagram of a creeping yield target constraint provided by an embodiment of the present application.
  • Figure 17 is a schematic diagram of a crawling target constraint provided by an embodiment of the present application.
  • Figure 18 is a schematic flow chart of a motion trajectory planning method provided by an embodiment of the present application.
  • Figure 19 is a schematic block diagram of a motion trajectory planning device provided by an embodiment of the present application.
  • Figure 20 is a schematic block diagram of a motion trajectory planning device provided by an embodiment of the present application.
  • Prefixes such as “first” and “second” are used in the embodiments of this application only to distinguish different description objects, and have no limiting effect on the position, order, priority, quantity or content of the described objects.
  • the use of ordinal words and other prefixes used to distinguish the described objects does not limit the described objects.
  • Words constitute redundant restrictions.
  • at least one means one or more
  • plural means two or more.
  • the character "/" generally indicates that the related objects are in an "or” relationship.
  • At least one of the following or similar expressions thereof refers to any combination of these items, including any combination of a single item (items) or a plurality of items (items).
  • at least one of a, b, or c can mean: a, b, c, a-b, a-c, b-c, or a-b-c, where a, b, c can be single or multiple .
  • FIG. 1 is a functional block diagram of the intelligent driving device 100 provided by the embodiment of the present application.
  • the smart driving device 100 may include a sensing system 120 , a display device 130 and a computing platform 150 , where the sensing system 120 may include several types of sensors that sense information about the environment around the smart driving device 100 .
  • the sensing system 120 may include a positioning system.
  • the positioning system may be a global positioning system (GPS), a Beidou system or other positioning systems, an inertial measurement unit (IMU), a lidar, a millimeter One or more of wave radar, ultrasonic radar and camera device.
  • GPS global positioning system
  • IMU inertial measurement unit
  • lidar a millimeter One or more of wave radar, ultrasonic radar and camera device.
  • the computing platform 150 may include processors 151 to 15n (n is a positive integer).
  • the processor is a circuit with signal processing capabilities.
  • the processor may be a circuit with instruction reading and execution capabilities.
  • CPU central processing unit
  • microprocessor microprocessor
  • GPU graphics processing unit
  • DSP digital signal processor
  • the processor can realize certain functions through the logical relationship of the hardware circuit. The logical relationship of the hardware circuit is fixed or can be reconstructed.
  • the processor is an application-specific integrated circuit (application-specific integrated circuit).
  • Hardware circuits implemented by ASIC or programmable logic device (PLD), such as field programmable gate array (FPGA).
  • PLD programmable logic device
  • FPGA field programmable gate array
  • the process of the processor loading the configuration file and realizing the hardware circuit configuration can be understood as the process of the processor loading instructions to realize the functions of some or all of the above units.
  • it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as a neural network processing unit (NPU), tensor processing unit (TPU), deep learning processing Unit (deep learning processing unit, DPU), etc.
  • the computing platform 150 may also include a memory, which is used to store instructions. Some or all of the processors 151 to 15n may call instructions in the memory and execute the instructions to implement corresponding functions.
  • the smart driving device 100 may include an advanced driving assist system (ADAS).
  • ADAS utilizes a variety of sensors in the perception system 120 (including but not limited to: lidar, millimeter wave radar, camera device, ultrasonic sensor, global positioning System, inertial measurement unit) obtains information from around the intelligent driving equipment, analyzes and processes the obtained information, and implements functions such as obstacle perception, target recognition, intelligent driving equipment positioning, path planning, driver monitoring/reminder, etc., thereby Improve the safety, automation and comfort of driving with intelligent driving equipment.
  • ADAS advanced driving assist system
  • FIG 2 shows a schematic diagram of the sensing range of various sensors.
  • the sensors can include, for example, lidar, millimeter wave radar, camera devices, and ultrasonic sensors in the perception system 120 shown in Figure 1.
  • Millimeter wave radar can be divided into long-range radar and Medium/short range radar.
  • the sensing range of lidar is about 80-150 meters
  • the sensing range of long-range millimeter wave radar is about 1-250 meters
  • the sensing range of medium/short-range millimeter wave radar is about 30-120 meters
  • the sensing range of ultrasonic radar is about 50-200 meters
  • the sensing range of ultrasonic radar is about 0-5 meters.
  • L0 level is no automation
  • L1 level is driving support
  • L2 level is partial automation
  • L3 level is conditional automation
  • L4 level is high automation
  • L5 level is complete automation.
  • the tasks of monitoring and responding to road conditions from L1 to L3 are jointly completed by the driver and the system, and the driver is required to take over dynamic driving tasks.
  • L4 and L5 levels allow the driver to completely transform into a passenger role.
  • ADAS can implement mainly include but are not limited to: adaptive cruise, automatic emergency braking, automatic parking, blind spot monitoring, front intersection traffic warning/braking, rear intersection traffic warning/braking, and front collision warning. , lane departure warning, lane keeping assist, rear collision avoidance warning, traffic sign recognition, traffic jam assist, highway assist, etc.
  • L0-L5 autonomous driving levels
  • the methods for planning autonomous driving trajectories are usually based on FSM and game-based object decision planning.
  • the above decision planning will be greatly affected by the trajectory prediction accuracy of the game target.
  • the game target is a pedestrian or an illegal object.
  • Unreasonable trajectory prediction of the game goal will cause the autonomous vehicle to have problems such as unexpected, frequent light and heavy braking, false braking, and continued conservative yielding, or cause the game goal and the autonomous vehicle to stop and/or start at the same time.
  • Typical scenarios are roads with mixed traffic of motor vehicles and non-motor vehicles or intersections without traffic lights, resulting in low traffic efficiency of autonomous vehicles and poor user driving experience.
  • embodiments of the present application provide a method and device for motion trajectory planning and intelligent driving equipment, which can perform vehicle trajectory planning based on the semantic-level decision-making tags of the game strategy and the derivation trajectory of the game goal, while ensuring safety and complying with traffic regulations. Under the premise of realizing the behavior of the vehicle actively maintaining low-speed creeping during the autonomous driving process to test the game target.
  • the vehicle When the vehicle maintains the low-speed creeping state, it is in a low-speed state and is not easily affected by the prediction accuracy of the game target trajectory, so that it can Avoiding problems such as frequent false braking, light braking, heavy braking, and false braking will help improve the user's driving experience.
  • non-motor vehicles involved in this application refer to vehicles that are driven by human or animal power and travel on the road, and although they are driven by a power unit, their designed maximum speed, empty vehicle mass, and overall dimensions comply with relevant national standards. Motorized wheelchairs, electric bicycles and other means of transportation for disabled persons; motor vehicles refer to means of transportation driven or towed by their own power devices.
  • the "creep” involved in the embodiments of this application is a state of traveling at a speed lower than or equal to a certain speed.
  • the "certain speed” may be 10 km/h, or it may be 5 km/h. , or it can be other values. It should be understood that when the self-vehicle is in a creeping state, the planned movement trajectory of the self-vehicle is not easily affected by the prediction accuracy of the game target trajectory.
  • Figure 3 shows a system architecture diagram of trajectory planning provided by an embodiment of the present application.
  • the trajectory planning system includes a perception module, a decision planning module, a control module, a parameter identification module and an actuator.
  • the perception module may include one or more camera devices in the perception system 120 shown in Figure 1, or one or more radar sensors, used to collect the surrounding environment information of the intelligent driving equipment, and the real-time motion parameters of the intelligent driving equipment. etc., the perception module can also process the collected surrounding environment information, and establish a world model composed of roads, obstacles, etc. for the downstream modules (i.e., decision-making planning module and control module); the decision-making planning module and control module can be as shown in Figure 1 One or more processors in the computing platform 150.
  • the decision planning module is used to determine the game strategy of the intelligent driving device based on the surrounding environment information, generate the derivation trajectory of the game goal, and generate the derivation trajectory according to the game strategy and the game goal.
  • Determine the planned motion trajectory and the control module is used to calculate the corresponding control quantities based on the planned motion trajectory, and output the above control quantities to the actuator.
  • the actuator executes the control quantity, the intelligent driving equipment is controlled to drive according to the planned motion trajectory.
  • the actuator may also include the steering and braking control systems in the intelligent driving device 100 .
  • the decision planning module may include a game goal screening module, an interactive game decision module and a motion planning module.
  • the game goal screening module is used to screen game goals;
  • the interactive game decision-making module is used to deduce the derivation trajectory of the game goal and the derivation trajectory of the self-car based on the screened game goals, and based on the derivation trajectory of the game goal and the derivation trajectory of the self-car.
  • the trajectory generates a game strategy (or semantic-level decision label);
  • the motion planning module is used to plan the vehicle's motion trajectory based on the game strategy and the derivation trajectory of the game goal, and generate the planned motion trajectory.
  • FIG. 5 shows a schematic flow chart of a motion trajectory planning method 500 provided by the embodiment of the present application.
  • This method 500 can be applied to the intelligent driving device shown in Figure 1, or can also be used in the intelligent driving equipment shown in Figure 3 or Figure 4. system execution as shown.
  • the method 500 may be performed by the game goal screening module in FIG. 4 .
  • the following uses the intelligent driving device as a vehicle as an example to introduce method 500. It should be understood that the steps or operations of the motion trajectory planning method shown in FIG. 5 are only exemplary, and the embodiments of the present application may also perform other operations or modifications of each operation in FIG. 5 .
  • the method 500 includes:
  • S501 Obtain the motion status information and road information of the interactive object whose distance from the own vehicle is less than or equal to the preset distance.
  • interactive objects include but are not limited to intelligent driving devices, pedestrians, electric vehicles, bicycles, and other objects.
  • the above-mentioned preset distance may be 100 meters (meter, m), or may be 200 m, or may be other distances.
  • the motion status information of the interactive object may include, but is not limited to, the predicted motion trajectory of the interactive object, and second motion parameters, such as position information, speed, acceleration, etc.
  • the motion status information of the interactive object may be calculated by the vehicle through the position information and/or speed information of the interactive object measured by a radar sensor, etc.; or, the motion status information of the interactive object may be calculated by the vehicle communicating with the infrastructure ( Vehicle to infrastructure, V2I) vehicle to outside information exchange (vehicle to everything, V2X), or it can be received through vehicle to vehicle communication (vehicle to vehicle, V2V), or it can be received through other methods .
  • V2I Vehicle to infrastructure
  • V2X vehicle to everything
  • V2V vehicle to vehicle to vehicle communication
  • V2V vehicle to vehicle communication
  • the self-vehicle can directly receive the predicted movement trajectory of the interactive object through V2I, V2V, or V2X; in another example, the predicted movement trajectory of the interactive object can be based on the position information and heading angle information of the interactive object received by the self-vehicle. etc. are calculated, in which the position information and heading angle information can be included in the basic vehicle safety messages (BSM) information, or can also be included in other information.
  • BSM basic vehicle safety messages
  • the predicted movement trajectory of the interactive object mentioned above can be understood as the predicted movement trajectory of the interactive object within a certain period of time in the future.
  • the "future period" may be 10 seconds, or may be 20 seconds.
  • road information includes but is not limited to lane line information of the current road, guidance rules for one or more lanes in the current road, etc.
  • the road information can be received by the own vehicle through V2I, V2V, or V2X, or it can also be collected and calculated by the own vehicle through radar sensors and/or camera devices.
  • S502 Determine the interactive object whose predicted movement trajectory intersects with the planned driving path of the vehicle as the preliminary screening game target.
  • the planned driving path can be obtained through the control module of the own vehicle.
  • the motion state of the interactive object can be determined based on the predicted motion trajectory of the interactive object, and further, the preliminary screening game goal can be determined based on the motion state of the interactive object. For example, when the interactive object travels along the predicted motion trajectory, the motion state of the interactive object is considered clear, and when the predicted motion trajectory of the interactive object intersects with the planned driving path of the own vehicle, the interactive object is considered to be the target of the preliminary screening game.
  • the self-vehicle is going straight, and the interactive object is interfered by obstacles and needs to occupy the self-vehicle driving lane, or the interactive object merges from the right side of the self-vehicle.
  • the predicted motion of the interactive object can be determined
  • the self-vehicle goes straight at an intersection without indicator lights, and the interactive object turns left in the opposite lane of the intersection; or the self-vehicle turns left at an intersection without indicator lights, and the interactive object turns left in the opposite lane of the intersection.
  • the vehicle is traveling straight in the opposite direction, it can be considered that the predicted movement trajectory of the interactive object intersects with the planned driving path of the own vehicle, and the interactive object can be determined as the preliminary screening game target.
  • the new predicted movement trajectory of the interactive object can be determined based on the driving intention of the interactive object, and then the new prediction can be judged. Whether the motion trajectory intersects with the planned driving path of the own vehicle.
  • the driving intention of the interactive object can be determined according to a behavior prediction method based on semantic intention. Furthermore, when it is determined that the driving intention of the interactive object intersects with the motion trajectory of the own vehicle, the interactive object is determined to be the preliminary screening game target.
  • S503 Determine whether the preliminary screening game target is a dangerous target.
  • sending the autonomous driving decision-making and trajectory planning process means not performing the trajectory planning process provided by the embodiments of this application. It should be understood that during the specific implementation process, the self-vehicle can perform other driving decisions and trajectory planning.
  • the "dangerous target” involved in the embodiments of the present application refers to one that is too close in space and/or time to the own vehicle, so that it is not suitable to target the target including creeping targets provided by the embodiments of the present application.
  • Interactive objects for trajectory planning of execution strategies It should be understood that other autonomous driving decisions and/or trajectory planning may be performed for this dangerous target, such as performing automatic emergency braking.
  • the initial screening game target is a dangerous target by performing CIPV screening on the preliminary screening game target.
  • the schematic diagram of CIPV screening is shown in Figure 7. Based on safety considerations, an area that changes with speed and is symmetrical about the longitudinal symmetry plane of the vehicle (such as an isosceles trapezoid or vase shape) is set directly in front of the vehicle's driving direction to screen dangerous targets. .
  • the length of the CIPV screening area in the longitudinal direction that is, parallel to the direction of the own vehicle (the S direction in Figure 7), consists of three parts: ef, fg and gd, where ef is a constant threshold used to characterize the self-vehicle.
  • a certain longitudinal range from the front of the car to the rear is also the dangerous target screening area;
  • fg is the constant speed threshold of the own vehicle, which is used to represent when the own vehicle is traveling at the current speed, from when the driver receives the emergency stop signal to when the intelligent driving equipment is braked The distance traveled within the reaction time;
  • gd is the braking distance threshold of the own vehicle, which is used to represent the distance required for the vehicle to brake at the maximum deceleration under the current working conditions.
  • L ef represents the length of the ef segment
  • longitude_const_thresh represents the longitudinal constant threshold
  • L fg represents the length of the fg segment
  • v ego_current represents the speed of the vehicle at the current moment
  • t reaction represents the time when the driver receives the emergency stop signal to the intelligent driving device.
  • the reaction time when braking L gd represents the length of segment gd
  • a max represents the maximum acceleration of the vehicle decelerating under the current working conditions.
  • the CIPV screening area transverse direction that is, the direction perpendicular to the longitudinal symmetry plane of the own vehicle (the L direction in Figure 7), is determined by the lengths of ab and dc. Both ab and dc are transverse constant thresholds, and the length of ab can be greater than dc.
  • the transverse length of the CIPV screening area is from ab to dc along the longitudinal direction, and the transverse length can be reduced linearly or nonlinearly along the longitudinal direction.
  • the schematic diagram of the CIPV screening area shown in Figure 7 shows the linear reduction of the transverse length.
  • the longitudinal constant threshold such as ef
  • the transverse constant threshold such as ab, dc
  • the degree of change in transverse length reduction self-vehicle speed, self-vehicle acceleration, driver reaction time, etc.
  • Change the shape of the CIPV screening area to adapt to the vehicle model of the initial screening game target and the aggressiveness of the own car. For example, the larger the car model of the initial screening game target, the longer the length of ed, and/or the longer the length of ab and/or cd; or, the higher the radicalness of the own car, the shorter the length of ed, and/or Or the shorter the length of ab and/or cd.
  • L obj_l represents the lateral distance between the preliminary screening game target and the own vehicle
  • V obj_l represents the lateral speed of the preliminary screening game target.
  • V obj_s represents the longitudinal speed of the initial screening game target
  • V ego_s represents the longitudinal speed of the own vehicle.
  • the preliminary screening game target must meet the following conditions: (a) The preliminary screening game target is within the CIPV screening area; (b) The ttl of the preliminary screening game target is less than or equal to the first preset threshold ttl_thresh; (c) ⁇ V of the gaming target is less than or equal to the second preset threshold ⁇ V_thresh.
  • the first preset threshold ttl_thresh is determined based on the longitudinal position, car type and radicalness of the vehicle in the initial screening game target, and can be dynamically adjusted based on one or more of the above information;
  • the second preset threshold ⁇ V_thresh is based on the initial screening
  • the model of the game target and the degree of radicalness of the own car are determined, and can be dynamically adjusted according to the vehicle type of the initial screening game target and/or the degree of radicalness of the own car.
  • the trajectory planning based on the creep strategy of this application will not be performed on this "dangerous target” later, but instead, for example, automatic emergency braking (autonomous emergency braking) will be performed. AEB) and other operations; if the preliminary screening game target is determined not to be a "dangerous target”, then the trajectory planning based on the creep strategy of this application will continue to be performed on the preliminary screening game target.
  • S504 Determine the game strategy of the own car based on the preliminary screening game goals.
  • the preliminary screening game goals in this step are the "game goals" that need to be considered when performing trajectory planning for the self-vehicle based on the creeping strategy in subsequent embodiments (for example, method 800).
  • the method of motion trajectory planning provided by the embodiment of the present application can initially screen out the game goals that need to be considered when planning the trajectory of the self-vehicle based on the creep strategy based on trajectory conflicts, intention conflicts, and CIPV methods, and can ensure that it is relatively safe. Game strategy decision-making and trajectory planning under the environment will help improve driving safety.
  • FIG. 8 shows a schematic flow chart 800 of a motion trajectory planning method provided by an embodiment of the present application.
  • This method 800 can be applied to the intelligent driving device shown in Figure 1, or can also be used in the intelligent driving equipment shown in Figure 3 or Figure 4. system execution as shown.
  • the method 800 may be executed by the interactive game decision-making module in FIG. 4 .
  • the method 800 can be executed after the above-mentioned method 500.
  • the method 800 can be regarded as an extension of S504. It should be understood that the steps or operations of the motion trajectory planning method shown in FIG. 8 are only exemplary illustrations, and the embodiments of the present application may also perform other operations or modifications of each operation in FIG. 8 .
  • the method 800 includes:
  • the sampling space is a space used to describe the possible states of the own vehicle and/or the game target.
  • the above possible states include but are not limited to kinematic states such as speed, acceleration, acceleration change rate, etc.
  • the above possibilities may also include other kinematic states of the self-vehicle and/or game goals. It should be understood that this possible state space has an upper limit of state values and a lower limit of state values.
  • the upper limit of state values is the maximum value of the state that the vehicle may reach, such as the maximum acceleration that the own vehicle and/or the game target can achieve, the maximum Speed, maximum acceleration change rate; the lower limit of this state value is the minimum value of the state that the own vehicle and/or the game goal may reach, such as the minimum acceleration, minimum speed, and minimum acceleration change rate that the vehicle can achieve.
  • a longitudinal acceleration sampling space is generated.
  • a lateral offset sampling space is generated.
  • the "longitudinal” reference in the embodiments of this application can be understood as the direction parallel to the longitudinal symmetry plane of the intelligent driving device in a plane parallel to the ground, that is, the direction of travel of the intelligent driving device;
  • the reference in the embodiments of this application to " “Lateral direction” can be understood as the direction perpendicular to the longitudinal symmetry plane of the intelligent driving device, that is, in a plane parallel to the ground, perpendicular to the direction of travel of the intelligent driving device.
  • the above-mentioned game goals can be understood as preliminary screening game goals that are not dangerous goals in the above-mentioned embodiments, or they can also be game goals screened out through other methods.
  • the samples in the lateral offset sampling space are used to represent the offset of the trajectory of the own vehicle or game target in the direction perpendicular to the longitudinal symmetry plane of the intelligent driving device.
  • Multiple longitudinal sampling spaces can be sampled within each lateral offset sampling space.
  • the longitudinal sampling space of the own vehicle can be divided into the longitudinal creep acceleration sampling space of the own vehicle and the longitudinal non-creep acceleration sampling space. The sampling intervals and sampling intervals of the two can be the same.
  • sampling in the longitudinal creep acceleration sampling space and sampling in the longitudinal non-creep acceleration sampling space can be distinguished by using "labels" and other methods.
  • the longitudinal acceleration sampling space of the game target is composed of the longitudinal non-creep acceleration sampling space.
  • the horizontal and vertical sampling space of the own vehicle and the horizontal and vertical sampling space of the game target are crossed and combined to form the final sampling space, as shown in Figure 9.
  • the lateral offset sampling intervals of the self-vehicle and the game target shown in Figure 9 are [minEgoLateraloffset, maxEgoLateraloffset] and [minObjLateraloffset, maxObjLateraloffset] respectively.
  • the self-vehicle longitudinal non-creep acceleration sampling space and longitudinal creep acceleration are The interval of the sampling space is [-4,3]m/s 2
  • the interval of the longitudinal acceleration sampling space of the game target is [-4,3]m/s 2
  • the sampling interval is all 1m/s 2
  • we can get 16*8 128 longitudinal sampling space
  • the specific sampling intervals of the horizontal and vertical sampling spaces can be other intervals and can be determined according to hardware computing performance and accuracy requirements.
  • Sampling based on the sampling space performs lateral path deduction on the self-vehicle based on the current position of the self-vehicle, and performs longitudinal trajectory deduction on the self-vehicle based on the current acceleration of the self-vehicle.
  • the lateral path and longitudinal trajectory of the vehicle are deduced respectively.
  • the lateral path includes the offset path curve shape, as shown in Figure 10;
  • the longitudinal trajectory includes the corresponding time (t ) and the relationship between the longitudinal distance (s), and finally generate multiple sets of derivation trajectories with time information based on the transverse paths and longitudinal trajectories.
  • s e is the longitudinal position (unit: m) of the vehicle when performing transverse path deduction, and this position changes with the deduction time.
  • l(s e ) is the lateral offset (unit: m) corresponding to the longitudinal position of the own vehicle s e ;
  • s e StartThresh is the longitudinal starting position of the own vehicle (unit: m).
  • the longitudinal starting position of the own vehicle can be understood is the position of the self-vehicle in the direction parallel to the longitudinal symmetry plane of the intelligent driving device at the moment when trajectory deduction is started; curEgoLateraloffset is the offset corresponding to the current orientation of the self-vehicle (unit: m); s e CubicCurveThresh is the end of the cubic curve connection longitudinal position (unit: m); egolateralOffset is the lateral offset (unit: m) of the end of the own vehicle's cubic curve connection, corresponding to the own vehicle's minEgoLateralOffset or maxEgoLateralOffset; s e StartThresh and s e CubicCurveThresh can be based on the vehicle model and radical The degree is dynamically adjusted.
  • the values of s e StartThresh and/or s e CubicCurveThresh are larger; and/or when the vehicle is more aggressive, the values of s e StartThresh and/or s e CubicCurveThresh are larger. Small.
  • a cubic curve is used to connect the path, and the tangent direction of the cubic curve at (s e StartThresh, curEgoLateralOffset) and (s e CubicCurveThresh, egolateralOffset) is consistent with intelligent driving
  • the driving direction of the equipment is parallel, where a, b, c, and d are coefficients of a cubic polynomial. Their specific values can be determined according to the vehicle model and degree of radicalness. This application does not limit their specific values.
  • the change of the longitudinal acceleration of the own vehicle with the deduction time can be shown in Figure 11, where the horizontal axis represents the deduction time t, and the vertical axis represents the acceleration of the own vehicle.
  • the longitudinal position of the self-vehicle at a certain moment during the trajectory deduction process can be determined based on the derived longitudinal acceleration and speed of the self-vehicle.
  • the longitudinal positions of the self-vehicle at multiple moments during the trajectory deduction process constitute the longitudinal trajectory of the self-vehicle.
  • the longitudinal acceleration deduction of the own vehicle can include the following four sections: (1) Delay section (t ⁇ [0, delayTime)): longitudinal trajectory deduction based on the current acceleration of the own vehicle currentAcc; (2) Uniform acceleration change Rate section (t ⁇ [delayTime, jerkChangeTime), hereinafter referred to as the uniform Jerk section): determine the acceleration change rate based on the acceleration at the beginning of the uniform Jerk section and the longitudinal sampling acceleration targetAcc, and perform longitudinal trajectory deduction according to the uniform Jerk; (3) Uniform acceleration section ( t ⁇ [jerkChangeTime,speedLimitTime)): Maintain the longitudinal sampling acceleration targetAcc for trajectory deduction; (4) Uniform speed section (t ⁇ speedLimitTime): When the derivation speed reaches the upper bound speed or lower bound speed, maintain the upper bound speed or lower bound speed for trajectory deduction deduction.
  • the above-mentioned uniform jerk section may be a process in which acceleration linearly increases, or may also be a process in which acceleration linearly decreases.
  • the upper bound speed of the longitudinal trajectory deduction can be the upper limit speed set by the own vehicle or the maximum speed limit of the road section the own vehicle travels, and the lower bound speed of the longitudinal trajectory deduction Can be 0.
  • the upper bound speed of the longitudinal trajectory deduction can be the creep speed
  • the lower speed limit of the longitudinal trajectory deduction can be 0
  • the creeping speed may be preset by the system, or may be set by the user himself. For example, it may be 5 km/h or 10 km/h, or may be other values. This is the case in the embodiment of the present application. No specific limitation is made.
  • deducing the longitudinal trajectory of the self-vehicle may only include one, two, or three of the above stages.
  • the longitudinal trajectory deduction can be performed starting from (3) the uniform acceleration section.
  • the trajectory can be deduced from (4) the uniform velocity segment.
  • Sampling based on the sampling space performs horizontal path deduction on the game target based on the current position of the game target, and performs longitudinal trajectory deduction on the game target based on the current acceleration of the game target.
  • the game target trajectory deduction method can be the same as the self-vehicle trajectory deduction method.
  • the following formula is used to deduce the lateral path of the game goal:
  • s o is the longitudinal position of the game target (unit: m), which changes with the deduction time; l (s o ) is the lateral offset corresponding to the longitudinal position of the game target (unit: m); s o StartThresh is the longitudinal position of the game target starting position; curObjLateralOffset is the offset corresponding to the current orientation of the game target (unit: m); s o CubicCurveThresh is the longitudinal position where the cubic curve connection ends (unit: m); objlateralOffset is the lateral offset where the cubic curve connection of the game target ends Shift (unit: m), corresponding to the minObjLateralOffset or maxObjLateralOffset of the game target; s o StartThresh and s o CubicCurveThresh can be dynamically adjusted according to the model and aggressiveness of the game target.
  • the values of so StartThresh and/or so CubicCurveThresh are larger; and/or when the degree of radicalness of the game target is high, the values of so StartThresh and/or so CubicCurveThresh are relatively large. Small.
  • a cubic curve is used to connect the path, and the cubic curve is at (s o StartThresh,objcurLateralOffset) and (s o CubicCurveTheesh,objlateralOffset) in the tangent direction to the intelligent driving device.
  • a, b, c, and d are the coefficients of the cubic polynomial. The values of these coefficients can be the same as the values of a, b, c, and d in the self-vehicle lateral path derivation formula, or they can be different.
  • the method of deducing the longitudinal trajectory of the game target can refer to the description in the derivation of the longitudinal trajectory of the self-vehicle.
  • the longitudinal acceleration deduction of the game target can include the following four sections: (1) Delay section (t ⁇ [0, delayTime)): perform longitudinal trajectory deduction based on the current acceleration currentAcc of the game target; (2) ) Uniform Jerk segment (t ⁇ [delayTime, jerkChangeTime)): Determine the acceleration change rate based on the acceleration at the beginning of the uniform Jerk segment and the longitudinal sampling acceleration targetAcc, and conduct longitudinal trajectory deduction according to uniform Jerk; (3) Uniform acceleration segment (t ⁇ [jerkChangeTime) ,speedLimitTime)): Keep the longitudinal sampling acceleration targetAcc for trajectory deduction; (4) Uniform speed section (t ⁇ speedLimitTime): When the deduction speed reaches the upper bound speed or lower bound speed, keep the upper bound speed or lower bound speed for trajectory deduction.
  • the upper bound speed of the longitudinal trajectory deduction of the game target can be the maximum speed limit of the road section traveled by the game target, and the lower bound speed of the longitudinal trajectory deduction can be 0.
  • the above-mentioned uniform jerk section may be a process in which acceleration linearly increases, or may also be a process in which acceleration linearly decreases.
  • the above four stages of deducing the longitudinal trajectory of the game goal are only exemplary.
  • the derivation of the longitudinal trajectory of the game goal may only include one, two, or three of the above stages.
  • the deduced trajectory of the self-vehicle can be determined based on the deduced lateral path and longitudinal trajectory of the self-vehicle (or game goal), and the derivation trajectory is the point where the coordinates of the self-car (or game goal) change with time.
  • a set of components that is, the result of the fusion of longitudinal trajectories and lateral offsets (or game goals).
  • the longitudinal sampling acceleration and lateral offset of a self-vehicle can be determined, as well as the longitudinal sampling acceleration and lateral offset of a gaming target. It should be understood that the derivation trajectory of the own vehicle and the derivation trajectory of the game target determined based on the above sampling constitute a derivation trajectory pair.
  • the strategy cost (hereinafter referred to as cost) is evaluated for all derivation trajectory pairs of the self-vehicle and the game goal.
  • cost the strategy cost
  • strategic benefit evaluation can be carried out from one or more of five dimensions such as safety, comfort, passability, right of way, and offset.
  • the safety cost can be determined based on the minimum distance between the self-vehicle derivation trajectory and the game target derivation trajectory based on the derivation trajectory, while taking into account the head direction and speed of the self-vehicle and the game target. The smaller the minimum distance with speed and direction, the greater the safety cost.
  • the collision type is distinguished as the first type of collision or the second type of collision according to the motion state and position and posture of the self-vehicle and the game target at the time of collision.
  • the first type of collision is abbreviated as "E2O” and the second type of collision is abbreviated as "O2E”.
  • the collision is determined to be a first-type collision; in the game When the target collides with the side or rear of the own vehicle in the direction of travel, the collision is deemed to be a type II collision. Or, when the game target is stationary, and the game target collides with the game target, no matter which part of the game target the collision position is, the collision will be considered a first-type collision; when the game target is stationary, the game target collides with the game target. In the event of a collision, no matter which part of the vehicle the collision location is, the collision will be deemed to be a type 2 collision.
  • the side panels of the vehicle include the left side panel and the right side panel of the vehicle body.
  • the left and right sides of the vehicle body (A, B, C pillars) and the outer covering are called the left and right sides of the vehicle body.
  • the area indicated by the dotted line shown in FIG. 12 is the left side of the vehicle body.
  • the A-pillar is set in the front of the side panel, connected to the front panel, and also provides the installation of the front door hinge and front windshield glass. It is the supporting column for the rotation of the front side door switch; the B-pillar is set in the middle of the side panel and provides the front seat belts.
  • the installation of the front door latch and the rear side door hinge is the supporting column for the rotation of the rear side door switch (some small commercial vehicles only have left and right front doors, and the B-pillar is located at the rear of the side panel); the C-pillar is set at the rear of the side panel, serving as a rear
  • the seat belts and rear side door lock buckles are installed, and are the support and installation pillars for the rear window glass or rear door; the D-pillar, together with the top cover rear cross member and other components, form the rear door frame, and are the pillars that provide support for the rear triangle window and rear door frame.
  • the above-mentioned vehicle body side panels are included in a side panel assembly, and the side panel assembly may also include inner panels and reinforcements.
  • the side panels of the vehicle also include left and right side panels of the vehicle body and wheels.
  • the left side panel of the vehicle includes the left side panel of the vehicle body and the left side wheel of the vehicle (for example, including the left front wheel and the left rear wheel);
  • the right side panel of the vehicle includes the right side panel of the vehicle body and the right side wheel of the vehicle (for example, Including right front wheel and right rear wheel).
  • the rear of the vehicle may be provided with a rear bumper or one side of the rear license plate.
  • the safety cost is 0.
  • the safety cost is an approximately infinite value
  • the safety cost can be determined based on the speed of the game target at the collision location.
  • the security cost calculation formula can be as follows:
  • Cost safety represents the safety cost
  • v represents the speed of the game target at the collision position when a collision occurs based on the vehicle's deduction trajectory and the game goal's deduction trajectory
  • vThresMin represents the occurrence of the collision between the vehicle's deduction trajectory and the game goal's deduction trajectory.
  • the lower speed threshold for safety cost penalty is applied along with the game target speed
  • vThresMax represents the upper speed threshold for safety cost penalty according to the game target speed when the derivation trajectory of the own vehicle collides with the game target's derivation trajectory
  • m is Weight value.
  • the weight value is related to the type of game goal.
  • the comfort cost of the deduced trajectory pair of the self-vehicle and the game target can be determined based on the acceleration change rate (jerk value) of the self-vehicle (or game target) in the deduced trajectory of the self-vehicle (or game target).
  • the calculation formula of the comfort cost of the own vehicle (or game target) can be:
  • Cost comfortable represents the comfort cost
  • jerk represents the acceleration change rate of the self-vehicle (or game goal)
  • jerkThresMin represents the lower limit threshold of the acceleration change rate of the self-vehicle (or game goal);
  • jerkThresMax represents the acceleration change rate of the self-vehicle (or game goal).
  • Acceleration change rate upper limit threshold when jerk is less than jerkThresMin, Cost comfortable takes 0; when jerk is greater than jerkThresMax, Cost comfortable takes 1. It should be understood that the better the comfort, the smaller the comfort cost.
  • the comfort cost consists of the self-vehicle comfort cost and the game target comfort cost. The weights of the self-vehicle comfort cost and the game target comfort cost in the comfort cost can be different.
  • the passability cost calculation method for the derivation trajectory pair of the self-vehicle and the game target can be as follows.
  • the calculation formula of the passability cost of the own vehicle (or game target) can be:
  • Cost passability represents the passability cost
  • deltaT represents the time when the current derivation trajectory of the self-vehicle (or game target) passes the conflict point, and the horizontal sampling of the self-vehicle (or game goal) corresponds to the current orientation and the longitudinal sampling is 0
  • tThresMin represents the lower threshold of the time difference
  • tThresMax represents the upper threshold of the time difference.
  • the passability cost is composed of the self-vehicle passability cost and the game goal passability cost.
  • the self-vehicle passability cost and the game goal passability cost can have different weights in the passability cost.
  • the right-of-way relationship between the self-vehicle and the game goal is determined based on the pair of deduced trajectories of the self-vehicle and the game goal. If the interactive game causes the intelligent driving equipment with high rights of way to change its motion state, then the change in the motion state should be punished with a higher right of way cost, that is, the intelligent driving equipment with high right of way tends not to change the current motion state.
  • the calculation of the right-of-way cost for the deduced trajectory pair of the self-vehicle and the game target can be as follows.
  • the calculation formula for the cost of the own vehicle (or game target) right-of-way can be:
  • Cost roadRight represents the road right cost
  • acc represents the current longitudinal sampling acceleration of the high-right-of-way intelligent driving device
  • accThresMin represents the lower limit threshold of longitudinal acceleration
  • accThresMax represents the upper limit threshold of longitudinal acceleration.
  • Cost roadRight takes 0; when acc is greater than accThresMin, Cost roadRight takes 1.
  • the right-of-way cost consists of the self-vehicle right-of-way cost and the game target right-of-way cost. The weight of the self-vehicle right-of-way cost and the game target right-of-way cost in the right-of-way cost can be different.
  • each cost can be weighted and summed to obtain the final cost of the current pair of derivation trajectories of the self-vehicle and the game goal.
  • one or more of the above costs can be used, or the above costs can be deformed and/or expanded on the basis of .
  • the initial game strategy is determined based on the derivation trajectory pair with the smallest strategy cost among all the derivation trajectory pairs of the self-car and the game target. For the vertical game strategy in the initial game strategy, if according to the trajectory deduction, the vehicle passes the conflict area before the game goal, then the vertical game strategy of the initial game strategy is the go-go game goal, and vice versa, the yield game goal is. In some possible implementations, it is determined whether the vertical game strategy is a creeping strategy or a non-creeping strategy based on the longitudinal sampling acceleration.
  • the longitudinal game strategy is the creep strategy
  • the longitudinal sampling acceleration of the own vehicle belongs to the own vehicle non-creep acceleration sampling space
  • the longitudinal game strategy is It is a non-creeping strategy.
  • the lateral game strategy in the initial game strategy if the lateral sampling offset of the own vehicle is not the offset corresponding to the current direction (curEgoLateralOffset), then the lateral game strategy of the optimal solution is the detour game goal, otherwise, the game goal is ignored.
  • the decision label of the grab game goal is GRABWAY (abbreviated as GW), and the decision label of the yield game goal is YIELD (abbreviated as YD);
  • the creep strategy if according to Trajectory deduction, if the vehicle passes the conflict area before the game target, it belongs to creeping and grabbing, otherwise, it belongs to creeping and giving way.
  • the decision label of the creeping and grabbing game goal is CREEP-FORWARD-GRABWAY (abbreviated as CFG)
  • the decision label of the creep-yielding game objective is CREEP-FORWARD-YIELD (abbreviated as CFY).
  • the decision label of bypassing the game goal is BYPASS (abbreviated as BP), and the decision label of ignoring the game goal is IGNORE (abbreviated as IG).
  • BYPASS abbreviated as BP
  • IGNORE abbreviated as IGNORE
  • the difference between the creeping strategy and the non-crawling strategy is that: when the motion planning module is planning the trajectory according to the creeping strategy, it tries to make the vehicle pass through the conflict area at a creeping speed; when planning the trajectory according to the creeping strategy, When planning the yield trajectory using the creep strategy, try to make the vehicle drive at creeping speed during the period when the game target passes through the conflict area; that is, it can be understood that the creep strategy instructs the motion planning module to provide the self-vehicle based on the creeping speed based on the creeping speed.
  • the vehicle plans its trajectory in the conflict area, or the trajectory of its own vehicle during the period when the game target passes through the conflict area; while the motion planning module is planning the trajectory of rushing or yielding according to the non-creep strategy, when the vehicle passes through the conflict area
  • the driving speed when the game target passes through the conflict area, or the driving speed of the vehicle during the period when the game target passes through the conflict area. It may rush to the game target at a very high speed, or it may stop and give way to the game target.
  • S806 Stabilize the initial game strategy to generate a final game strategy.
  • the non-creeping label jumps to the creeping label (YD/GW->CFY/CFG), the creeping label and the creeping yield
  • the difficulty of jumping between row labels (CFY ⁇ ->CFG) and jumping between creeping labels and non-creeping labels (CFY/CFG->YD/GW) increases in sequence.
  • intervalTime is the time interval between the previous frame and the current frame
  • minTimerThres, midTimerThres and maxTimerThres are all timer time thresholds and increase sequentially
  • ydToCfyTimer means that the decision label jumps from YD to CFY timer
  • ydToCfgTimer represents the timer for the decision label to jump from YD to CFG
  • cfyToYdTimer represents the timer for the decision label to jump from CFY to YD
  • cfyToGwTimer represents the timer for the decision label to jump from CFY to GW
  • cfyToCfgTimer represents the timer for the decision label to jump from CFY to CFG.
  • cfgToYdTimer represents the timer for the decision label to jump from CFG to YD
  • cfgToGwTimer represents the timer for the decision label to jump from CFG to GW
  • cfgToCfyTimer represents the timer for the decision label to jump from CFG to CFY
  • gwToCfyTimer represents the timer for the decision label to jump from GW to CFY
  • gwToCfgTimer indicates the timer for the decision tag to be jumped from GW to CFG.
  • minTimerThres, midTimerThres, and maxTimerThres can be 0.1 seconds, 0.3 seconds, and 0.5 seconds respectively, or the minTimerThres, midTimerThres, and maxTimerThres can also take other values.
  • the method of motion trajectory planning provided by the embodiment of the present application is based on refined horizontal and vertical sampling trajectories, considering the rationality of the path in the transverse direction, and distinguishing between creep acceleration sampling space and non-creep acceleration sampling space in the longitudinal direction. Sampling, the upper bound speed and lower bound speed of the trajectory derivation, can narrow the gap between the deduced trajectory and the actual motion trajectory. Under the premise of ensuring safety and complying with traffic regulations, the vehicle actively maintains low-speed crawling to test the game target.
  • the game target behavior is too aggressive, causing the time or space distance between the vehicle and the game target to be lower than the safe distance, Then it can ensure that the own vehicle transfers from creeping to the right of way (non-creeping and yielding); if the game target behavior is conservative, allowing the own vehicle to have room to grab the game goal, it can make the own vehicle find the right time to change from creeping to grabbing the right of way. (Not crawling and rushing).
  • the motion planning module After the interactive game decision-making module shown in Figure 4 generates the final game strategy, the motion planning module generates a speed curve that meets the kinematic requirements of the intelligent driving equipment based on the final game strategy.
  • the motion planning module determines the time period when the first position space of the game target occupies the second position space of the own car based on the planned driving path of the own vehicle and the deduced trajectory of the game goal, and uses a collision detection algorithm to generate a time-based Space constraints.
  • the game target occupies the space from s1 to s2 of the vehicle's planned driving path between time t1 and t2 (i.e., the above-mentioned "time period") (i.e., the above-mentioned "second position space").
  • t1 is the conflict start time and t2 is the conflict end time.
  • the starting position of creeping and the creeping duration are determined, and then the longitudinal trajectory of the self-vehicle is planned. If the final game strategy is to creep and give way, the position where creeping and giving way ends can be at s1, and the creeping duration can be between t1 and t2. Then, the starting point of creeping and giving way can be determined based on the position where creeping and giving way ends. For example, the creeping distance is determined based on the creeping speed and creeping duration. The creeping distance is determined based on the creeping distance and the position where the crawling ends. The current position of the car is used to plan the speed of the car from its current position to the position where it starts to creep.
  • the starting position of crawling can be s1, and the vehicle will end crawling at s2 at time t1, which is determined based on the creeping speed and creeping distance. Creeping duration, and then determine the time to start creeping based on the creeping duration and t1 time, and based on the current location of the vehicle, the location and time when the vehicle starts to creep, plan the vehicle from its current position to the time it starts to creep. The speed of grabbing the position.
  • the uncertainty of the game target behavior is considered, and a safe distance in time and space is introduced.
  • this design can increase the spatial distance between the vehicle and the game target, and increase the creeping safety to a certain extent. Further, as shown in (b) of FIG. 16 , within the creeping duration CFTime, the intelligent driving device keeps traveling at the creeping speed Cf_Velocity.
  • TimeGap1, TimeGap2, and TimeGap can be 0.5 seconds respectively, or they can be other values.
  • the three values can be the same or different; the safeDis, safeDis1, and safeDis2 can be 0.5 meters respectively, or they can be respectively is 1 meter, or it can be other values.
  • the three values can be the same or different.
  • Figure 18 shows a schematic flow chart of a motion trajectory planning method 1800 provided by the embodiment of the present application.
  • the method shown in this process can be applied to the intelligent driving equipment shown in Figure 1, or can also be used through Figure 3 or Figure 4 The system shown performs.
  • the method 1800 includes:
  • the first motion parameters include the speed and/or acceleration of the first intelligent driving device.
  • the first intelligent driving device may be the self-vehicle in the above embodiment, or may be other objects that can move autonomously, such as an intelligent robot.
  • the planned driving path of the first intelligent driving device may be obtained from the regulation module; the first motion parameter may be obtained from the perception system, and the first motion parameter may include but is not limited to speed, acceleration, The correspondence between the location and time of an intelligent driving device, etc.
  • S1820 Obtain the predicted motion trajectory and second motion parameters of the first object, where the second motion parameters include the speed and/or acceleration of the first object.
  • the subject that obtains the predicted movement trajectory and the second movement parameter of the first object and the subject that obtains the planned driving path and the first movement parameter of the first intelligent driving device may be the same subject.
  • the same subject may It is the first intelligent driving equipment.
  • the first object may be the game target in the above embodiment, or may be other objects that may affect the planned motion trajectory of the first intelligent driving device;
  • the second motion parameters may include but are not limited to speed, acceleration, The correspondence between the position of an object and time, etc.
  • the method of obtaining the predicted motion trajectory and the second motion parameter of the first object may refer to the description in the above embodiment, for example, the description in S501 of method 500, which will not be described again here.
  • the method for determining that the planned driving path intersects with the predicted movement trajectory may refer to the description in the above embodiment, for example, the description in S502 of method 500, which will not be described again.
  • the first game strategy may be the final game strategy in the above embodiment, and more specifically, it may be the creep strategy, CFY or CFG in the vertical game strategy in the above embodiment.
  • the first preset speed is greater than or equal to 3km/h and less than or equal to 15km/h.
  • the first preset speed may be the "creep speed" in the above embodiment, or may be other speeds that make the planned movement trajectory of the first intelligent driving device less susceptible to the influence of the first object trajectory prediction accuracy.
  • the method of determining the planned motion trajectory according to the first motion parameter and the second motion parameter may refer to the description in the above embodiments, for example, the description in method 800, which will not be described again here.
  • S1840 Control the first intelligent driving device to drive according to the planned motion trajectory.
  • determining the planned motion trajectory based on the first motion parameter and the second motion parameter includes: determining the first derivation trajectory of the first intelligent driving device based on the first sampled acceleration and the first motion parameter, and the third A sampled acceleration is the possible acceleration of the first intelligent driving device determined in the sampling space; the second deduced trajectory of the first object is determined based on the second sampled acceleration and the second motion parameter, and the second sampled acceleration is The possible acceleration of the first object determined in the sampling space; the first derivation trajectory and the second derivation trajectory indicate that the first intelligent driving device and the first object will not collide, or the first object will not collide with the first object in the first derivation trajectory.
  • the first derivation trajectory and the second derivation trajectory are determined according to the first derivation trajectory.
  • the first moment is the moment after the current moment.
  • the first sampling acceleration and the second sampling acceleration may be the longitudinal sampling acceleration in the above embodiment. More specifically, the first sampling acceleration may be the longitudinal sampling acceleration in the above creeping longitudinal sampling space.
  • the first derivation trajectory and the second derivation trajectory may be the derivation trajectories of the method 800 in the above embodiment.
  • the specific method of determining the first derivation trajectory and the second derivation trajectory may refer to S802 to S804 in the method 800. Description will not be repeated here.
  • the method for determining whether a collision occurs between the first intelligent driving device and the first object and the type of collision according to the first derivation trajectory and the second derivation trajectory may refer to the description in the above method 1100, which will not be described again here. .
  • step of determining the "deduction trajectory” involved in the embodiment of the present application can be completed by the interactive game decision-making module in the above embodiment, and the step of determining the "planned motion trajectory” involved in the embodiment of the present application can be This is accomplished by the motion planning module in the above embodiment.
  • the first motion parameter includes the speed and/or acceleration of the first intelligent driving device
  • determining the first derivation trajectory of the first intelligent driving device according to the first sampled acceleration and the first motion parameter includes:
  • the first sub-deduction trajectory is determined based on the first sampled acceleration and the speed and/or acceleration of the first intelligent driving device.
  • the speed of the first intelligent driving device at the end of the first sub-deduction trajectory is the first preset speed.
  • Determine the second sub-deduction trajectory according to the first preset speed, and the end point of the first sub-deduction trajectory is the starting point of the second sub-deduction trajectory;
  • the first sub-derivation trajectory may include the time delay segment (t ⁇ [0, delayTime)), the uniform acceleration change rate segment (t ⁇ [delayTime, jerkChangeTime) and the uniform acceleration segment (t ⁇ [ The trajectory of at least one segment in jerkChangeTime, speedLimitTime));
  • the second sub-deduction trajectory may be the trajectory of the constant speed segment in the above embodiment.
  • the first derivation trajectory also includes a third sub-deduction trajectory
  • the third sub-deduction trajectory may include a trajectory in which the driving speed is reduced from the first preset speed to zero (ie, braking).
  • determining the planned motion trajectory based on the first deduction trajectory and the second deduction trajectory includes: determining a first game strategy based on the first deduction trajectory and the second deduction trajectory, and the first game strategy is used to Instruct the first intelligent driving device to rush or yield to the first object at the first preset speed; when the first duration is greater than the first time threshold, according to the first game strategy, the planned driving path and the second The deduced trajectory determines the planned motion trajectory.
  • the duration of the first game strategy may be the duration of any one of the timers ydToCfgTimer, ydToCfyTimer, gwToCfgTimer, and gwToCfyTimer in the above embodiment; or, the duration of the first game strategy may be the duration of cfgToCfyTimer in the above embodiment. Or the duration of any timer in cfyToCfgTimer.
  • determining the planned motion trajectory according to the first game strategy, the planned driving path and the second deduced trajectory includes: determining the first position of the first object according to the planned driving path and the second deduced trajectory. The time period when the space occupies the second position space of the first intelligent driving device; based on the first game strategy, the planned motion trajectory is determined according to the second position space and the time period, and the planned motion trajectory includes the first intelligent driving device The trajectory of the driving device traveling at the first preset speed in the second location space, or the planned motion trajectory includes the trajectory of the first intelligent driving device traveling at the first preset speed within the time period.
  • the sub-planned motion trajectory may be a trajectory in which the intelligent driving device maintains the creep speed Cf_Velocity during the creep duration CFTime as shown in Figure 16 or Figure 17 in the above embodiment.
  • the method before determining the first game strategy based on the first deduction trajectory and the second deduction trajectory, the method further includes: determining the strategy cost of a pair of deduction trajectories composed of the first deduction trajectory and the second deduction trajectory. Minimum.
  • the policy cost may be the policy cost in the above embodiment.
  • the method before obtaining the predicted motion trajectory and the second motion parameter of the first object, the method further includes: determining that the distance between the first object and the first intelligent driving device is less than or equal to a first distance threshold, and Greater than or equal to the second distance threshold.
  • the first distance threshold may be the preset distance in S501 of the above method 500.
  • the second distance threshold may be determined according to the CIPV screening in the above embodiment, or may be determined by other methods.
  • the method of motion trajectory planning provided by the embodiment of the present application can realize the behavior of the self-vehicle actively maintaining creeping speed to test the game target on the premise of ensuring safety and complying with traffic regulations, making the planned movement trajectory of the self-vehicle less susceptible to influence.
  • the influence of game goals In particular, in narrow lanes where motor vehicles and non-motor vehicles are mixed, or at intersections without indicator lights, the vehicle will no longer experience unexpected frequent mis-braking, mis-braking, continuous conservative yielding, and simultaneous starts and stops. and other issues, which will help improve the efficiency of self-vehicle traffic and improve user driving experience.
  • Figure 19 shows a schematic block diagram of a motion trajectory planning device 1900 provided by an embodiment of the present application.
  • the device 1900 includes an acquisition unit 1910 and a processing unit 1920.
  • the acquisition unit 1910 can implement corresponding communication functions, and the processing unit 1920 is used for data processing.
  • the device 1900 may also include a storage unit, which may be used to store instructions and/or data, and the processing unit 1920 may read the instructions and/or data in the storage unit, so that the device implements the foregoing method embodiments. .
  • the apparatus 1900 may comprise means for performing the method in Figure 5, Figure 8 or Figure 18. Moreover, each unit in the device 1900 and the above-mentioned other operations and/or functions are respectively intended to implement the corresponding processes of the method embodiments in Figure 5, Figure 8 or Figure 18.
  • the acquisition unit 1910 can be used to execute S1810 and S1820 in the method 1800
  • the processing unit 1920 can be used to execute S1830 and S1840 in the method 1800.
  • the device 1900 obtains the unit 1910, which is used to obtain the planned driving path and the first motion parameter of the first intelligent driving device.
  • the first motion parameter includes the speed and/or acceleration of the first intelligent driving device; obtain the first The predicted movement trajectory and the second movement parameter of the object, the second movement parameter including the speed and/or acceleration of the first object;
  • the processing unit 1920 is configured to, when the planned driving path and the predicted movement trajectory intersect, calculate the The first movement parameter and the second movement parameter determine the planned movement trajectory of the first intelligent driving device.
  • the planned movement trajectory includes the movement of the first intelligent driving device to rush or yield to the first object at a first preset speed. Trajectory; control the first intelligent driving device to drive according to the planned motion trajectory.
  • the processing unit 1920 is specifically configured to: determine the first derivation trajectory of the first intelligent driving device according to the first sampled acceleration and the first motion parameter, where the first sampled acceleration is in the sampling space. The determined acceleration that the first intelligent driving device may achieve; determine the second derivation trajectory of the first object based on the second sampled acceleration and the second motion parameter, the second sampled acceleration is the first determined in the sampling space.
  • the possible acceleration of an object when the first derivation trajectory and the second derivation trajectory indicate that the first intelligent driving device and the first object will not collide, or the first object is in the driving direction of the first object
  • the planned motion trajectory is determined based on the first derivation trajectory and the second derivation trajectory, and the first time is a time after the current time.
  • the processing unit 1920 is specifically configured to: determine a first sub-derivation trajectory based on the first sampled acceleration and the speed and/or acceleration of the first intelligent driving device.
  • the speed of the first intelligent driving device at the end point is the first preset speed;
  • the second sub-deduction trajectory is determined based on the first preset speed, and the end point of the first sub-deduction trajectory is the starting point of the second sub-deduction trajectory;
  • the first deduction trajectory includes the first sub-deduction trajectory and the second sub-deduction trajectory.
  • the processing unit 1920 is specifically configured to: determine a first game strategy according to the first derivation trajectory and the second derivation trajectory, and the first game strategy is used to instruct the first intelligent driving device to use the The first preset speed grabs or gives way to the first object; when the first duration is greater than the first time threshold, the planned motion trajectory is determined based on the first game strategy, the planned driving path and the second derivation trajectory.
  • the processing unit 1920 is specifically configured to: determine when the first position space of the first object occupies the second position space of the first intelligent driving device according to the planned driving path and the second derivation trajectory. time period; based on the first game strategy, determine the planned motion trajectory according to the second position space and the time period, the planned motion trajectory includes the first intelligent driving device in the second position space with the first preset The trajectory of speed travel, or the planned movement trajectory includes the trajectory of the first intelligent driving device traveling at the first preset speed within the time period.
  • the processing unit 1920 is specifically configured to determine the minimum strategic cost of the pair of derivation trajectories composed of the first derivation trajectory and the second derivation trajectory.
  • the processing unit 1920 is specifically configured to: determine the first derivation trajectory according to the lateral offset of the first intelligent driving device, the first sampled acceleration and the first motion parameter, the lateral offset is the offset perpendicular to the traveling direction of the first intelligent driving device.
  • the processing unit 1920 is further configured to determine that the distance between the first object and the first intelligent driving device is less than or equal to a first distance threshold and greater than or equal to a second distance threshold.
  • the first preset speed is greater than or equal to 3km/h and less than or equal to 15km/h.
  • each unit in the above device is only a division of logical functions.
  • the units may be fully or partially integrated into a physical entity, or may be physically separated.
  • the unit in the device can be implemented in the form of a processor calling software; for example, the device includes a processor, the processor is connected to a memory, instructions are stored in the memory, and the processor calls the instructions stored in the memory to implement any of the above methods.
  • the processor is, for example, a general-purpose processor, such as a CPU or a microprocessor
  • the memory is a memory within the device or a memory outside the device.
  • the units in the device can be implemented in the form of hardware circuits, and some or all of the functions of the units can be implemented through the design of the hardware circuits, which can be understood as one or more processors; for example, in one implementation,
  • the hardware circuit is an ASIC, which realizes the functions of some or all of the above units through the design of the logical relationship of the components in the circuit; for another example, in another implementation, the hardware circuit can be implemented through PLD, taking FPGA as an example. It can include a large number of logic gate circuits, and the connection relationships between the logic gate circuits can be configured through configuration files to realize the functions of some or all of the above units. All units of the above device may be fully realized by the processor calling software, or may be fully realized by hardware circuits, or part of the units may be realized by the processor calling software, and the remaining part may be realized by hardware circuits.
  • the processor is a circuit with signal processing capabilities.
  • the processor may be a circuit with instruction reading and execution capabilities, such as a CPU, a microprocessor, a GPU, or DSP, etc.; in another implementation, the processor can realize certain functions through the logical relationship of the hardware circuit. The logical relationship of the hardware circuit is fixed or can be reconstructed.
  • the processor is a hardware circuit implemented by ASIC or PLD. For example, FPGA.
  • the process of the processor loading the configuration file and realizing the hardware circuit configuration can be understood as the process of the processor loading instructions to realize the functions of some or all of the above units.
  • it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as NPU, TPU, DPU, etc.
  • each unit in the above device can be one or more processors (or processing circuits) configured to implement the above method, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA , or a combination of at least two of these processor forms.
  • processors or processing circuits
  • each unit in the above device may be integrated together in whole or in part, or may be implemented independently. In one implementation, these units are integrated together and implemented as a system-on-a-chip (SOC).
  • SOC may include at least one processor for implementing any of the above methods or implementing the functions of each unit of the device.
  • the at least one processor may be of different types, such as a CPU and an FPGA, or a CPU and an artificial intelligence processor. CPU and GPU etc.
  • the above-mentioned acquisition unit 1910 may include the game goal screening module shown in FIG. 4
  • the above-mentioned processing unit 1920 may include the interactive game decision-making module and/or motion planning module shown in FIG. 4 .
  • each operation performed by the above-mentioned acquisition unit 1910 and processing unit 1920 may be performed by the same processor, or may be performed by different processors, for example, by multiple processors.
  • one or more processors may be connected to one or more sensors in the sensing system 120 in FIG. 1 , acquire the first motion parameters of the intelligent driving device from the one or more sensors, and perform the first motion The parameters are processed to obtain the deduction trajectory of the intelligent driving equipment.
  • one or more processors can also be connected to the power system of the intelligent driving device to control the intelligent driving device to drive according to the planned movement trajectory.
  • the one or more processors described above may be processors provided in a vehicle machine, or may also be processors provided in other vehicle-mounted terminals.
  • the device 1900 may be a chip provided in a vehicle machine or other vehicle-mounted terminal.
  • the above-mentioned device 1900 may be the computing platform 150 as shown in FIG. 1 provided in an intelligent driving device.
  • FIG. 20 is a schematic block diagram of a motion trajectory planning device according to an embodiment of the present application.
  • the device 2000 for motion trajectory planning shown in FIG. 20 may include: a processor 2010, a transceiver 2020, and a memory 2030.
  • the processor 2010, the transceiver 2020 and the memory 2030 are connected through an internal connection path.
  • the memory 2030 is used to store instructions.
  • the processor 2010 is used to execute the instructions stored in the memory 2030, so that the transceiver 2020 receives/sends some parameters.
  • the memory 2030 can be coupled with the processor 2010 through an interface or integrated with the processor 2010 .
  • transceiver 2020 may include but is not limited to a transceiver device such as an input/output interface to implement communication between the device 2000 and other devices or communication networks.
  • each step of the above method can be completed by instructions in the form of hardware integrated logic circuits or software in the processor 2010 .
  • the methods disclosed in conjunction with the embodiments of the present application can be directly implemented by a hardware processor, or executed by a combination of hardware and software modules in the processor.
  • the software module can be located in random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, registers and other mature storage media in this field.
  • the storage medium is located in the memory 2030.
  • the processor 2010 reads the information in the memory 2030 and completes the steps of the above method in combination with its hardware. To avoid repetition, it will not be described in detail here.
  • the processor 2010 may use a general-purpose CPU, microprocessor, ASIC, GPU or one or more integrated circuits to execute relevant programs to implement the motion trajectory planning method of the method embodiment of the present application.
  • the processor 2010 may also be an integrated circuit chip with signal processing capabilities.
  • each step of the motion trajectory planning method of the present application can be completed by instructions in the form of hardware integrated logic circuits or software in the processor 2010 .
  • the above-mentioned processor 2010 can also be a general-purpose processor, DSP, ASIC, FPGA or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component.
  • Each method, step and logical block diagram disclosed in the embodiment of this application can be implemented or executed.
  • a general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.
  • the steps of the method disclosed in conjunction with the embodiments of the present application can be directly implemented by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor.
  • Software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, registers and other mature storage media in this field.
  • the storage medium is located in the memory 2030.
  • the processor 2010 reads the information in the memory 2030 and performs the motion trajectory planning method of the method embodiment of the present application in conjunction with its hardware.
  • the memory 2030 may be a read-only memory (ROM), a static storage device, a dynamic storage device or a random access memory (RAM).
  • ROM read-only memory
  • RAM random access memory
  • the transceiver 2020 uses a transceiver device such as but not limited to a transceiver to implement communication between the device 2000 and other devices or communication networks.
  • the above-mentioned transceiver 2010 may include the game goal screening module shown in FIG. 4
  • the above-mentioned processor 2020 may include the interactive game decision-making module and/or motion planning module shown in FIG. 4 .
  • An embodiment of the present application also provides an intelligent driving device, which may include the above device 1900 or the above device 2000.
  • Embodiments of the present application also provide a computer-readable medium.
  • the computer-readable medium stores program codes or instructions.
  • the processor implements the above-mentioned FIG. 5 and FIG. 8 or the method in Figure 18.
  • An embodiment of the present application also provides a chip, including: at least one processor and a memory.
  • the at least one processor is coupled to the memory and is used to read and execute instructions in the memory to execute the above-mentioned Figure 5.
  • the size of the sequence numbers of the above-mentioned processes does not mean the order of execution.
  • the execution order of each process should be determined by its functions and internal logic, and should not be used in the embodiments of the present application.
  • the implementation process constitutes any limitation.
  • the disclosed systems, devices and methods can be implemented in other ways.
  • the device embodiments described above are only illustrative.
  • the division of the units is only a logical function division. In actual implementation, there may be other division methods.
  • multiple units or components may be combined or can be integrated into another system, or some features can be ignored, or not implemented.
  • the coupling or direct coupling or communication connection between each other shown or discussed may be through some interfaces, and the indirect coupling or communication connection of the devices or units may be in electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or they may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.
  • each functional unit in each embodiment of the present application can be integrated into one processing unit, each unit can exist physically alone, or two or more units can be integrated into one unit.
  • the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.
  • the technical solution of the present application is essentially or the part that contributes to the existing technology or the part of the technical solution can be embodied in the form of a software product.
  • the computer software product is stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application.
  • the aforementioned storage media include: U disk, mobile hard disk, ROM, RAM, magnetic disk or optical disk and other media that can store program codes.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Atmospheric Sciences (AREA)
  • Automation & Control Theory (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)

Abstract

一种运动轨迹规划的方法、装置以及智能驾驶设备。该方法包括：获取第一智能驾驶设备的规划行驶路径和第一运动参数（S1810）；获取第一物体的预测运动轨迹和第二运动参数（S1820）；在该规划行驶路径和该预测运动轨迹存在交叉时，根据该第一运动参数和该第二运动参数确定该第一智能驾驶设备的规划运动轨迹，该规划运动轨迹包括该第一智能驾驶设备以第一预设速度抢行或让行该第一物体的运动轨迹（S1830）；控制该第一智能驾驶设备按照该规划运动轨迹行驶（S1840）。该方法可以应用于智能车辆、智能机器人等能够自主移动的物体中，能够降低自主移动物体非预期地频繁误轻重刹、误刹停等问题，有助于提高自主移动物体的通行效率。

Description

运动轨迹规划的方法、装置以及智能驾驶设备 技术领域
本申请涉及智能驾驶领域，更具体地，涉及一种运动轨迹规划的方法、装置以及智能驾驶设备。
背景技术
在车辆自动驾驶领域中，车辆的决策模块为障碍物标注标签，为运动规划模块提供合理输入，保证自动驾驶车辆对周围障碍物进行合理响应。传统的基于有限状态机(finite state machine，FSM)的决策规划通常受目标轨迹预测精度的影响，如目标为行人或者非机动车时，由于目标运动的自由度较大，使得运动方向不明确，导致轨迹预测结果跳跃性较大，预测的目标的运动轨迹往往频繁入侵自动驾驶车辆的行驶道路。在一些场景中，例如在机动车、非机动车混行的道路或者无指示灯交叉路口，对目标不合理的轨迹预测会使自动驾驶车辆存在非预期地、频繁地轻重刹、误刹停以及持续保守让行等问题，导致自动驾驶车辆通行效率低，用户驾乘体验差。
发明内容
本申请提供一种运动轨迹规划的方法和装置以及智能驾驶设备，能够降低智能驾驶设备非预期地频繁误轻重刹、误刹停、持续保守让行以及陷入同起同停等问题，有助于提高智能驾驶设备的通行效率，改善用户驾乘体验。
本申请涉及的智能驾驶设备可以包括路上交通工具、水上交通工具、空中交通工具、工业设备、农业设备、或娱乐设备等。例如智能驾驶设备可以为车辆，该车辆为广义概念上的车辆，可以是交通工具(如商用车、乘用车、摩托车、飞行车、火车等)，工业车辆(如：叉车、挂车、牵引车等)，工程车辆(如挖掘机、推土车、吊车等)，农用设备(如割草机、收割机等)，游乐设备，玩具车辆等，本申请实施例对车辆的类型不作具体限定。再如，智能驾驶设备可以为飞机、或轮船等交通工具。
第一方面，提供了一种运动轨迹规划的方法，应用于能够自主移动的第一智能驾驶设备，该方法可以包括：获取第一智能驾驶设备的规划行驶路径和第一运动参数，该第一运动参数包括该第一智能驾驶设备的速度和/或加速度；获取第一物体的预测运动轨迹和第二运动参数，该第二运动参数包括该第一物体的速度和/或加速度；在该规划行驶路径和该预测运动轨迹存在交叉时，根据该第一运动参数和该第二运动参数确定该第一智能驾驶设备的规划运动轨迹，该规划运动轨迹包括该第一智能驾驶设备以第一预设速度抢行或让行该第一物体的运动轨迹；控制该第一智能驾驶设备按照该规划运动轨迹行驶。
在一些可能的实现方式中，第一预设速度可以为使第一智能驾驶设备的驾乘人员感觉安全的较低速度，可以根据驾乘人员的激进程度进行调整。
在一些可能的实现方式中，该第一预设速度大于或等于3公里每小时，且小于或等于 15公里每小时。示例性地，该第一预设速度可以为5km/h，或者也可以为10km/h，或者也可以为其他速度。
需要说明的是，本申请中，第一智能驾驶设备是抢行第一物体还是让行第一物体，根据先通过冲突区域的物体为哪一物体确定。若第一智能驾驶设备在第一物体之前抢先通过冲突区域，则认为第一智能驾驶设备抢行第一物体；若第一智能驾驶设备在第一物体通过冲突区域后再通过，则认为第一智能驾驶设备让行第一物体。其中，冲突区域为预测出的第一智能驾驶设备的规划路径与第一物体的推演轨迹发生重叠的区域。
在上述技术方案中，能够在保证安全和遵守交通法规的前提下，实现第一智能驾驶设备主动保持较低速度来试探第一物体的行为，使得第一智能驾驶设备的规划运动轨迹不易受到第一物体预测轨迹精度的影响。在第一智能驾驶设备让行第一物体时，第一智能驾驶设备可以以第一预设速度行驶，无需刹停让行，有助于提高第一智能驾驶设备行驶的平稳性，在第一智能驾驶设备内承载有乘客时，能够提高乘客的舒适性，改善乘客的驾乘体验；此外，由于第一智能驾驶设备让行第一物体时无需刹停，还能提高第一智能驾驶设备的通行效率。在第一智能驾驶设备以第一预设速度抢行第一物体时，由于第一智能驾驶设备以较低速度行驶，使得第一物体可以根据第一智能驾驶设备的运动轨迹调整自身的运动轨迹，能够避免第一智能驾驶设备持续保持让行第一物体，有助于提高第一智能驾驶设备的通行效率。特别地，在机动车、非机动车混行的窄道或者无指示灯交叉路口等场景下，第一智能驾驶设备不会发生非预期地频繁误轻重刹、误刹停、持续保守让行以及陷入同起同停等问题，有助于提高第一智能驾驶设备行驶的平稳性以及通行效率。
示例性地，第一智能驾驶设备可以为能够自主移动的车辆，或者也可以为其他能够自主移动的物体，例如智能机器人等。
示例性地，第一物体可以为另一智能驾驶设备，包括可以自主移动的智能驾驶设备，和/或受人工控制的智能驾驶设备，该智能驾驶设备可以为机动车，或者也可以为非机动车。该第一物体也可以为人，或者也可以为除智能驾驶设备以外其他能够自主移动的物体。
示例性地，该第一智能驾驶设备的规划行驶路径可以是第一智能驾驶设备的规控模块生成的、期望第一智能驾驶设备在未来一段时长内行驶的路径。应理解，该路径可以只含空间位置信息。
示例性地，该第一物体的预测运动轨迹可以是第一智能驾驶设备的轨迹预测模块根据获取的第一物体的位置，以及速度和/或加速度等预测出的第一物体在未来一段时长内运动的轨迹。在一些可能的实现方式中，该第一物体的预测运动轨迹也可以是由云端或其他设施发送给第一智能驾驶设备的。
应理解，上述“未来一段时长”可以为10秒，或者也可以为20秒，或者也可以为其他时长。
应理解，在该规划行驶路径和该预测运动轨迹存在交叉时，为保证第一智能驾驶设备和第一物体交互时的安全以及第一智能驾驶设备内用户的驾乘体验，需要规划第一智能驾驶设备的行驶轨迹。则可以根据第一智能驾驶设备的第一运动参数，以及第一物体的运动参数确定第一博弈策略，进而确定第一智能驾驶设备的规划运动轨迹。
在一些可能的实现方式中，第一运动参数中仅包含第一智能驾驶设备的速度时，则可以根据第一智能驾驶设备的速度推测出第一智能驾驶设备的加速度；第二运动参数中仅包 含第一物体的速度时，则可以根据第一物体的速度推测出第一物体的加速度。第一运动参数中仅包含第一智能驾驶设备的加速度时，则可以根据第一智能驾驶设备的规划行驶路径中的位置信息，结合加速度推测出第一智能驾驶设备的速度；第二运动参数中仅包含第一物体的加速度时，则可以根据第一物体的预测运动轨迹中的位置信息，结合加速度推测出第一物体的速度。
结合第一方面，在第一方面的某些实现方式中，该根据该第一运动参数和该第二运动参数确定规划运动轨迹，包括：根据第一采样加速度和该第一运动参数确定该第一智能驾驶设备的第一推演轨迹，该第一采样加速度为在采样空间中确定的该第一智能驾驶设备可能达到的加速度；根据第二采样加速度和该第二运动参数确定该第一物体的第二推演轨迹，该第二采样加速度为在所述采样空间中确定的该第一物体可能达到的加速度；在该第一推演轨迹和该第二推演轨迹指示该第一智能驾驶设备和该第一物体不会发生碰撞，或者该第一物体在该第一物体的行驶方向上与该第一智能驾驶设备的侧围或尾部在第一时刻发生碰撞时，根据该第一推演轨迹和该第二推演轨迹确定该规划运动轨迹，该第一时刻为当前时刻之后的时刻。
应理解，在该采样空间中可以获取多组第一采样加速度和第二采样加速度对，进一步地，确定多组第一推演轨迹和第二推演轨迹对。进一步地，可以根据多组第一推演轨迹和第二推演轨迹对中自动驾驶策略代价最小的一对，确定第一智能驾驶设备的规划运动轨迹。
示例性地，在该采样空间中，第一采样加速度和第二采样加速度的最小值可以为-4m/s 2，最大值可以为3m/s 2。上述在该采样空间中获取的多组第一采样加速度和第二采样加速度对中，两组第一采样加速度之间的采样间隔可以为1m/s 2，或者也可以为其他数值；两组第二采样加速度之间的采样间隔可以为1m/s 2，或者也可以为其他数值。应理解，第二推演轨迹是推演的第一物体可能的运动轨迹。
在一些可能的实现方式中，以第一智能驾驶设备和第一物体均为车辆为例，在第一智能驾驶设备和第一物体均处于行驶状态时，“该第一物体在该第一物体的行驶方向上与该第一智能驾驶设备的侧围或尾部发生碰撞”可以理解为，第一物体的车头与第一智能驾驶设备的车身除车头以外的部位发生碰撞；或者，第一物体处于倒车状态，则为第一物体的车尾与第一智能驾驶设备的车身除车头以外的部位发生碰撞。
在一些可能的实现方式中，根据第一推演轨迹和第二推演轨迹可以确定第一智能驾驶设备与第一物体发生碰撞(对第一智能驾驶设备的碰撞部位不作限定)，且碰撞发生时，第一智能驾驶设备处于静止，则也可以根据该第一推演轨迹和该第二推演轨迹确定该规划运动轨迹。
在一些可能的实现方式中，根据第一推演轨迹和第二推演轨迹可以确定第一智能驾驶设备与第一物体发生碰撞，且碰撞方式为该第一智能驾驶设备在该第一智能驾驶设备的行驶方向上与该第一物体的侧围或尾部发生碰撞时，则忽略该第一推演轨迹和该第二推演轨迹，即不再根据该第一推演轨迹和该第二推演轨迹确定该规划运动轨迹。
在上述技术方案中，可以根据第一智能驾驶设备与第一物体是否发生碰撞，以及碰撞类型确定第一智能驾驶设备的博弈策略，有助于提高第一智能驾驶设备行驶安全性。
结合第一方面，在第一方面的某些实现方式中，该第一运动参数包括该第一智能驾驶设备的速度和/或加速度，该根据第一采样加速度和该第一运动参数确定该第一智能驾驶 设备的第一推演轨迹，包括：根据该第一采样加速度，以及该第一智能驾驶设备的速度和/或加速度确定第一子推演轨迹，该第一子推演轨迹的终点处该第一智能驾驶设备的速度为第一预设速度；根据该第一预设速度确定第二子推演轨迹，该第一子推演轨迹的终点为该第二子推演轨迹的起点；根据该第一子推演轨迹和该第二子推演轨迹，确定该第一推演轨迹。
需要说明的是，上述“第一子推演轨迹的终点”并非第一智能驾驶设备停止运动的点，该“第一子推演轨迹的终点”仅作为区分第一推演轨迹中特征不同的两段子推演轨迹的一个特征，该“第一子推演轨迹的终点”也可以被称为“第二子推演轨迹的起点”。
在一些可能的实现方式中，第一采样加速度为负值，且第一智能驾驶设备的当前加速度与第一采样加速度不相等时，则第一子推演轨迹可以包括第一智能驾驶设备的加速度从当前加速度变化为第一采样加速度的轨迹。在第一智能驾驶设备的加速度变为第一采样加速度时，第一智能驾驶设备的速度还未减速至第一预设速度，则第一子推演轨迹还可以包括第一智能驾驶设备以第一采样加速度行驶，以使第一智能驾驶设备的速度减小至第一预设速度的轨迹。
在一些可能的实现方式中，第一采样加速度为正值，且第一智能驾驶设备的当前加速度与第一采样加速度不相等时，则第一子推演轨迹可以包括第一智能驾驶设备的加速度从当前加速度变化为第一采样加速度的轨迹。在第一智能驾驶设备的加速度变为第一采样加速度时，第一智能驾驶设备的速度还未加速至第一预设速度，则第一子推演轨迹还可以包括第一智能驾驶设备以第一采样加速度行驶，以使第一智能驾驶设备的速度增加至第一预设速度的轨迹。
在一些可能的实现方式中，第一采样加速度与第一智能驾驶设备当前加速度相等，且第一智能驾驶设备的当前速度与第一预设速度不相等，则第一子推演轨迹可以包括第一智能驾驶设备以当前加速度(即第一采样加速度)行驶，以使第一预设速度由当前速度变化为第一预设速度的轨迹。
在一些可能的实现方式中，第一采样加速度与第一智能驾驶设备当前加速度不相等，且第一智能驾驶设备的当前速度与第一预设速度相等，则第一子推演轨迹可以为空。即直接根据当前速度(即第一预设速度)确定第二子推演轨迹。
在一些可能的实现方式中，还可以根据该第一采样加速度，以及该第一智能驾驶设备的速度和/或加速度确定第一子推演轨迹，该第一子推演轨迹的终点处该第一智能驾驶设备的速度为第二速度，该第二速度为该第一智能驾驶设备所能行驶的最高速度或该第一智能驾驶设备行驶路段的最高限速；根据该该第二速度确定第二子推演轨迹；该第一推演轨迹包括该第一子推演轨迹和该第二子推演轨迹。
示例性地，第一采样加速度为正值，且第一智能驾驶设备的当前加速度与第一采样加速度不相等时，则第一子推演轨迹可以包括第一智能驾驶设备的加速度从当前加速度变化为第一采样加速度的轨迹。在第一智能驾驶设备的加速度变为第一采样加速度时，第一智能驾驶设备的速度还未加速至第二速度，则第一子推演轨迹还可以包括第一智能驾驶设备以第一采样加速度行驶，以使第一智能驾驶设备的速度增加至第二速度的轨迹。
在一些可能的实现方式中，也可以根据第二采样加速度和该第一运动参数确定该第一智能驾驶设备的第一推演轨迹。具体可以包括：根据该第二采样加速度，以及该第一智能 驾驶设备的速度和/或加速度确定第三子推演轨迹，该第三子推演轨迹的终点处该第一智能驾驶设备的速度为第三速度或零，该第三速度为该第一智能驾驶设备所能行驶的最高速度或该第一智能驾驶设备行驶路段的最高限速；在该第三子推演轨迹的终点处该第一智能驾驶设备的速度为该第三速度时，根据该第三速度确定第四子推演轨迹；该第一推演轨迹包括该第三子推演轨迹；或者，该第一推演轨迹包括该第三子推演轨迹和第四子推演轨迹，该第三子推演轨迹的终点为该第四子推演轨迹的起点。
应理解，上述技术方案中涉及的“子推演轨迹”可以理解为在沿第一智能驾驶设备前进方向上第一智能驾驶设备的位置随时间变化的点组成的集合，每个点中不包含垂直于第一智能驾驶设备前进方向上的位置。示例性地，若第一智能驾驶设备为智能驾驶设备，则“第一智能驾驶设备前进方向”可以为在平行于地面的平面中，平行于智能驾驶设备纵向对称平面的方向。
在上述技术方案中，在进行沿第一智能驾驶设备前进方向上的第一智能驾驶设备的轨迹规划时，使第一智能驾驶设备的速度最低降至第一预设速度，可以避免第一智能驾驶设备的速度减至零，能够有效避免第一智能驾驶设备频繁刹停，有助于提高第一智能驾驶设备的通行效率；在第一智能驾驶设备内承载有乘客时，还能够改善乘客的驾乘体验。
结合第一方面，在第一方面的某些实现方式中，该根据第三采样加速度和该第二运动参数确定该第一物体的第二推演轨迹，包括：根据该第三采样加速度，以及该第一物体的速度和/或加速度确定第五子推演轨迹，该第五子推演轨迹的终点处该第一物体的速度为第四速度或零，该第四速度为该第一物体所能行驶的最高速度或该第一物体行驶路段的最高限速；在该第五子推演轨迹的终点处该第一智能驾驶设备的速度为该第四速度时，根据该第四速度确定第六子推演轨迹；该第二推演轨迹包括该第五子推演轨迹；或者，该第二推演轨迹包括该第五子推演轨迹和第六子推演轨迹。
结合第一方面，在第一方面的某些实现方式中，该根据该第一推演轨迹和该第二推演轨迹确定该规划运动轨迹，包括：根据该第一推演轨迹和该第二推演轨迹确定第一博弈策略，该第一博弈策略用于指示该第一智能驾驶设备以该第一预设速度抢行或让行该第一物体；在第一时长大于第一时间阈值时，根据该第一博弈策略、该规划行驶路径和该第二推演轨迹确定该规划运动轨迹。
示例性地，第一时间阈值可以为0.1秒，或者也可以为0.3秒，或者第一时间阈值还可以取其他数值。
应理解，在第一智能驾驶设备和第一物体运动的过程中，可以根据第一智能驾驶设备的第一运动参数和第一物体的第二运动参数实时推演多组推演轨迹对，根据不同的推演轨迹对可以确定不同的博弈策略。在本申请中，博弈策略可以包括：第一智能驾驶设备以第一预设速度抢行第一物体；第一智能驾驶设备以第一预设速度让行第一物体；第一智能驾驶设备以非蠕行的方式抢行或让行第一物体。
在一些可能实现方式中，第一智能驾驶设备的控制装置每隔固定时长确定一次博弈策略。示例性地，该固定时长可以为20毫秒，或者也可以为50毫秒。
在根据推演轨迹对确定博弈策略时，当前帧(或当前周期)确定的博弈策略与上一帧(或上一周期)确定的博弈策略可能不同，为了防止博弈策略频繁跳变导致的规划运动轨迹的不稳定，本申请提出在确定第一博弈策略持续的时长大于第一时间阈值时，再根据该 第一博弈策略确定规划运动轨迹。假设第一博弈策略为第一智能驾驶设备以第一预设速度抢行第一物体、第二博弈策略为第一智能驾驶设备以非蠕行的方式抢行或让行第一物体、第三博弈策略决策为第一智能驾驶设备以第一预设速度让行第一物体为，以固定时长为20毫秒、第一时间阈值为0.1秒为例，若确定博弈策略为第一博弈策略之前，确定的博弈策略为第二博弈策略，则需要连续5次确定博弈策略为第一博弈策略，才会进一步地根据该第一博弈策略、该规划行驶路径和该第二推演轨迹确定该规划运动轨迹。
在一些可能的实现方式中，可以根据确定博弈策略为第一博弈策略之前的博弈策略确定第一时间阈值的取值。在一些可能的实现方式中，若确定第一博弈策略之前确定的博弈策略为第二博弈策略，则第一时间阈值为第一阈值；若确定第一博弈策略之前确定的博弈策略为第三博弈策略，则第一时间阈值为第二阈值。其中，第一阈值小于第二阈值。示例性地，第一阈值可以为0.1秒，第二阈值可以为0.3秒。
在一些可能的实现方式中，若确定第一博弈策略之后确定的博弈策略为第二博弈策略，且第二博弈策略持续的时长大于第三阈值，则根据该第二博弈策略确定该规划运动轨迹。其中，该第三阈值大于第二阈值。示例性地，则第一时间阈值为第二阈值。其中，第一阈值小于第二阈值。示例性地，该第三阈值可以为0.5秒。
在上述技术方案中，博弈策略发生变化时，可以根据变化后博弈策略持续的时长确定是否根据该博弈策略重新进行轨迹规划，在变化后博弈策略持续的时长不满足条件时，不重新规划运动轨迹，能够避免博弈策略频繁跳变，有助于提高规划运动轨迹的稳定性。
结合第一方面，在第一方面的某些实现方式中，该根据该第一博弈策略、该规划行驶路径和该第二推演轨迹确定该规划运动轨迹，包括：根据该规划行驶路径和该第二推演轨迹确定该第一物体的第一位置空间占据该第一智能驾驶设备的第二位置空间时的时间段；基于该第一博弈策略，根据该第二位置空间和该时间段确定该规划运动轨迹，该规划运动轨迹包括该第一智能驾驶设备在该第二位置空间以该第一预设速度行驶的轨迹，或者，该规划运动轨迹包括该第一智能驾驶设备在该时间段内以该第一预设速度行驶的轨迹。
在一些可能的实现方式中，基于该第一博弈策略，根据该第二位置空间和该时间段确定该规划运动轨迹，包括：根据该第一博弈策略、该第二位置空间和距离阈值safedis确定第三位置空间，根据该时间段和时间阈值TimeGap确定第一时间段；根据该第一时间段、该第一预设速度和该第三位置空间确定该规划运动轨迹。
结合第一方面，在第一方面的某些实现方式中，该根据该第一推演轨迹和该第二推演轨迹确定该第一博弈策略之前，该方法还包括：确定该第一推演轨迹和该第二推演轨迹组成的推演轨迹对的策略代价最小。
示例性地，策略代价可以包括安全性策略代价，还可以包括舒适性策略代价、通过性策略代价和路权策略代价中的至少一个。具体地，安全性策略代价用于表征移动物体行驶安全性，安全性越低，则策略代价越高；舒适性策略代价用于表征移动物体内用户的舒适性，一般该移动物体的加速度变化率越大，则舒适性越差，则策略代价越高；通过性策略代价用于表征移动物体通过冲突点的通过性，示例性地，第一智能驾驶设备的第一推演轨迹通过冲突点的时间与标定时间之间的差值越大，则通过性越差，则策略代价越高，其中，标定时间为以第一智能驾驶设备的横向偏移为0，且采样加速度为0时，推演的第一智能驾驶设备的轨迹通过冲突点的时间；路权策略代价用于表征是否使路权较高的物体的运动 状态发生改变，若第一推演轨迹和第二推演轨迹对使得路权较高的物体的运动状态发生改变，则策略代价较高。
应理解，上述“移动物体”可以为第一智能驾驶设备，也可以为第一物体。
在一些可能的实现方式中，策略代价可以只包括安全性策略代价；或者，策略代价除了安全性策略代价以外，还可以包括第一智能驾驶设备和第一物体的舒适性策略代价、通过性策略代价和路权策略代价，其中各策略代价所占权重可以不同。在一些可能的实现方式中，各策略代价所占权重还可以随行驶场景变化。
在一些可能的实现方式中，策略代价的最小值可以为0。
以第一物体为智能驾驶设备为例，在根据该第一推演轨迹和该第二推演轨迹，确定该第一物体在该第一物体的行驶方向上与该第一智能驾驶设备在偏离该第一智能驾驶设备行驶方向的方向上发生碰撞时，可以结合碰撞位置处第一物体的速度确定全性策略代价。应理解，若策略代价仅包括安全性策略代价，则在推演出多组第一推演轨迹和第二推演轨迹对时，取安全性策略代价最低的第一推演轨迹和第二推演轨迹对确定第一博弈策略。
在上述技术方案中，可以评估第一推演轨迹和第二推演轨迹对的策略代价，并根据策略代价确定使用哪个第一推演轨迹/第二推演轨迹对进行第一智能驾驶设备的轨迹规划，有助于提高第一智能驾驶设备的规划运动轨迹的合理性。还可以进一步地，结合行驶场景和多个策略代价评价维度对规划运动轨迹进行更为准确的评价。
结合第一方面，在第一方面的某些实现方式中，该根据第一采样加速度和该第一运动参数确定该第一智能驾驶设备的第一推演轨迹，包括：根据该第一智能驾驶设备的横向偏移、该第一采样加速度和该第一运动参数确定该第一推演轨迹，该横向偏移为垂直于该第一智能驾驶设备行驶方向上的偏移。
在一些可能的实现方式中，根据该第一采样加速度和该第一运动参数确定第一智能驾驶设备的子推演轨迹(或者也可以被称为纵向轨迹)，即该第一智能驾驶设备在平行于该第一智能驾驶设备行驶方向上的轨迹。
进一步地，根据该子推演轨迹，结合第一智能驾驶设备的横向偏移可以确定出第一智能驾驶设备的第一推演轨迹，该第一推演轨迹为第一智能驾驶设备的坐标随时间变化的点组成的集合，即第一智能驾驶设备的纵向轨迹和第一智能驾驶设备的横向偏移融合后的结果。
应理解，对于第一物体，也可以根据第一物体的横向偏移、该第二采样加速度和该第二运动参数确定该第二推演轨迹。
结合第一方面，在第一方面的某些实现方式中，该获取第一物体的预测运动轨迹和第二运动参数之前，该方法还包括：确定该第一物体与该第一智能驾驶设备之间的距离小于或等于第一距离阈值，且大于或等于第二距离阈值。
示例性地，该第一距离阈值可以为100米，或者也可以为200米，或者也可以为其他数值。
示例性地，第二距离阈值可以是根据最近路径车辆(closest in-path vehicle，CIPV)筛选确定的。应理解，若通过CIPV筛选确定第二距离阈值，则第一物***于第一智能驾驶设备周围的不同方位时，该第二距离阈值可能不同。在一些可能的实现方式中，该第二距离阈值也可以为其他时间距离和/或空间距离，例如，若为时间距离，则可以为2秒，或 者3秒，或者也可以为其他数值；若为空间距离，则可以为1.5米，或者2米，或者也可以为其他数值。
在一些可能的实现方式中，与该第一智能驾驶设备之间的距离小于或等于第一距离阈值的第一物体有两个或两个以上，则需要确定该两个或两个以上的第一物体与第一智能驾驶设备之间的距离均大于或等于第二距离阈值。也就是说，当第一智能驾驶设备周围有任意一个第一物体与第一智能驾驶设备之间的距离小于第二距离阈值，则不进行第一智能驾驶设备的自动驾驶决策以及轨迹规划。
在上述技术方案中，通过限定第一物体与第一智能驾驶设备之间的最大距离，有助于剔除距离过远的第一物体，能够减少轨迹规划过程中的计算量。通过限定第一物体与第一智能驾驶设备之间的最小距离，能够防止第一物体与第一智能驾驶设备距离过近，导致没有足够空间进行减速行驶的情况发生，有助于提高自动驾驶过程的安全性。
第二方面，提供了一种运动轨迹规划的方法，该方法可以包括：获取第一智能驾驶设备的规划行驶路径和第一运动参数；获取第一物体的预测运动轨迹和第二运动参数；在该规划行驶路径和该预测运动轨迹存在交叉时，根据该第一运动参数和该第二运动参数确定该第一智能驾驶设备的规划运动轨迹，该规划运动轨迹用于指示该第一智能驾驶设备以第一预设速度通过冲突区域，或者用于指示该第一智能驾驶设备在该第一物体通过冲突区域时以该第一预设速度行驶，其中，该冲突区域为该第一智能驾驶设备的规划行驶路径与该第一物体的推演轨迹发生重叠的区域；控制该第一智能驾驶设备按照该规划运动轨迹行驶。
第三方面，提供了一种运动轨迹规划的方法，该方法包括：获取第一智能驾驶设备的规划行驶路径和第一运动参数；获取第一物体的预测运动轨迹和第二运动参数；在该规划行驶路径和该预测运动轨迹存在交叉时，根据该第一运动参数和该第二运动参数确定第一博弈策略，该第一博弈策略用于指示该第一智能驾驶设备以第一预设速度抢行或让行该第一物体；根据该第一博弈策略确定该第一智能驾驶设备的规划运动轨迹。
需要说明的是，第一博弈策略决策第一智能驾驶设备以第一预设速度抢行或让行第一物体，是期望第一智能驾驶设备以第一预设速度通过冲突区域，或者第一智能驾驶设备在第一物体通过冲突区域时以第一预设速度行驶。但是在具体实现过程中，第一智能驾驶设备通过冲突区域时的速度不一定达到第一预设速度，或者也可能大于第一预设速度。以下结合两大类情况，具体说明第一智能驾驶设备在实际行驶过程中可能发生的场景。
第一智能驾驶设备在第一物体之前抢先通过冲突区域的情况，可以进一步细化为：
1、第一智能驾驶设备在从低于第一预设速度的行驶速度向第一预设速度加速的过程中，通过冲突区域。进一步地，由于第一智能驾驶设备在通过冲突区域后，第一智能驾驶设备与第一物体的交互结束，因此，在第一智能驾驶设备通过冲突区域后，可以决策规划第一智能驾驶设备以高于第一预设速度的行驶速度继续行驶。
2、第一智能驾驶设备在到达冲突区域之前，速度已经达到第一预设速度(可能是从高于第一预设速度的行驶速度减速至第一预设速度，也可能是从低于第一预设速度的行驶速度加速至第一预设速度)，则第一智能驾驶设备以第一预设速度通过冲突区域。进一步地，由于第一智能驾驶设备在通过冲突区域后，第一智能驾驶设备与第一物体的交互结束，因此，在第一智能驾驶设备通过冲突区域后，可以决策规划第一智能驾驶设备以高于第一预设速度的行驶速度继续行驶。
3、第一智能驾驶设备在从高于第一预设速度的行驶速度向第一预设速度减速的过程中，通过冲突区域。进一步地，由于第一智能驾驶设备在通过冲突区域后，第一智能驾驶设备与第一物体的交互结束，因此，在第一智能驾驶设备通过冲突区域后，可以决策规划第一智能驾驶设备结束减速状态，控制其以高于第一预设速度的行驶速度继续行驶。
可以理解的是，在上述前两种情况中，该第一智能驾驶设备以小于或等于第一预设速度的行驶速度抢行第一物体；在第三种情况中，第一智能驾驶设备以高于第一预设速度的行驶速度抢行第一物体。
第一智能驾驶设备在第一物体通过冲突区域后再通过的情况，可以进一步细化为：
1、第一智能驾驶设备在向冲突区域行驶的过程中，处于从低于第一预设速度的行驶速度向第一预设速度加速的状态，此时第一物体通过冲突区域。进一步地，由于第一物体在通过冲突区域后，第一智能驾驶设备与第一物体的交互结束，因此，在确定第一物体通过冲突区域后，可以决策规划第一智能驾驶设备以高于第一预设速度的行驶速度继续行驶。
2、第一智能驾驶设备以第一预设速度在向冲突区域行驶的过程中，则第一物体通过冲突区域。进一步地，由于第一物体在通过冲突区域后，第一智能驾驶设备与第一物体的交互结束，因此，在确定第一物体通过冲突区域后，可以决策规划第一智能驾驶设备以高于第一预设速度的行驶速度继续行驶。
3、第一智能驾驶设备在向冲突区域行驶的过程中，处于从高于第一预设速度的行驶速度向第一预设速度减速的状态，此时第一物体通过冲突区域。进一步地，由于第一物体在通过冲突区域后，第一智能驾驶设备与第一物体的交互结束，因此，在确定第一物体通过冲突区域后，可以决策规划第一智能驾驶设备以高于第一预设速度的行驶速度继续行驶。
可以理解的是，在上述前两种情况中，第一智能驾驶设备以小于或等于第一预设速度的行驶速度让行第一物体；在第三种情况中，第一智能驾驶设备以高于第一预设速度的行驶速度让行第一物体。
也就是说，无论第一博弈策略决策第一智能驾驶设备以第一预设速度抢行第一物体，还是让行第一物体，第一智能驾驶设备在通过冲突区域时，其速度可能是第一预设速度，或者，也可能高于或低于第一预设速度。
需要说明的是，上述“第一智能驾驶设备与第一物体的交互结束”之后，该第一物体的运动行为将不再影响第一智能驾驶设备的运动轨迹的决策与规划。应理解，在第一智能驾驶设备与该第一物体交互结束之后，可能出现其他与第一智能驾驶设备距离过近的物体，则第一智能驾驶设备在通过冲突区域后，也可能及时刹停，而非以高于第一预设速度的行驶速度继续行驶。
第四方面，提供了一种运动轨迹规划的装置，该装置可以包括：获取单元，用于获取该第一智能驾驶设备的规划行驶路径和第一运动参数；获取第一物体的预测运动轨迹和第二运动参数；处理单元，用于在该规划行驶路径和该预测运动轨迹存在交叉时，根据该第一运动参数和该第二运动参数确定该第一智能驾驶设备的规划运动轨迹，该规划运动轨迹包括该第一智能驾驶设备以第一预设速度抢行或让行该第一物体的运动轨迹；控制该第一智能驾驶设备按照该规划运动轨迹行驶。
结合第四方面，在第四方面的某些实现方式中，该处理单元具体用于：根据第一采样加速度和该第一运动参数确定该第一智能驾驶设备的第一推演轨迹，该第一采样加速度为 在采样空间中确定的该第一智能驾驶设备可能达到的加速度；根据第二采样加速度和该第二运动参数确定该第一物体的第二推演轨迹，该第二采样加速度为在采样空间中确定的该第一物体可能达到的加速度；在该第一推演轨迹和该第二推演轨迹指示该第一智能驾驶设备和该第一物体不会发生碰撞，或者该第一物体在该第一物体的行驶方向上与该第一智能驾驶设备在偏离该第一智能驾驶设备行驶方向的方向上在第一时刻发生碰撞时，根据该第一推演轨迹和该第二推演轨迹确定该规划运动轨迹，该第一时刻为当前时刻之后的时刻。
结合第四方面，在第四方面的某些实现方式中，该第一运动参数包括该第一智能驾驶设备的速度和/或加速度，该处理单元具体用于：根据该第一采样加速度，以及该第一智能驾驶设备的速度和/或加速度确定第一子推演轨迹，该第一子推演轨迹的终点处该第一智能驾驶设备的速度为第一预设速度；根据该第一预设速度确定第二子推演轨迹；该第一推演轨迹包括该第一子推演轨迹和该第二子推演轨迹。
结合第四方面，在第四方面的某些实现方式中，该处理单元具体用于：根据该第一推演轨迹和该第二推演轨迹确定第一博弈策略，该第一博弈策略用于指示该第一智能驾驶设备以该第一预设速度抢行或让行该第一物体；在第一时长大于第一时间阈值时，根据该第一博弈策略、该规划行驶路径和该第二推演轨迹确定该规划运动轨迹。
结合第四方面，在第四方面的某些实现方式中，该处理单元具体用于：根据该规划行驶路径和该第二推演轨迹确定该第一物体的第一位置空间占据该第一智能驾驶设备的第二位置空间时的时间段；基于该第一博弈策略，根据该第二位置空间和该时间段确定该规划运动轨迹，该规划运动轨迹包括该第一智能驾驶设备在该第二位置空间以该第一预设速度行驶的轨迹，或者，该规划运动轨迹包括该第一智能驾驶设备在该时间段内以该第一预设速度行驶的轨迹。
结合第四方面，在第四方面的某些实现方式中，该处理单元具体用于：确定该第一推演轨迹和该第二推演轨迹组成的推演轨迹对的策略代价最小。
结合第四方面，在第四方面的某些实现方式中，该处理单元具体用于：根据该第一智能驾驶设备的横向偏移、该第一采样加速度和该第一运动参数确定该第一推演轨迹，该横向偏移为垂直于该第一智能驾驶设备行驶方向上的偏移。
结合第四方面，在第四方面的某些实现方式中，该处理单元还用于：确定该第一物体与该第一智能驾驶设备之间的距离小于或等于第一距离阈值，且大于或等于第二距离阈值。
第五方面，提供了一种运动轨迹规划的装置，该装置可以包括：获取单元，用于获取第一智能驾驶设备的规划行驶路径和第一运动参数；获取第一物体的预测运动轨迹和第二运动参数；处理单元，用于在该规划行驶路径和该预测运动轨迹存在交叉时，根据该第一运动参数和该第二运动参数确定该第一智能驾驶设备的规划运动轨迹，该规划运动轨迹用于指示该第一智能驾驶设备以第一预设速度通过冲突区域，或者用于指示该第一智能驾驶设备在该第一物体通过冲突区域时以该第一预设速度行驶，其中，该冲突区域为该第一智能驾驶设备的规划行驶路径与该第一物体的推演轨迹发生重叠的区域；控制该第一智能驾驶设备按照该规划运动轨迹行驶。
第六方面，提供了一种运动轨迹规划的装置，该装置包括：获取模块，用于获取第一智能驾驶设备的规划行驶路径和第一运动参数；获取第一物体的预测运动轨迹和第二运动参数；处理模块，用于在该规划行驶路径和该预测运动轨迹存在交叉时，根据该第一运动 参数和该第二运动参数确定第一博弈策略，该第一博弈策略用于指示该第一智能驾驶设备以第一预设速度抢行或让行该第一物体；根据该第一博弈策略确定该第一智能驾驶设备的规划运动轨迹。
第七方面，提供了一种运动轨迹规划的装置，该装置包括：存储器，用于存储程序；处理器，用于执行存储器存储的程序，当存储器存储的程序被执行时，处理器用于执行上述第一方面至第三方面中任一种可能实现方式中的方法。
第八方面，提供了一种智能驾驶设备，该智能驾驶设备包括上述第四方面至上述第七方面中任一种实现方式中的装置。
结合第八方面，在第八方面的某些实现方式中，该智能驾驶设备为车辆。
第九方面，提供了一种计算机程序产品，上述计算机程序产品包括：计算机程序代码，当上述计算机程序代码在计算机上运行时，使得计算机执行上述第一方面至第三方面中任一种可能实现方式中的方法。
需要说明的是，上述计算机程序代码可以全部或部分存储在第一存储介质上，其中第一存储介质可以与处理器封装在一起的，也可以与处理器单独封装，本申请实施例对此不作具体限定。
第十方面，提供了一种计算机可读介质，上述计算机可读介质存储由程序代码，当上述计算机程序代码在计算机上运行时，使得计算机执行上述第一方面至第三方面中任一种可能实现方式中的方法。
第十一方面，提供了一种芯片，该芯片包括处理器，用于调用存储器中存储的计算机程序或计算机指令，以使得该处理器执行上述第一方面至第三方面中任一种可能实现方式中的方法。
结合第十一方面，在一种可能的实现方式中，该处理器通过接口与存储器耦合。
结合第十一方面，在一种可能的实现方式中，该芯片***还包括存储器，该存储器中存储有计算机程序或计算机指令。
附图说明
图1是本申请实施例提供的智能驾驶设备的功能性框图示意。
图2是本申请实施例提供的各种传感器感测范围示意图。
图3是本申请实施例提供的一种运动轨迹规划的方法实施所需的***架构示意图。
图4是本申请实施例提供的一种运动轨迹规划的方法实施所需的***架构示意图。
图5是本申请实施例提供的一种运动轨迹规划的方法的示意性流程图。
图6是本申请实施例提供的一种运动轨迹规划的方法的应用场景的示意图。
图7是本申请实施例提供的一种CIPV的示意图。
图8是本申请实施例提供的一种运动轨迹规划的方法的示意性流程图。
图9是本申请实施例提供的一种采样空间的示意图。
图10是本申请实施例提供的一种横向路径推演示意图。
图11是本申请实施例提供的一种纵向轨迹推演示意图。
图12是本申请实施例提供的一种车辆侧面的示意图。
图13是本申请实施例提供的博弈策略应用场景的示意图。
图14是本申请实施例提供的博弈策略稳定处理的示意性流程图。
图15是本申请实施例提供的一种自车与博弈目标运动轨迹的示意图。
图16是本申请实施例提供的一种蠕行让行目标约束的示意图。
图17是本申请实施例提供的一种蠕行抢行目标约束的示意图。
图18是本申请实施例提供的一种运动轨迹规划的方法的示意性流程图。
图19是本申请实施例提供的一种运动轨迹规划的装置的示意性框图。
图20是本申请实施例提供的一种运动轨迹规划的装置的示意性框图。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行描述。其中，在本申请实施例的描述中，除非另有说明，“/”表示或的意思，例如，A/B可以表示A或B；本文中的“和/或”仅仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。
本申请实施例中采用诸如“第一”、“第二”的前缀词，仅仅为了区分不同的描述对象，对被描述对象的位置、顺序、优先级、数量或内容等没有限定作用。本申请实施例中对序数词等用于区分描述对象的前缀词的使用不对所描述对象构成限制，对所描述对象的陈述参见权利要求或实施例中上下文的描述，不应因为使用这种前缀词而构成多余的限制。此外，在本实施例的描述中，除非另有说明，“至少一个”是指一个或者多个，“多个”的含义是两个或两个以上。字符“/”一般表示前后关联对象是一种“或”的关系。“以下至少一项(个)”或其类似表达，是指的这些项中的任意组合，包括单项(个)或复数项(个)的任意组合。例如，a，b，或c中的至少一项(个)，可以表示：a，b，c，a-b，a-c，b-c，或a-b-c，其中a，b，c可以是单个，也可以是多个。
图1是本申请实施例提供的智能驾驶设备100的一个功能框图示意。智能驾驶设备100可以包括感知***120、显示装置130和计算平台150，其中，感知***120可以包括感测关于智能驾驶设备100周边的环境的信息的若干种传感器。例如，感知***120可以包括定位***，定位***可以是全球定位***(global positioning system，GPS)，也可以是北斗***或者其他定位***、惯性测量单元(inertial measurement unit，IMU)、激光雷达、毫米波雷达、超声雷达以及摄像装置中的一种或者多种。
智能驾驶设备100的部分或所有功能可以由计算平台150控制。计算平台150可包括处理器151至15n(n为正整数)，处理器是一种具有信号的处理能力的电路，在一种实现中，处理器可以是具有指令读取与运行能力的电路，例如中央处理单元(central processing unit，CPU)、微处理器、图形处理器(graphics processing unit，GPU)(可以理解为一种微处理器)、或数字信号处理器(digital signal processor，DSP)等；在另一种实现中，处理器可以通过硬件电路的逻辑关系实现一定功能，该硬件电路的逻辑关系是固定的或可以重构的，例如处理器为专用集成电路(application-specific integrated circuit，ASIC)或可编程逻辑器件(programmable logic device，PLD)实现的硬件电路，例如现场可编辑逻辑门阵列(filed programmable gate array，FPGA)。在可重构的硬件电路中，处理器加载配置文档，实现硬件电路配置的过程，可以理解为处理器加载指令，以实现以上部分或全部单元的功能的过程。此外，还可以是针对人工智能设计的硬件电路，其可以理解为一 种ASIC，例如神经网络处理单元(neural network processing unit，NPU)、张量处理单元(tensor processing unit，TPU)、深度学***台150还可以包括存储器，存储器用于存储指令，处理器151至15n中的部分或全部处理器可以调用存储器中的指令，执行指令，以实现相应的功能。
智能驾驶设备100可以包括高级驾驶辅助***(advanced driving assistant system，ADAS)，ADAS利用感知***120中的多种传感器(包括但不限于：激光雷达、毫米波雷达、摄像装置、超声波传感器、全球定位***、惯性测量单元)从智能驾驶设备周围获取信息，并对获取的信息进行分析和处理，实现例如障碍物感知、目标识别、智能驾驶设备定位、路径规划、驾驶员监控/提醒等功能，从而提升智能驾驶设备驾驶的安全性、自动化程度和舒适度。
图2示出各种传感器感测范围示意图，传感器可以包括例如图1所示的感知***120中的激光雷达、毫米波雷达、摄像装置、超声波传感器，其中毫米波雷达可以分为长距雷达和中/短距雷达。目前，激光雷达的感测范围约在80-150米，长距毫米波雷达的感测范围约为1-250米，中/短距毫米波雷达的感测范围约在30-120米，摄像头的感测范围约在50-200米，超声波雷达的感测范围约在0-5米。
在不同的自动驾驶等级(L0-L5)下，基于人工智能算法和多传感器所获取的信息，ADAS可以实现不同等级的自动驾驶辅助，上述的自动驾驶等级(L0-L5)是基于汽车工程师协会(society of automotive engineers，SAE)的分级标准的。其中，L0级为无自动化；L1级为驾驶支援；L2级为部分自动化；L3级为有条件自动化；L4级为高度自动化；L5级为完全自动化。L1至L3级监测路况并做出反应的任务都由驾驶员和***共同完成，并需要驾驶员接管动态驾驶任务。L4和L5级可以让驾驶员完全转变为乘客的角色。目前，ADAS可以实现的功能主要包括但不限于：自适应巡航、自动紧急刹车、自动泊车、盲点监测、前方十字路***通警示/制动、后方十字路***通警示/制动、前车碰撞预警、车道偏离预警、车道保持辅助、后车防撞预警、交通标识识别、交通拥堵辅助、高速公路辅助等。应当理解的是：上述的各种功能在不同的自动驾驶等级(L0-L5)下可以有具体的模式，自动驾驶等级越高，对应的模式越智能。
如上所述，当前技术背景下，规划自动驾驶轨迹的方法通常是基于FSM和基于博弈的物体决策规划，上述决策规划会受博弈目标的轨迹预测精度的影响较大，如博弈目标为行人或者非机动车时，由于运动的自由度较大，使得运动方向更加不明确，导致轨迹预测结果跳跃性较大，使得预测的博弈目标的运动轨迹频繁入侵自动驾驶车辆的行驶道路。对博弈目标不合理的轨迹预测会导致自动驾驶车辆存在非预期地、频繁轻重刹、误刹停、以及持续保守让行等问题，或者使得博弈目标和自动驾驶车辆同时停止和/或同时起步，典型场景为机动车、非机动车混行的道路或者无交通指示灯路口，导致自动驾驶车辆通行效率低，用户驾乘体验差。鉴于此，本申请实施例提供一种运动轨迹规划的方法和装置以及智能驾驶设备，能够基于博弈策略的语义级决策标签以及博弈目标的推演轨迹进行车辆的轨迹规划，在保证安全和遵守交通法规的前提下，实现车辆在自动驾驶过程中主动保持低速蠕行来试探博弈目标的行为，车辆在保持低速蠕行状态时，是一种低速状态且不易受博弈目标轨迹预测精度的影响，从而能够避免频繁误刹、轻刹、重刹、以及误刹停等问题，有助于改善用户的驾乘体验。
需要说明的是，本申请中涉及的非机动车是指，以人力或者畜力驱动，上道路行驶的交通工具，以及虽有动力装置驱动但设计最高时速、空车质量、外形尺寸符合有关国家标准的残疾人机动轮椅车、电动自行车等交通工具；机动车是指由自带动力装置驱动或牵引的交通工具。
应理解，本申请实施例涉及的“蠕行”，是以低于或等于某一速度行驶的状态，示例性地，该“某一速度”可以为10km/h，或者也可以为5km/h，或者也可以为其他数值。应理解，在自车处于蠕行状态时，自车的规划运动轨迹不易受博弈目标轨迹预测精度的影响。
图3示出了本申请实施例提供的一种轨迹规划的***架构图。如图3所示，轨迹规划的***包括感知模块、决策规划模块、控制模块、参数辨识模块和执行器。其中，感知模块可以包括图1所示的感知***120中的一种或多种摄像装置，或者一种或多种雷达传感器，用于采集智能驾驶设备的周围环境信息，智能驾驶设备实时运动参数等，感知模块还可以对采集的周围环境信息进行处理，为下游模块(即决策规划模块和控制模块)建立道路、障碍物等构成的世界模型；决策规划模块、控制模块可以为图1所示的计算平台150中的一个或多个处理器，具体地，决策规划模块用于根据周围环境信息确定智能驾驶设备的博弈策略，生成博弈目标的推演轨迹，并根据博弈策略和博弈目标的推演轨迹确定规划运动轨迹，控制模块用于根据规划运动轨迹计算相应的控制量，并将上述控制量输出到执行器。在执行器执行控制量时，控制智能驾驶设备按规划运动轨迹行驶。在一些可能的实现方式中，执行器也可以包括智能驾驶设备100中的转向、制动控制***。
其中，如图4所示，决策规划模块可以包括博弈目标筛选模块、交互博弈决策模块和运动规划模块。其中，博弈目标筛选模块，用于进行博弈目标筛选；交互博弈决策模块用于根据筛选的博弈目标推演博弈目标的推演轨迹和自车的推演轨迹，并根据博弈目标的推演轨迹和自车的推演轨迹生成博弈策略(或称语义级决策标签)；运动规划模块用于根据博弈策略和博弈目标的推演轨迹进行自车的运动轨迹规划，生成规划运动轨迹。
以下结合图5至图17说明以上三个模块的详细工作流程。
图5示出了本申请实施例提供的一种运动轨迹规划的方法500的示意性流程图，该方法500可以应用于图1所示的智能驾驶设备中，也可以由图3或图4所示的***执行。示例性地，该方法500可以由图4中的博弈目标筛选模块执行。以下以该智能驾驶设备为车辆为例，介绍方法500。应理解，图5示出的运动轨迹规划的方法的步骤或操作仅为示例性说明，本申请实施例还可以执行其他操作或者图5中的各个操作的变形。该方法500包括：
S501，获取与自车之间的距离小于或等于预设距离的交互物体的运动状态信息和道路信息。
示例性地，交互物体包括但不限于智能驾驶设备、行人、电动车、自行车，以及其他的物体。
示例性地，上述预设距离可以为100米(meter，m)，或者可以为200m，或者也可以为其他距离。
示例性地，交互物体的运动状态信息可以包括但不限于交互物体的预测运动轨迹，以及第二运动参数，如位置信息、速度、加速度等。示例性地，交互物体的运动状态信息可以是自车通过雷达传感器等测量的交互物体的位置信息和/或速度信息计算的；或者，交 互物体的运动状态信息可以是通过车与基础设施通信(vehicle to infrastructure，V2I)车对外界的信息交换(vehicle to everything，V2X)接收的，或者也可以为通过车与车通信(vehicle to vehicle，V2V)接收的，或者也可以是通过其他方式接收的。
示例性地，自车可以通过V2I，或V2V，或V2X直接接收交互物体的预测运动轨迹；又一示例，交互物体的预测运动轨迹可以是自车根据接收的交互物体的位置信息和航向角信息等计算出的，其中，位置信息和航向角信息等可以包含在车辆基本安全消息(basic vehicle safety messages，BSM)信息中，或者也可以包含在其他信息中。
应理解，上述交互物体的预测运动轨迹可以理解为预测的交互物体在未来一段时长内的运动轨迹。示例性地，该“未来一段时长”可以为10秒，或者也可以为20秒。
示例性地，道路信息包括但不限于当前道路的车道线信息、当前道路中一个或多个车道的导向规则等。该道路信息可以为自车通过V2I，或V2V，或V2X接收的，或者也可以为自车根据通过雷达传感器和/或摄像装置采集并计算得出的。
S502，确定预测运动轨迹与自车的规划行驶路径有交叉的交互物体为初筛博弈目标。
示例性地，该规划行驶路径可以为通过自车的规控模块获取的。
在一些可能的实现方式中，可以根据交互物体的预测运动轨迹确定交互物体的运动状态，进一步地，结合交互物体的运动状态确定初筛博弈目标。例如，在交互物体沿着预测运动轨迹行驶，认为交互物体运动状态明确，则交互物体的预测运动轨迹与自车的规划行驶路径有交叉时，即认为交互物体为初筛博弈目标。
例如，如图6中的(a)所示，自车直行，交互物体受到障碍物干扰需占用自车行驶车道，或者交互物体从自车右侧汇入，此时可以确定交互物体的预测运动轨迹与自车的规划行驶路径有交叉时，确定该交互物体为初筛博弈目标。或者，如图6中的(b)所示，自车在无指示灯路口直行，而交互物体在该路口的对向车道左转；或者自车在无指示灯路口左转，而交互物体在对向直行等，均可认为交互物体的预测运动轨迹与自车的规划行驶路径有交叉，则可以将该交互物体确定为初筛博弈目标。
又一示例，交互物体实际运动轨迹与预测运动轨迹不同，则认为交互物体的运动状态不明确，进一步地，可以根据交互物体的驾驶意图确定交互物体的新的预测运动轨迹，进而判断新的预测运动轨迹与自车的规划行驶路径是否有交叉。
示例性地，可以根据基于语义意图的行为预测方法确定交互物体的驾驶意图。进一步地，在确定交互物体的驾驶意图与自车运动轨迹有交叉时，认定该交互物体为初筛博弈目标。
例如，如图6中的(c)所示，在自车在无指示灯路口直行，交互物体在自车左侧，可能直行，也可以左转或右转，则可以根据交互物体的转向灯情况确定交互物体的驾驶意图。进一步地，在交互物体直行或左转时，与自车规划行驶路径有交叉，此时可认定该交互物体为初筛博弈目标。
S503，判断初筛博弈目标是否为危险目标。
具体地，若初筛博弈目标不属于危险目标，则执行S504，根据初筛博弈目标确定自车的博弈策略，即针对该初筛博弈目标进入自动驾驶博弈决策及轨迹规划；否则，执行S505，结束自动驾驶决策及轨迹规划流程。
需要说明的是，上述“结束自动驾驶决策及轨迹规划流程”是指不进行本申请实施例 提供的轨迹规划流程。应理解，在具体实现过程中，自车可以执行其它驾驶决策及轨迹规划。
还需说明的是，本申请实施例中涉及的“危险目标”是指与自车之间的空间距离和/或时间距离过近，以至于不适合针对其进行本申请实施例提供的包括蠕行策略的轨迹规划的交互物体。应理解，针对该危险目标可以进行其他自动驾驶决策和/或轨迹规划，例如执行自动紧急刹车。
示例性地，可以通过对初筛博弈目标进行CIPV筛选，确定初筛博弈目标是否为危险目标。CIPV筛选示意图如图7所示，基于安全考虑在自车行驶方向的正前方设置一个随速度变化，且关于自车纵向对称平面对称的区域(例如等腰梯形或者花瓶形状)进行危险目标的筛选。如图7所示，CIPV筛选区域纵向，即平行于自车行驶的方向(图7中的S方向)长度由ef、fg以及gd三部分组成，其中，ef是常量阈值，用于表征自车车头向车尾方向一定纵向范围也是危险目标筛选区域；fg是自车匀速行驶阈值，用于表征自车以当前车速行驶时，驾驶员接到紧急停车信号时至对智能驾驶设备进行制动时的反应时间内行驶的距离；gd是自车制动距离阈值，用于表征自车在当前工况下以最大减速度刹停所需的距离。CIPV筛选区域纵向长度的公式化表达如下：
Figure PCTCN2022113426-appb-000001
其中，L ef表示ef段的长度，longitude_const_thresh表示纵向常量阈值；L fg表示fg段的长度，v ego_current表示自车当前时刻的速度，t reaction表示驾驶员接到紧急停车信号时至对智能驾驶设备进行制动时的反应时间；L gd表示gd段的长度，a max表示自车在当前工况下减速的最大加速度。
CIPV筛选区域横向，即垂直于自车的纵向对称平面的方向(图7中的L方向)范围由ab和dc的长度确定，ab和dc均是横向常量阈值，且ab的长度可以大于dc。CIPV筛选区域的横向长度从ab沿着纵向方向到dc，横向长度可以沿着纵向方向线性或者非线性缩小，其中，图7所示CIPV筛选区域示意图为横向长度线性缩小。
在一些可能的实现方式中，可以通过调整纵向常量阈值(如ef)、横向常量阈值(如ab、dc)、横向长度缩小变化程度、自车速度、自车加速度、驾驶员反应时间等来动态改变CIPV筛选区域的形状，从而对初筛博弈目标的车型以及自车的激进程度倾向进行适配。示例性地，初筛博弈目标的车型越大，则ed长度越长，和/或ab和/或cd的长度越长；或者，自车的激进程度越高，则ed长度越短，和/或ab和/或cd的长度越短。
进一步地，根据上述CIPV筛选区域计算初筛博弈目标在横向上，即垂直于智能驾驶设备纵向对称平面的方向上，与自车发生轨迹冲突或意图冲突的时长ttl，计算公式如下：
Figure PCTCN2022113426-appb-000002
其中，L obj_l表示初筛博弈目标横向上与自车的距离；V obj_l表示初筛博弈目标横向上的速度。
进一步地，计算初筛博弈目标在纵向上，即平行于智能驾驶设备行驶的方向，与自车之间的速度差ΔV，计算公式如下：
ΔV＝V obj_s-V ego_s，
其中，V obj_s表示初筛博弈目标纵向的速度，V ego_s表示自车纵向的速度。初筛博弈目标被确定为危险目标需要同时满足如下条件：(a)初筛博弈目标在CIPV筛选区域内；(b)初筛博弈目标的ttl小于或等于第一预设阈值ttl_thresh；(c)博弈目标的ΔV小于或等于第二预设阈值ΔV_thresh。其中，第一预设阈值ttl_thresh是根据初筛博弈目标的纵向位置、车型以及自车激进程度确定的，可以根据上述信息中一个或多个进行动态调整；第二预设阈值ΔV_thresh是根据初筛博弈目标的车型以及自车激进程度确定的，可以根据初筛博弈目标的车型和/或自车激进程度进行动态调整。
应理解，如果初筛博弈目标被判定是“危险目标”，那么后续就不对这个“危险目标”进行本申请的基于蠕行策略的轨迹规划，而是执行例如自动紧急制动(autonomous emergency braking，AEB)等操作；如果初筛博弈目标被判定不是“危险目标”，那么则对该初筛博弈目标继续进行本申请的基于蠕行策略的轨迹规划。
S504，根据初筛博弈目标确定自车的博弈策略。
需要说明的是，在此步骤中的初筛博弈目标即为后续实施例(例如方法800)中，基于蠕行策略对自车进行轨迹规划时需要考虑的“博弈目标”。
S505，结束自动驾驶决策及轨迹规划流程。
本申请实施例提供的一种运动轨迹规划的方法，能够基于轨迹冲突、意图冲突以及CIPV方法初步筛选出基于蠕行策略对自车进行轨迹规划时需要考虑的博弈目标，能够保证在相对安全的环境下进行博弈策略决策及轨迹规划，有助于提高行车安全性。
图8示出了本申请实施例提供的一种运动轨迹规划的方法的示意性流程图800，该方法800可以应用于图1所示的智能驾驶设备中，也可以由图3或图4所示的***执行。示例性地，该方法800可以由图4中的交互博弈决策模块执行。在一些可能的实现方式中，该方法800可以在上述方法500之后执行，例如，该方法800可以视为对S504的扩展。应理解，图8示出的运动轨迹规划的方法的步骤或操作仅为示例性说明，本申请实施例还可以执行其他操作或者图8中的各个操作的变形。该方法800包括：
S801，生成采样空间。
在本申请实施例中，采样空间为用于描述自车和/或博弈目标的可能性状态的空间，上述可能性状态包括但不限于速度、加速度、加速度变化率等运动学状态，上述可能性状态也可以包括自车和/或博弈目标的其他运动学状态。应理解，该可能性状态空间具有状态取值上限和状态取值下限，该状态取值上限为车辆可能达到的状态的最大取值，例如自车和/或博弈目标能够达到的最大加速度、最大速度、最大加速度变化率；该状态取值下限为自车和/或博弈目标可能达到的状态的最小取值，例如车辆能够达到的最小加速度、最小速度、最小加速度变化率。在对自车和博弈目标进行轨迹推演时，根据在采样空间中采集的自车和博弈目标的可能性状态推演自车和博弈目标的轨迹。本申请中，以采样空间中的可能性状态为加速度为例进行说明。
示例性地，基于自车与博弈目标的运动状态信息，考虑道路限速、加速度变化率(Jerk值)、博弈目标类型等纵向特征信息，生成纵向加速度采样空间。基于道路边界、静止障碍物、智能驾驶设备运动学等特征信息，生成横向偏移采样空间。
需要说明的是，本申请实施例中涉及“纵向”可以理解为在平行于地面的平面中，平行于智能驾驶设备纵向对称平面的方向，即智能驾驶设备行驶方向；本申请实施例中涉及“横 向”可以理解为垂直于智能驾驶设备纵向对称平面的方向，即在平行于地面的平面中，垂直于智能驾驶设备行驶方向。
示例性地，上述博弈目标可以理解为上述实施例中的不属于危险目标的初筛博弈目标，或者也可以是通过其他方式筛选出的博弈目标。
在一些可能的实现方式中，对博弈目标和自车分别取最小横向偏移、当前横向偏移和最大横向偏移，可以得到3*3＝9个横向偏移采样空间。横向偏移采样空间中的采样用于表示自车或博弈目标的轨迹在垂直于智能驾驶设备纵向对称平面方向上的偏移。每个横向偏移采样空间中可以进行多个纵向采样空间采样。示例性地，自车的纵向采样空间可以分为自车纵向蠕行加速度采样空间和纵向非蠕行加速度采样空间，二者采样区间和采样间隔可以相同，区别在于：在进行轨迹推演时，基于自车的纵向非蠕行加速度和纵向蠕行加速度，进行的轨迹推演的上界和下界速度不同。在具体实现过程中，可以通过使用“标签”等方式，对在纵向蠕行加速度采样空间中的采样和纵向非蠕行加速度采样空间中的采样进行区分。博弈目标的纵向加速度采样空间由纵向非蠕行加速度采样空间构成。将自车的横纵向采样空间与博弈目标的横纵向采样空间交叉组合，形成最终的采样空间，如图9所示。具体地，图9所示的自车和博弈目标的横向偏移采样区间分别为[minEgoLateraloffset，maxEgoLateraloffset]和[minObjLateraloffset，maxObjLateraloffset]，图9中自车纵向非蠕行加速度采样空间和纵向蠕行加速度采样空间的区间均为[-4,3]m/s 2，博弈目标的纵向加速度采样空间的区间为[-4,3]m/s 2，取采样间隔均为1m/s 2，可以得到16*8＝128的纵向采样空间，自车与博弈目标构成的采样空间可以包括9*128＝1152个采样空间。应理解，横纵向采样空间的具体采样间隔可以为其他间隔，可以根据硬件计算性能及精度需求确定。例如，纵向采样间隔可以2m/s 2，或者也可以为0.5m/s 2；再例如，横向采样间隔还可以继续细化为5*5＝25的横向偏移采样空间。
S802，基于采样空间的采样根据自车当前位置对自车进行横向路径推演，根据自车当前加速度对自车进行纵向轨迹推演。
在一些可能的实现方式中，基于采样空间的采样，分别推演自车的横向路径和纵向轨迹，其中，横向路径包括偏移的路径曲线形状，如图10所示；纵向轨迹包括相应时刻(t)的纵向距离(s)之间的关系，最终根据横向路径和纵向轨迹生成带时间信息的多组推演轨迹。
示例性地，基于图7所示的SL坐标，使用如下公式推演自车的横向路径：
Figure PCTCN2022113426-appb-000003
其中，s e是自车在进行横向路径推演时刻所处的纵向位置(单位：m)，该位置随推演时刻变化。l(s e)是自车纵向位置s e对应的横向偏移(单位：m)；s eStartThresh是自车纵向的起始位置(单位：m)，该自车纵向的起始位置可以理解为在开始轨迹推演的时刻，自车在平行于智能驾驶设备纵向对称平面的方向上的位置；curEgoLateraloffset是自车的当前朝向对应的偏移(单位：m)；s eCubicCurveThresh是三次曲线连接结束的纵向位置(单位：m)；egolateralOffset是自车三次曲线连接结束的横向偏移(单位：m)，对应自车的 minEgoLateralOffset或者maxEgoLateralOffset；s eStartThresh和s eCubicCurveThresh可以根据自车的车型以及激进程度进行动态调整。示例性地，自车车型为大型车时，s eStartThresh和/或s eCubicCurveThresh的数值较大；和/或自车的激进程度较高时，s eStartThresh和/或s eCubicCurveThresh的数值较小。从(s eStartThresh,curEgoLateralOffset)到(s eCubicCurveThresh,egolateralOffset)使用三次曲线进行路径的连接，且三次曲线在(s eStartThresh,curEgoLateralOffset)和(s eCubicCurveThresh,egolateralOffset)处的切线方向与智能驾驶设备的行驶方向平行，其中，a、b、c、d为三次多项式的系数，其具体取值可以根据自车的车型以及激进程度确定，本申请对其具体取值不作限定。
在进行自车纵向轨迹推演过程中，自车纵向加速度随推演时间变化可以如图11所示，其中，横轴代表推演时刻t，纵轴代表自车的加速度。示例性地，自车在轨迹推演过程中某时刻的纵向位置可以根据推演的自车纵向加速度以及自车速度确定，自车在轨迹推演过程中多个时刻的纵向位置构成自车的纵向轨迹。如图11所示，自车纵向加速度推演可以包括如下四段：(1)时延段(t∈[0,delayTime))：基于自车当前加速度currentAcc进行纵向轨迹推演；(2)匀加速度变化率段(t∈[delayTime,jerkChangeTime)，以下简称匀Jerk段)：根据匀Jerk段开始的加速度和纵向采样加速度targetAcc确定加速度变化率，按匀Jerk进行纵向轨迹推演；(3)匀加速度段(t∈[jerkChangeTime,speedLimitTime))：保持纵向采样加速度targetAcc进行轨迹推演；(4)匀速段(t≥speedLimitTime)：推演速度达到上界速度或者下界速度时，则保持上界速度或者下界速度进行轨迹推演。需要说明的是，上述匀jerk段可以是加速度线性增加的过程，或者也可以是加速度线性减小的过程。还需说明的是，在纵向采样加速度为纵向非蠕行加速度时，纵向轨迹推演的上界速度可以为自车设置的上限速度或者自车所行驶路段的最高限速，纵向轨迹推演的下界速度可以为0。对于纵向采样加速度为纵向蠕行加速度时，当自车当前速度低于蠕行速度，纵向轨迹推演的上界速度可以为蠕行速度，纵向轨迹推演的下界速度可以为0；当自车当前速度高于或者等于蠕行速度时，纵向轨迹推演的上界速度可以为自车设置的上限速度或者道路最高限速，纵向轨迹推演的下界速度可以为蠕行速度。示例性地，该蠕行速度可以为***预设的，或者也可以为用户自己设定的，例如，可以为5km/h或10km/h，或者也可以为其他数值，本申请实施例对此不作具体限定。
应理解，上述自车纵向轨迹推演的四个阶段仅为示例性说明，在具体实现过程中，对自车纵向轨迹进行推演可以仅包括上述阶段中的一个或两个或三个。一示例，在自车当前加速度currentAcc等于纵向采样加速度targetAcc时，可以从(3)匀加速度段开始进行纵向轨迹推演。又一示例，在自车当前加速度currentAcc小于或等于纵向采样加速度targetAcc，且自车当前速度达到上界速度或者下界速度时，则可以从(4)匀速度段进行轨迹推演。
S803，基于采样空间的采样根据博弈目标当前位置对博弈目标进行横向路径推演，根据博弈目标当前加速度对博弈目标进行纵向轨迹推演。
在一些可能的实现方式中，博弈目标轨迹推演的方法可以与自车轨迹推演方法相同。示例性地，基于图7所示的SL坐标，使用如下公式推演博弈目标的横向路径：
Figure PCTCN2022113426-appb-000004
其中，s o是博弈目标纵向位置(单位：m)，该位置随推演时刻变化；l(s o)是博弈目标纵向位置对应的横向偏移(单位：m)；s oStartThresh是博弈目标纵向的起始位置；curObjLateralOffset是博弈目标的当前朝向对应的偏移(单位：m)；s oCubicCurveThresh是三次曲线连接结束的纵向位置(单位：m)；objlateralOffset是博弈目标三次曲线连接结束的横向偏移(单位：m)，对应博弈目标的minObjLateralOffset或者maxObjLateralOffset；s oStartThresh和s oCubicCurveThresh可以根据博弈目标的车型以及激进程度进行动态调整。示例性地，博弈目标车型为大型车时，s oStartThresh和/或s oCubicCurveThresh的数值较大；和/或博弈目标的激进程度较高时，s oStartThresh和/或s oCubicCurveThresh的数值较小。从(s oStartThresh,curObjLateralOffset)到(s oCubicCurveThresh,objlateralOffset)使用三次曲线进行路径的连接且三次曲线在(s oStartThresh,objcurLateralOffset)和(s oCubicCurveTheesh,objlateralOffset)处切线方向与智能驾驶设备的朝向平行，a、b、c、d为三次多项式的系数，该系数取值可以与自车横向路径推演公式中的a、b、c、d取值相同，也可以不同。
在一些可能的实现方式中，博弈目标纵向轨迹推演的方法可以参照自车纵向轨迹推演中的描述。示例性地，如图11所示，博弈目标纵向加速度推演可以包括如下四段：(1)时延段(t∈[0,delayTime))：基于博弈目标当前加速度currentAcc进行纵向轨迹推演；(2)匀Jerk段(t∈[delayTime,jerkChangeTime))：根据匀Jerk段开始的加速度和纵向采样加速度targetAcc确定加速度变化率，按匀Jerk进行纵向轨迹推演；(3)匀加速度段(t∈[jerkChangeTime,speedLimitTime))：保持纵向采样加速度targetAcc进行轨迹推演；(4)匀速段(t≥speedLimitTime)：推演速度达到上界速度或者下界速度时，则保持上界速度或者下界速度进行轨迹推演。示例性地，博弈目标的纵向轨迹推演的上界速度可以为博弈目标所行驶路段的最高限速，纵向轨迹推演的下界速度可以为0。需要说明的是，上述匀jerk段可以是加速度线性增加的过程，或者也可以是加速度线性减小的过程。
应理解，上述博弈目标纵向轨迹推演的四个阶段仅为示例性说明，在具体实现过程中，对博弈目标纵向轨迹进行推演可以仅包括上述阶段中的一个或两个或三个。
应理解，根据推演的自车(或博弈目标)的横向路径和纵向轨迹可以确定出自车(或博弈目标)的推演轨迹，该推演轨迹为自车(或博弈目标)的坐标随时间变化的点组成的集合，即(或博弈目标)的纵向轨迹和横向偏移融合后的结果。
S804，对自车和博弈目标的推演轨迹对进行策略代价评价。
应理解，基于图9所示的采样空间，在采样空间中每采一个样，可以确定一个自车的纵向采样加速度和横向偏移，以及一个博弈目标的纵向采样加速度和横向偏移。应理解，基于上述采样确定的自车的推演轨迹和博弈目标的推演轨迹构成推演轨迹对。
示例性地，对自车和博弈目标所有的推演轨迹对进行策略代价(以下简cost)评价，策略代价越小，则策略收益越高，则该策略被作为最优策略的可能性越高。在一些可能的实现方式中，可以从安全、舒适、通过性、路权、偏移等五个维度中的一个或多个进行策略收益评价。
在一些可能的实现方式中，可以基于推演轨迹对中自车推演轨迹和博弈目标推演轨迹之间的最小距离，同时考虑自车和博弈目标的车头朝向和速度，确定安全性cost。随速度 以及随朝向的最小距离越小则安全性cost越大。当随速度以及随朝向的最小距离小于一定阈值时，则判定自车和博弈目标在轨迹推演的最小距离处发生了碰撞。根据碰撞时自车和博弈目标的运动状态以及位置姿态区分碰撞类型为第一类碰撞还是第二类碰撞。为便于描述，本申请实施例中，将第一类碰撞简记为“E2O”，将第二类碰撞简记为“O2E”。
示例性地，在自车和博弈目标均处于行驶状态，在自车在其行驶方向上与该博弈目标的侧围或尾部发生碰撞时，则将该碰撞认定为是第一类碰撞；在博弈目标在该其行驶方向上与自车的侧围或尾部发生碰撞时，则将该碰撞认定为是第二类碰撞。或者，博弈目标处于静止时，自车与博弈目标发生碰撞，无论碰撞位置为博弈目标的哪个部位，均将该碰撞认定为是第一类碰撞；自车处于静止时，博弈目标与自车发生碰撞，无论碰撞位置为自车的哪个部位，均将该碰撞认定为是第二类碰撞。
应理解，车辆的侧围包括车身左侧围和车身右侧围。示例性地，车身左右两侧(A、B、C柱)、外覆盖件所构成的左、右两侧称为车身左、右侧围。示例性地，图12所示的虚线所指示区域为车辆的车身左侧围。其中，A柱设置在侧围前部，与前围板连接，还提供前门铰链和前风挡玻璃的安装，是前侧门开关转动的支撑立柱；B柱设置在侧围中部，提供前排安全带、前门锁扣以及后侧门铰链的安装，是后侧门开关的转动的支撑立柱(部分小型商用车，只有左右前门，B柱位于侧围后部)；C柱设置在侧围后部，为后排安全带、后侧门锁扣提供安装，是后窗玻璃或后门的支撑安装立柱；D柱与顶盖后横梁等部件一起构成后门框，为后三角窗和后门框提供支撑的立柱。在一些可能的实现方式中，上述车身侧围包含于侧围总成中，侧围总成还可以包括内板及加强件。
在一些可能的实现方式中，车辆的侧围还包括车身左、右侧围和车轮。例如，车辆的左侧围包括车身左侧围和车辆左侧车轮(例如，包括左侧前车轮和左侧后车轮)；车辆的右侧围包括车身右侧围和车辆右侧车轮(例如，包括右侧前车轮和右侧后车轮)。
应理解，车辆的尾部可以为车辆设置后保险杆、后牌照板的一侧。
需要说明的是，当根据自车的推演轨迹和博弈目标的推演轨迹确定未发生碰撞时，安全性cost为0。当碰撞类型为第一类碰撞时，安全性cost为一个近似无穷大的值；当碰撞类型为第二类碰撞时，可以根据碰撞位置处博弈目标的速度确定安全性cost。
示例性地，安全性cost计算公式可以如下所示：
Figure PCTCN2022113426-appb-000005
其中，Cost safety表示安全性cost；v表示根据自车的推演轨迹和博弈目标的推演轨迹确定发生碰撞时，碰撞位置处博弈目标的速度，vThresMin表示自车的推演轨迹与博弈目标的推演轨迹发生碰撞时，随博弈目标速度进行安全性cost惩罚的速度下限阈值；vThresMax表示自车的推演轨迹与博弈目标的推演轨迹发生碰撞时，随博弈目标速度进行安全性cost惩罚的速度上限阈值；m为权重值，在一些可能的实现方式中，该权重值与博弈目标的类型相关。
又一示例中，可以根据自车(或博弈目标)的推演轨迹中，自车(或博弈目标)的加速度变化率(jerk值)确定自车和博弈目标的推演轨迹对的舒适性cost。自车(或博弈目标)舒适性cost的计算公式可以为：
Figure PCTCN2022113426-appb-000006
其中，Cost comfortable表示舒适性cost，jerk表示根据自车(或博弈目标)的加速度变化率，jerkThresMin表示自车(或博弈目标)的加速度变化率下限阈值；jerkThresMax表示自车(或博弈目标)的加速度变化率上限阈值。在一些可能的实现方式中，在jerk小于jerkThresMin时，Cost comfortable取0；在jerk大于jerkThresMax时，Cost comfortable取1。应理解，舒适性越好，则舒适性cost越小。舒适性cost由自车舒适性cost和博弈目标舒适性cost组成，自车舒适性cost和博弈目标舒适性cost在舒适性cost中所占的权重可以不同。
再一示例中，自车和博弈目标的推演轨迹对的通过性cost计算方法可以如下所示。自车(或博弈目标)通过性cost的计算公式可以为：
Figure PCTCN2022113426-appb-000007
其中，Cost passability表示通过性cost；deltaT表示自车(或博弈目标)当前推演轨迹通过冲突点的时间，与自车(或博弈目标)以横向采样为当前朝向对应的偏移且纵向采样为0的推演轨迹通过冲突点的时间之间的差值，tThresMin表示时间差值的下限阈值；tThresMax表示时间差值的上限阈值。在一些可能的实现方式中，在deltaT小于tThresMin时，Cost passabiiity取0；在deltaT大于tThresMax时，Cost passability取1。应理解，通过性越好，则通过性cost越小。通过性cost由自车通过性cost和博弈目标通过性cost组成，自车通过性cost和博弈目标通过性cost在通过性cost中所占权重可以不同。
再一示例中，根据自车和博弈目标的推演轨迹对判断自车与博弈目标的路权关系。若交互博弈使得高路权的智能驾驶设备改变运动状态，则该运动状态的改变应该施以较高的路权cost惩罚，即高路权的智能驾驶设备倾向于不改变当前运动状态。自车和博弈目标的推演轨迹对的路权cost的计算可以如下所示。自车(或博弈目标)路权cost的计算公式可以为：
Figure PCTCN2022113426-appb-000008
其中，Cost roadRight表示路权cost；acc表示高路权智能驾驶设备当前的纵向采样加速度；accThresMin表示纵向加速度的下限阈值；accThresMax表示纵向加速度的上限阈值。在一些可能的实现方式中，在acc大于accThresMax时，Cost roadRight取0；在acc大于accThresMin时，Cost roadRight取1。路权cost由自车路权cost和博弈目标路权cost组成，自车路权cost和博弈目标路权cost在路权cost中所占权重可以不同。
需要说明的是，在根据自车和博弈目标的推演轨迹对进行策略cost评价时，可以将上述各cost加权求和，获得当前自车和博弈目标的推演轨迹对的最终cost。各cost权重分配可以为：安全性权重>路权权重＝通过性权重>舒适性权重，或者也可以为其他分配方式，本申请对此不作具体限定。
还需说明的是，在根据自车和博弈目标的推演轨迹对进行策略cost评价时，可以采用上述各cost中的一个或多个，或者也可以在上述cost的基础上做变形和/或扩展。
S805，根据策略代价最小的推演轨迹对确定初始博弈策略。
在一些可能的实现方式中，根据自车和博弈目标所有推演轨迹对中策略cost最小的推演轨迹对确定初始博弈策略。对于初始博弈策略中的纵向博弈策略，若根据轨迹推演，自车比博弈目标先过冲突区域，则初始博弈策略的纵向博弈策略为抢行博弈目标，反之，则 让行博弈目标。在一些可能的实现方式中，根据纵向采样加速度确定纵向博弈策略为蠕行策略，还是非蠕行策略。示例性地，在自车的纵向采样加速度属于自车蠕行加速度采样空间时，纵向博弈策略为蠕行策略，在自车的纵向采样加速度属于自车非蠕行加速度采样空间时，纵向博弈策略为非蠕行策略。对于初始博弈策略中的横向博弈策略，若自车横向采样偏移不为当前朝向对应的偏移(curEgoLateralOffset)，则最优解的横向博弈策略为绕行博弈目标，反之，则忽略博弈目标。对于纵向博弈策略中的非蠕行策略，抢行博弈目标的决策标签为GRABWAY(简记为GW)，让行博弈目标的决策标签为YIELD(简记为YD)；对于蠕行策略，若根据轨迹推演，自车比博弈目标先过冲突区域，则属于蠕行抢行，反之，则属于蠕行让行，蠕行抢行博弈目标的决策标签为CREEP-FORWARD-GRABWAY(简记为CFG)，蠕行让行博弈目标的决策标签为CREEP-FORWARD-YIELD(简记为CFY)。对于横向博弈策略，绕行博弈目标的决策标签为BYPASS(简记为BP)，忽略博弈目标的决策标签为IGNORE(简记为IG)。示例性地，如图13中的(a)所示，初始博弈策略为绕行-抢行策略BP-GW；图13中的(b)所示，初始博弈策略为忽略-蠕行让行策略IG-CFY。
还需说明的是，蠕行策略和非蠕行策略的区别在于：运动规划模块在根据蠕行策略进行抢行的轨迹规划时，尽可能使自车以蠕行速度通过冲突区域；在根据蠕行策略进行让行的轨迹规划时，尽可能使自车在博弈目标通过冲突区域的时段内以蠕行速度行驶；即，可以理解为通过蠕行策略指示运动规划模块基于蠕行速度，为自车规划其在冲突区域的轨迹，或自车在博弈目标通过冲突区域时段内的轨迹；而运动规划模块在根据非蠕行策略进行抢行或让行的轨迹规划时，自车在通过冲突区域时的行驶速度，或自车在博弈目标通过冲突区域时段内的行驶速度并没有一个基准值，其可能以很高的速度抢行博弈目标，也可能刹停让行博弈目标。
S806，对初始博弈策略进行稳定处理生成最终博弈策略。
在一些可能的实现方式中，基于分层状态机通过设置不同阈值的计时器使得非蠕行标签跳转蠕行标签(YD/GW->CFY/CFG)、蠕行抢行标签和蠕行让行标签相互跳转(CFY<->CFG)、蠕行标签跳非蠕行标签(CFY/CFG->YD/GW)的难度依次上升。博弈策略稳定处理的示意图如图14所示，其中，intervalTime为前一帧和当前帧的时间间隔；minTimerThres、midTimerThres和maxTimerThres均为计时器时间阈值且依次增大；ydToCfyTimer表示决策标签由YD跳CFY的计时器；ydToCfgTimer表示决策标签由YD跳CFG的计时器；cfyToYdTimer表示决策标签由CFY跳YD的计时器；cfyToGwTimer表示决策标签由CFY跳GW的计时器；cfyToCfgTimer表示决策标签由CFY跳CFG的计时器；cfgToYdTimer表示决策标签由CFG跳YD的计时器；cfgToGwTimer表示决策标签由CFG跳GW的计时器；cfgToCfyTimer表示决策标签由CFG跳CFY的计时器；gwToCfyTimer表示决策标签由GW跳CFY的计时器；gwToCfgTimer表示决策标签由GW跳CFG的计时器。
示例性地，minTimerThres、midTimerThres和maxTimerThres可以分别为0.1秒、0.3秒和0.5秒，或者，该minTimerThres、midTimerThres和maxTimerThres也可以取其他数值。
如图14中的(a)和(b)所示，在前一帧纵向决策标签为YD或GW时，若当前帧 纵向决策标签为CFY或CFG，则在相应计时器当前记录的时长上中叠加intervalTime，只有在相应计时器记录的时长大于计时器时间阈值minTimerThres时，纵向决策才会改为CFY或CFG，否则将当前帧纵向决策标签修改为前一帧纵向决策标签。
如图14中的(c)和(d)所示，在前一帧纵向决策标签为CFY或CFG时，若当前帧纵向决策标签为YD或GW，则在相应计时器当前记录的时长上中叠加intervalTime，只有在相应计时器时间大于计时器时间阈值maxTimerThres时，纵向决策才会改为YD或GW，否则将当前帧纵向决策标签修改为前一帧纵向决策标签。在前一帧纵向决策标签为CFY(或CFG)时，若当前帧纵向决策标签为CFG(或CFY)，则在相应计时器当前记录的时长上中叠加intervalTime，只有在相应计时器时间大于计时器时间阈值midTimerThres时，纵向决策才会改为CFG(或CFY)，否则将当前帧纵向决策标签修改为前一帧纵向决策标签。
本申请实施例提供的一种运动轨迹规划的方法，基于精细化的横纵向采样轨迹，在横向上考虑路径的合理性，在纵向上区分基于蠕行加速度采样空间以及非蠕行加速度采样空间进行采样，所进行轨迹推演的上界速度和下界速度，能够缩小推演轨迹与实际运动轨迹的差距。在保证安全和遵守交通法规的前提下，实现自车主动保持低速蠕行来试探博弈目标的行为，若博弈目标行为偏激进，使得自车与博弈目标的时间距离或者空间距离低于安全距离，则能够保证自车由蠕行转让行(非蠕行让行)；若博弈目标行为偏保守，使得自车有抢行博弈目标的空间，则能够使得自车寻找合适时机由蠕行转抢行(非蠕行抢行)。
本申请实施例中，在图4所示的交互博弈决策模块生成最终博弈策略后，运动规划模块根据该最终博弈策略生成满足智能驾驶设备运动学要求的速度曲线。以下以最终博弈策略为蠕行策略为例说明运动规划模块进行运动轨迹规划的方法。
示例性地，运动规划模块根据自车的规划行驶路径以及博弈目标的推演轨迹，确定博弈目标的第一位置空间占据自车的第二位置空间时的时间段，采用碰撞检测算法生成基于时间的空间约束。以图15所示的博弈场景为例，博弈目标在t1至t2时刻(即上述“时间段”)之间，占据自车规划行驶路径的s1至s2的空间(即上述“第二位置空间”)为自车的不可行空间，即图16中的(a)以及图17中的(a)的阴影部分所示的不可行空间，其中，t1为冲突开始时刻，t2为冲突结束时刻。
进一步地，根据最终博弈策略，确定开始蠕行的位置以及蠕行时长，进而为自车规划纵向轨迹。若最终博弈策略为蠕行让行，则结束蠕行让行的位置可以为s1处，蠕行时长可以为t1至t2之间的时长，进而根据结束蠕行让行的位置确定开始蠕行让行的位置，例如根据蠕行速度和蠕行时长确定蠕行距离，根据蠕行距离和结束蠕行让行的位置确定开始蠕行让行的位置，进而根据开始蠕行让行的位置和自车当前所处位置规划自车从当前所处位置至开始蠕行让行位置的速度。若最终博弈策略为蠕行抢行，则开始蠕行抢行的位置可以为s1处，并使得自车在t1时刻在s2处结束蠕行抢行，进而根据蠕行速度和蠕行行驶距离确定蠕行时长，进而根据蠕行时长和t1时刻确定开始蠕行的时刻，并根据自车当前所处位置、开始蠕行抢行的位置和时刻，规划自车从当前所处位置至开始蠕行抢行位置的速度。
在一些可能的实现方式中，考虑博弈目标行为的不确定性，引入时间以及空间上的安全距离。对于蠕行让行的场景，如图16中的(a)所示，对于时间上，引入TimeGap1和 TimeGap2，期望自车在距离冲突开始时间t1之前的TimeGap1内提前进入蠕行状态，在冲突结束时间t2的基础上增加TimeGap2结束蠕行状态，即蠕行时长CFTime＝TimeGap1+TimeGap2+(t2–t1)；在空间上，引入安全距离safeDis，期望让行的空间内，在原冲突的基础上保持额外的safeDis距离，此设计可以增加自车与博弈目标的空间距离，一定程度上增加蠕行安全性。进一步地，如图16的(b)所示，在蠕行时长CFTime内，智能驾驶设备保持蠕行速度Cf_Velocity行驶。对于蠕行抢行的场景，如图17中的(a)所示，对于时间上，引入TimeGap，使得期望自车在距离冲突开始时间t1之前的TimeGap内结束蠕行状态；在空间上，引入安全距离safeDis1和safeDis2，期望抢行的空间内，在原冲突的基础上保持额外的safeDis1和safeDis2距离，即蠕行距离CFdis＝safeDis1+safeDis2+(s2-s1)，蠕行时长CFTime＝CFTime/Cf_Velocity。进一步地，如图17的(b)所示，在蠕行时长CFTime内，智能驾驶设备保持蠕行速度Cf_Velocity行驶。此设计可以一定程度上保证博弈策略的尽量稳定，降低不确定性导致的博弈策略的跳变。
示例性地，该TimeGap1、TimeGap2、TimeGap可以分别为0.5秒，或者也可以为其他数值，三者的数值可以相同，也可以不同；该safeDis、safeDis1、safeDis2可以分别为0.5米，或者也可以分别为1米，或者也可以为其他数值，三者的数值可以相同，也可以不同。
图18示出了本申请实施例提供的一种运动轨迹规划的方法1800的示意性流程图，该流程所示方法可以应用于图1所示的智能驾驶设备，也可以通过图3或图4所示的***执行。该方法1800包括：
S1810，获取第一智能驾驶设备的规划行驶路径和第一运动参数，该第一运动参数包括该第一智能驾驶设备的速度和/或加速度。
示例性地，该第一智能驾驶设备可以为上述实施例中的自车，或者也可以为其他能够自主移动的物体，例如，智能机器人等。
示例性地，第一智能驾驶设备的规划行驶路径可以是从规控模块获取的；该第一运动参数可以是从感知***获取的，该第一运动参数可以包括但不限于速度、加速度、第一智能驾驶设备的位置与时间之间的对应关系等。
S1820，获取第一物体的预测运动轨迹和第二运动参数，该第二运动参数包括该第一物体的速度和/或加速度。
应理解，此处获取第一物体的预测运动轨迹和第二运动参数的主体，和获取第一智能驾驶设备的规划行驶路径和第一运动参数的主体可以为同一主体，例如，该同一主体可以为第一智能驾驶设备。
示例性地，该第一物体可以为上述实施例中的博弈目标，或者也可以为其他可能影响第一智能驾驶设备规划运动轨迹的物体；第二运动参数可以包括但不限于速度、加速度、第一物体的位置与时间之间的对应关系等
示例性地，获取该第一物体的预测运动轨迹和第二运动参数的方法可以参考上述实施例中的描述，例如，方法500的S501中的描述，在此不再赘述。
S1830，在该规划行驶路径和该预测运动轨迹存在交叉时，根据该第一运动参数和该第二运动参数确定该第一智能驾驶设备的规划运动轨迹，该规划运动轨迹包括该第一智能驾驶设备以第一预设速度抢行或让行该第一物体的运动轨迹。
示例性地，确定规划行驶路径和该预测运动轨迹存在交叉的方法可以参考上述实施例中的描述，例如，方法500的S502中的描述，在此不再赘述。
示例性地，第一博弈策略可以为上述实施例中的最终博弈策略，更具体地，可以为上述实施例中的纵向博弈策略中的蠕行策略，CFY或CFG。
在一些可能的实现方式中，该第一预设速度大于或等于3km/h，且小于或等于15km/h。
示例性地，该第一预设速度可以为上述实施例中的“蠕行速度”，或者也可以为其他使得第一智能驾驶设备的规划运动轨迹不易受第一物体轨迹预测精度影响的速度。
示例性地，根据该第一运动参数和该第二运动参数确定该规划运动轨迹的方法可以参考上述实施例中的描述，例如，方法800中的描述，在此不再赘述。
S1840，控制该第一智能驾驶设备按照该规划运动轨迹行驶。
可选地，该根据该第一运动参数和该第二运动参数确定规划运动轨迹，包括：根据第一采样加速度和该第一运动参数确定该第一智能驾驶设备的第一推演轨迹，该第一采样加速度为在采样空间中确定的该第一智能驾驶设备可能达到的加速度；根据第二采样加速度和该第二运动参数确定该第一物体的第二推演轨迹，该第二采样加速度为在采样空间中确定的该第一物体可能达到的加速度；在该第一推演轨迹和该第二推演轨迹指示该第一智能驾驶设备和该第一物体不会发生碰撞，或者该第一物体在该第一物体的行驶方向上与该第一智能驾驶设备在偏离该第一智能驾驶设备行驶方向的方向上在第一时刻发生碰撞时，根据该第一推演轨迹和该第二推演轨迹确定该第一博弈策略，该第一时刻为当前时刻之后的时刻。
示例性地，第一采样加速度和第二采样加速度可以为上述实施例中的纵向采样加速度，更具体地，该第一采样加速度可以为上述蠕行纵向采样空间中的纵向采样加速度。
示例性地，该第一推演轨迹和该第二推演轨迹可以为上述实施例方法800的推演轨迹，确定该第一推演轨迹和第二推演轨迹的具体方法可以参考方法800中S802至S804中的描述，在此不再赘述。
示例性地，根据第一推演轨迹和该第二推演轨迹确定该第一智能驾驶设备和该第一物体是否发生碰撞，以及碰撞类型的方法可以参考上述方法1100中的描述，在此不再赘述。
可以理解的是，本申请实施例中涉及的确定“推演轨迹”的步骤可以是由上述实施例中的交互博弈决策模块完成的，本申请实施例中涉及的确定“规划运动轨迹”的步骤可以是由上述实施例中的运动规划模块完成的。
可选地，该第一运动参数包括该第一智能驾驶设备的速度和/或加速度，该根据第一采样加速度和该第一运动参数确定该第一智能驾驶设备的第一推演轨迹，包括：根据该第一采样加速度，以及该第一智能驾驶设备的速度和/或加速度确定第一子推演轨迹，该第一子推演轨迹的终点处该第一智能驾驶设备的速度为第一预设速度；根据该第一预设速度确定第二子推演轨迹，该第一子推演轨迹的终点为该第二子推演轨迹的起点；根据该第一子推演轨迹和该第二子推演轨迹，确定该第一推演轨迹。
示例性地，该第一子推演轨迹可以包括上述实施例中时延段(t∈[0,delayTime))、匀加速度变化率段(t∈[delayTime,jerkChangeTime)和匀加速度段(t∈[jerkChangeTime,speedLimitTime))中至少一段的轨迹；该第二子推演轨迹可以为上述实施例中的匀速段的轨迹。
示例性地，确定该第一子推演轨迹和该第二子推演轨迹可以参考方法800中S802至S803中的描述，在此不再赘述。
在一些可能的实现方式中，该第一推演轨迹还包括第三子推演轨迹，该第三子推演轨迹可以包括从行驶速度由第一预设速度降至零(即刹停)的轨迹。
可选地，该根据该第一推演轨迹和该第二推演轨迹确定该规划运动轨迹，包括：根据该第一推演轨迹和该第二推演轨迹确定第一博弈策略，该第一博弈策略用于指示该第一智能驾驶设备以该第一预设速度抢行或让行该第一物体；在第一时长大于第一时间阈值时，根据该第一博弈策略、该规划行驶路径和该第二推演轨迹确定该规划运动轨迹。
示例性地，该第一博弈策略持续的时长可以为上述实施例中ydToCfgTimer、ydToCfyTimer、gwToCfgTimer、gwToCfyTimer中任一计时器的时长；或者，该第一博弈策略持续的时长可以为上述实施例中cfgToCfyTimer或cfyToCfgTimer中任一计时器的时长。
可选地，该根据该第一博弈策略、该规划行驶路径和该第二推演轨迹确定该规划运动轨迹，包括：根据该规划行驶路径和该第二推演轨迹确定该第一物体的第一位置空间占据该第一智能驾驶设备的第二位置空间时的时间段；基于该第一博弈策略，根据该第二位置空间和该时间段确定该规划运动轨迹，该规划运动轨迹包括该第一智能驾驶设备在该第二位置空间以该第一预设速度行驶的轨迹，或者，该规划运动轨迹包括该第一智能驾驶设备在该时间段内以该第一预设速度行驶的轨迹。
示例性地，该子规划运动轨迹可以为上述实施例中如图16或图17所示的在蠕行时长CFTime内，智能驾驶设备保持蠕行速度Cf_Velocity行驶的轨迹。
示例性地，更具体的根据该第一博弈策略、该规划行驶路径和该第二推演轨迹确定该规划运动轨迹的方法可以参考上述实施例中的描述，在此不再赘述。
可选地，该根据该第一推演轨迹和该第二推演轨迹确定该第一博弈策略之前，该方法还包括：确定该第一推演轨迹和该第二推演轨迹组成的推演轨迹对的策略代价最小。
示例性地，该策略代价可以为上述实施例中策略代价。
确定根据该第一物体的类型、该第一推演轨迹和该第二推演轨迹确定第一博弈策略的策略代价的方法可以参考方法800中S804中的描述，在此不再赘述。
可选地，该获取第一物体的预测运动轨迹和第二运动参数之前，该方法还包括：确定该第一物体与该第一智能驾驶设备之间的距离小于或等于第一距离阈值，且大于或等于第二距离阈值。
示例性地，该第一距离阈值可以为上述方法500中S501中的预设距离。
示例性地，该第二距离阈值可以是根据上述实施例中CIPV筛选确定的，或者也可以是通过其他方法确定的。
本申请实施例提供的一种运动轨迹规划的方法，能够在保证安全和遵守交通法规的前提下，实现自车主动保持蠕行速度来试探博弈目标的行为，使得自车的规划运动轨迹不易受到博弈目标的影响。特别地，在机动车、非机动车混行的窄道或者无指示灯交叉路口，自车不会再发生非预期地频繁误轻重刹、误刹停、持续保守让行以及陷入同起同停等问题，有助于提高自车通行效率，改善用户驾乘体验。
在本申请的各个实施例中，如果没有特殊说明以及逻辑冲突，各个实施例之间的术语 和/或描述具有一致性、且可以相互引用，不同的实施例中的技术特征根据其内在的逻辑关系可以组合形成新的实施例。
上文中结合图5至图18详细说明了本申请实施例提供的方法。下面将结合图19和图20详细说明本申请实施例提供的装置。应理解，装置实施例的描述与方法实施例的描述相互对应，因此，未详细描述的内容可以参见上文方法实施例，为了简洁，这里不再赘述。
图19示出了本申请实施例提供的一种运动轨迹规划的装置1900的示意性框图，该装置1900包括获取单元1910和处理单元1920。获取单元1910可以实现相应的通信功能，处理单元1920用于进行数据处理。
可选地，该装置1900还可以包括存储单元，该存储单元可以用于存储指令和/或数据，处理单元1920可以读取存储单元中的指令和/或数据，以使得装置实现前述方法实施例。
该装置1900可以包括用于执行图5、图8或图18中的方法的单元。并且，该装置1900中的各单元和上述其他操作和/或功能分别为了实现图5、图8或图18中的方法实施例的相应流程。
其中，当该装置1900用于执行图18中的方法1800时，获取单元1910可用于执行方法1800中的S1810和S1820，处理单元1920可用于执行方法1800中的S1830和S1840。
具体地，装置1900获取单元1910，用于获取该第一智能驾驶设备的规划行驶路径和第一运动参数，该第一运动参数包括该第一智能驾驶设备的速度和/或加速度；获取第一物体的预测运动轨迹和第二运动参数，该第二运动参数包括该第一物体的速度和/或加速度；处理单元1920，用于在该规划行驶路径和该预测运动轨迹存在交叉时，根据该第一运动参数和该第二运动参数确定该第一智能驾驶设备的规划运动轨迹，该规划运动轨迹包括该第一智能驾驶设备以第一预设速度抢行或让行该第一物体的运动轨迹；控制该第一智能驾驶设备按照该规划运动轨迹行驶。
在一些可能的实现方式中，该处理单元1920具体用于：根据第一采样加速度和该第一运动参数确定该第一智能驾驶设备的第一推演轨迹，该第一采样加速度为在采样空间中确定的该第一智能驾驶设备可能达到的加速度；根据第二采样加速度和该第二运动参数确定该第一物体的第二推演轨迹，该第二采样加速度为在该采样空间中确定的该第一物体可能达到的加速度；在该第一推演轨迹和该第二推演轨迹指示该第一智能驾驶设备和该第一物体不会发生碰撞，或者该第一物体在该第一物体的行驶方向上与该第一智能驾驶设备的侧围或尾部在第一时刻发生碰撞时，根据该第一推演轨迹和该第二推演轨迹确定该规划运动轨迹，该第一时刻为当前时刻之后的时刻。
在一些可能的实现方式中，该处理单元1920具体用于：根据该第一采样加速度，以及该第一智能驾驶设备的速度和/或加速度确定第一子推演轨迹，该第一子推演轨迹的终点处该第一智能驾驶设备的速度为第一预设速度；根据该第一预设速度确定第二子推演轨迹，该第一子推演轨迹的终点为该第二子推演轨迹的起点；该第一推演轨迹包括该第一子推演轨迹和该第二子推演轨迹。
在一些可能的实现方式中，该处理单元1920具体用于：根据该第一推演轨迹和该第二推演轨迹确定第一博弈策略，该第一博弈策略用于指示该第一智能驾驶设备以该第一预设速度抢行或让行该第一物体；在第一时长大于第一时间阈值时，根据该第一博弈策略、该规划行驶路径和该第二推演轨迹确定该规划运动轨迹。
在一些可能的实现方式中，该处理单元1920具体用于：根据该规划行驶路径和该第二推演轨迹确定该第一物体的第一位置空间占据该第一智能驾驶设备的第二位置空间时的时间段；基于该第一博弈策略，根据该第二位置空间和该时间段确定该规划运动轨迹，该规划运动轨迹包括该第一智能驾驶设备在该第二位置空间以该第一预设速度行驶的轨迹，或者，该规划运动轨迹包括该第一智能驾驶设备在该时间段内以该第一预设速度行驶的轨迹。
在一些可能的实现方式中，该处理单元1920具体用于：确定该第一推演轨迹和该第二推演轨迹组成的推演轨迹对的策略代价最小。
在一些可能的实现方式中，该处理单元1920具体用于：根据该第一智能驾驶设备的横向偏移、该第一采样加速度和该第一运动参数确定该第一推演轨迹，该横向偏移为垂直于该第一智能驾驶设备行驶方向上的偏移。
在一些可能的实现方式中，该处理单元1920还用于：确定该第一物体与该第一智能驾驶设备之间的距离小于或等于第一距离阈值，且大于或等于第二距离阈值。
在一些可能的实现方式中，该第一预设速度大于或等于3km/h，且小于或等于15km/h。
应理解，以上装置中各单元的划分仅是一种逻辑功能的划分，实际实现时可以全部或部分集成到一个物理实体上，也可以物理上分开。此外，装置中的单元可以以处理器调用软件的形式实现；例如装置包括处理器，处理器与存储器连接，存储器中存储有指令，处理器调用存储器中存储的指令，以实现以上任一种方法或实现该装置各单元的功能，其中处理器例如为通用处理器，例如CPU或微处理器，存储器为装置内的存储器或装置外的存储器。或者，装置中的单元可以以硬件电路的形式实现，可以通过对硬件电路的设计实现部分或全部单元的功能，该硬件电路可以理解为一个或多个处理器；例如，在一种实现中，该硬件电路为ASIC，通过对电路内元件逻辑关系的设计，实现以上部分或全部单元的功能；再如，在另一种实现中，该硬件电路为可以通过PLD实现，以FPGA为例，其可以包括大量逻辑门电路，通过配置文件来配置逻辑门电路之间的连接关系，从而实现以上部分或全部单元的功能。以上装置的所有单元可以全部通过处理器调用软件的形式实现，或全部通过硬件电路的形式实现，或部分通过处理器调用软件的形式实现，剩余部分通过硬件电路的形式实现。
在本申请实施例中，处理器是一种具有信号的处理能力的电路，在一种实现中，处理器可以是具有指令读取与运行能力的电路，例如CPU、微处理器、GPU、或DSP等；在另一种实现中，处理器可以通过硬件电路的逻辑关系实现一定功能，该硬件电路的逻辑关系是固定的或可以重构的，例如处理器为ASIC或PLD实现的硬件电路，例如FPGA。在可重构的硬件电路中，处理器加载配置文档，实现硬件电路配置的过程，可以理解为处理器加载指令，以实现以上部分或全部单元的功能的过程。此外，还可以是针对人工智能设计的硬件电路，其可以理解为一种ASIC，例如NPU、TPU、DPU等。
可见，以上装置中的各单元可以是被配置成实施以上方法的一个或多个处理器(或处理电路)，例如：CPU、GPU、NPU、TPU、DPU、微处理器、DSP、ASIC、FPGA，或这些处理器形式中至少两种的组合。
此外，以上装置中的各单元可以全部或部分可以集成在一起，或者可以独立实现。在一种实现中，这些单元集成在一起，以片上***(system-on-a-chip，SOC)的形式实现。 该SOC中可以包括至少一个处理器，用于实现以上任一种方法或实现该装置各单元的功能，该至少一个处理器的种类可以不同，例如包括CPU和FPGA，CPU和人工智能处理器，CPU和GPU等。
示例性地，上述获取单元1910可以包括图4所示的博弈目标筛选模块，上述处理单元1920可以包括图4所示的交互博弈决策模块和/或运动规划模块。
在具体实现过程中，上述获取单元1910和处理单元1920所执行的各项操作可以由同一个处理器执行，或者，也可以由不同的处理器执行，例如分别由多个处理器执行。一示例中，一个或多个处理器可以与图1中的感知***120中一个或多个传感器相连接，从一个或多个传感器中获取智能驾驶设备的第一运动参数，并对第一运动参数进行处理获得智能驾驶设备的推演轨迹。或者，一个或多个处理器还可以与智能驾驶设备的动力***相连接，控制智能驾驶设备按照规划运动轨迹行驶。示例性地，在具体实现过程中，上述一个或多个处理器可以设置在车机中的处理器，或者也可以为设置在其他车载终端中的处理器。示例性地，在具体实现过程中，上述装置1900可以为设置在车机或者其他车载终端中的芯片。示例性地，在具体实现过程中，上述装置1900可以为设置在智能驾驶设备中的如图1所示的计算平台150。
图20是本申请实施例的一种运动轨迹规划的装置的示意性框图。图20所示的运动轨迹规划的装置2000可以包括：处理器2010、收发器2020以及存储器2030。其中，处理器2010、收发器2020以及存储器2030通过内部连接通路相连，该存储器2030用于存储指令，该处理器2010用于执行该存储器2030存储的指令，以收发器2020接收/发送部分参数。可选地，存储器2030既可以和处理器2010通过接口耦合，也可以和处理器2010集成在一起。
需要说明的是，上述收发器2020可以包括但不限于输入/输出接口(input/output interface)一类的收发装置，来实现装置2000与其他设备或通信网络之间的通信。
在实现过程中，上述方法的各步骤可以通过处理器2010中的硬件的集成逻辑电路或者软件形式的指令完成。结合本申请实施例所公开的方法可以直接体现为硬件处理器执行完成，或者用处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器2030，处理器2010读取存储器2030中的信息，结合其硬件完成上述方法的步骤。为避免重复，这里不再详细描述。
处理器2010可以采用通用的CPU，微处理器，ASIC，GPU或者一个或多个集成电路，用于执行相关程序，以实现本申请方法实施例的运动轨迹规划的方法。处理器2010还可以是一种集成电路芯片，具有信号的处理能力。在具体实现过程中，本申请的运动轨迹规划的方法的各个步骤可以通过处理器2010中的硬件的集成逻辑电路或者软件形式的指令完成。上述处理器2010还可以是通用处理器、DSP、ASIC、FPGA或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成，或者用译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器 等本领域成熟的存储介质中。该存储介质位于存储器2030，处理器2010读取存储器2030中的信息，结合其硬件执行本申请方法实施例的运动轨迹规划的方法。
存储器2030可以是只读存储器(read-only memory，ROM)，静态存储设备，动态存储设备或者随机存取存储器(random access memory，RAM)。
收发器2020使用例如但不限于收发器一类的收发装置，来实现装置2000与其他设备或通信网络之间的通信。
示例性地，上述收发器2010可以包括图4所示的博弈目标筛选模块，上述处理器2020可以包括图4所示的交互博弈决策模块和/或运动规划模块。
本申请实施例还提供一种智能驾驶设备，该智能驾驶设备可以包括上述装置1900，或者上述装置2000。
本申请实施例还提供一种计算机可读介质，该计算机可读介质存储有程序代码或指令，当该计算机程序代码或指令被计算机的处理器执行时，使得该处理器实现上述图5、图8或图18中的方法。
本申请实施例还提供一种芯片，包括：至少一个处理器和存储器，所述至少一个处理器与所述存储器耦合，用于读取并执行所述存储器中的指令，以执行上述图5、图8或图18中的方法。
本申请将围绕包括多个设备、组件、模块等的***来呈现各个方面、实施例或特征。应当理解和明白的是，各个***可以包括另外的设备、组件、模块等，并且/或者可以并不包括结合附图讨论的所有设备、组件、模块等。此外，还可以使用这些方案的组合。
应理解，在本申请的各种实施例中，上述各过程的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本申请实施例的实施过程构成任何限定。
另外，在本申请实施例中，“示例的”、“例如”等词用于表示作例子、例证或说明。本申请中被描述为“示例”的任何实施例或设计方案不应被解释为比其它实施例或设计方案更优选或更具优势。确切而言，使用示例的一词旨在以具体方式呈现概念。
本申请实施例中，“相应的(corresponding，relevant)”和“对应的(corresponding)”有时可以混用，应当指出的是，在不强调其区别时，其所要表达的含义是一致的。
在本说明书中描述的参考“一个实施例”或“一些实施例”等意味着在本申请的一个或多个实施例中包括结合该实施例描述的特定特征、结构或特点。由此，在本说明书中的不同之处出现的语句“在一个实施例中”、“在一些实施例中”、“在其他一些实施例中”、“在另外一些实施例中”等不是必然都参考相同的实施例，而是意味着“一个或多个但不是所有的实施例”，除非是以其他方式另外特别强调。术语“包括”、“包含”、“具有”及它们的变形都意味着“包括但不限于”，除非是以其他方式另外特别强调。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的***、装 置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在本申请所提供的几个实施例中，应该理解到，所揭露的***、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个***，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。
所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、ROM、RAM、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (21)

  1. 一种运动轨迹规划的方法，其特征在于，包括：
    获取第一智能驾驶设备的规划行驶路径和第一运动参数，所述第一运动参数包括所述第一智能驾驶设备的速度和/或加速度；
    获取第一物体的预测运动轨迹和第二运动参数，所述第二运动参数包括所述第一物体的速度和/或加速度；
    在所述规划行驶路径和所述预测运动轨迹存在交叉时，根据所述第一运动参数和所述第二运动参数确定所述第一智能驾驶设备的规划运动轨迹，所述规划运动轨迹包括所述第一智能驾驶设备以第一预设速度抢行或让行所述第一物体的运动轨迹；
    控制所述第一智能驾驶设备按照所述规划运动轨迹行驶。
  2. 如权利要求1所述的方法，其特征在于，所述根据所述第一运动参数和所述第二运动参数确定所述第一智能驾驶设备的规划运动轨迹，包括：
    根据第一采样加速度和所述第一运动参数确定所述第一智能驾驶设备的第一推演轨迹，所述第一采样加速度为在采样空间中确定的所述第一智能驾驶设备可能达到的加速度；
    根据第二采样加速度和所述第二运动参数确定所述第一物体的第二推演轨迹，所述第二采样加速度为在所述采样空间中确定的所述第一物体可能达到的加速度；
    在所述第一推演轨迹和所述第二推演轨迹指示所述第一智能驾驶设备和所述第一物体不会发生碰撞，或者所述第一物体在所述第一物体的行驶方向上与所述第一智能驾驶设备的侧围或尾部在第一时刻发生碰撞时，根据所述第一推演轨迹和所述第二推演轨迹确定所述规划运动轨迹，其中，所述第一时刻为当前时刻之后的时刻。
  3. 如权利要求2所述的方法，其特征在于，所述根据第一采样加速度和所述第一运动参数确定所述第一智能驾驶设备的第一推演轨迹，包括：
    根据所述第一采样加速度，以及所述第一智能驾驶设备的速度和/或加速度确定第一子推演轨迹，所述第一子推演轨迹的终点处所述第一智能驾驶设备的速度为所述第一预设速度；
    根据所述第一预设速度确定第二子推演轨迹，所述第一子推演轨迹的终点为所述第二子推演轨迹的起点；
    根据所述第一子推演轨迹和所述第二子推演轨迹，确定所述第一推演轨迹。
  4. 如权利要求2或3所述的方法，其特征在于，所述根据所述第一推演轨迹和所述第二推演轨迹确定所述规划运动轨迹，包括：
    根据所述第一推演轨迹和所述第二推演轨迹确定第一博弈策略，所述第一博弈策略用于指示所述第一智能驾驶设备以所述第一预设速度抢行或让行所述第一物体；
    在所述第一博弈策略持续的时长大于第一时间阈值时，根据所述第一博弈策略、所述规划行驶路径和所述第二推演轨迹确定所述规划运动轨迹。
  5. 如权利要求4所述的方法，其特征在于，所述根据所述第一博弈策略、所述规划行驶路径和所述第二推演轨迹确定所述规划运动轨迹，包括：
    根据所述规划行驶路径和所述第二推演轨迹确定所述第一物体的第一位置空间占据 所述第一智能驾驶设备的第二位置空间时的时间段；
    基于所述第一博弈策略，根据所述第二位置空间和所述时间段确定所述规划运动轨迹，所述规划运动轨迹包括所述第一智能驾驶设备在所述第二位置空间以所述第一预设速度行驶的轨迹，或者，所述规划运动轨迹包括所述第一智能驾驶设备在所述时间段内以所述第一预设速度行驶的轨迹。
  6. 如权利要求4所述的方法，其特征在于，所述根据所述第一推演轨迹和所述第二推演轨迹确定所述第一博弈策略之前，所述方法还包括：
    确定所述第一推演轨迹和所述第二推演轨迹组成的推演轨迹对的策略代价最小。
  7. 如权利要求2至6中任一项所述的方法，其特征在于，所述根据第一采样加速度和所述第一运动参数确定所述第一智能驾驶设备的第一推演轨迹，包括：
    根据所述第一智能驾驶设备的横向偏移、所述第一采样加速度和所述第一运动参数确定所述第一推演轨迹，所述横向偏移为垂直于所述第一智能驾驶设备行驶方向上的偏移。
  8. 如权利要求1至7中任一项所述的方法，其特征在于，所述获取第一物体的预测运动轨迹和第二运动参数之前，所述方法还包括：
    确定所述第一物体与所述第一智能驾驶设备之间的距离小于或等于第一距离阈值，且大于或等于第二距离阈值。
  9. 如权利要求1至8中任一项所述的方法，其特征在于，所述第一预设速度大于或等于3公里每小时，且小于或等于15公里每小时。
  10. 一种运动轨迹规划的装置，其特征在于，包括获取单元和处理单元：
    所述获取单元，用于：
    获取第一智能驾驶设备的规划行驶路径和第一运动参数，所述第一运动参数包括所述第一智能驾驶设备的速度和/或加速度；以及获取第一物体的预测运动轨迹和第二运动参数，所述第二运动参数包括所述第一物体的速度和/或加速度；
    所述处理单元，用于：
    在所述规划行驶路径和所述预测运动轨迹存在交叉时，根据所述第一运动参数和所述第二运动参数确定所述第一智能驾驶设备的规划运动轨迹，所述规划运动轨迹包括所述第一智能驾驶设备以第一预设速度抢行或让行所述第一物体的运动轨迹；以及
    控制所述第一智能驾驶设备按照所述规划运动轨迹行驶。
  11. 如权利要求10所述的装置，其特征在于，所述处理单元用于：
    根据第一采样加速度和所述第一运动参数确定所述第一智能驾驶设备的第一推演轨迹，所述第一采样加速度为在采样空间中确定的所述第一智能驾驶设备可能达到的加速度；
    根据第二采样加速度和所述第二运动参数确定所述第一物体的第二推演轨迹，所述第二采样加速度为在所述采样空间中确定的所述第一物体可能达到的加速度；
    在所述第一推演轨迹和所述第二推演轨迹指示所述第一智能驾驶设备和所述第一物体未发生碰撞，或者所述第一物体在所述第一物体的行驶方向上与所述第一智能驾驶设备的侧围或尾部在第一时刻发生碰撞时，根据所述第一推演轨迹和所述第二推演轨迹确定所述规划运动轨迹，其中，所述第一时刻为当前时刻之后的时刻。
  12. 如权利要求11所述的装置，其特征在于，所述第一运动参数包括所述第一智能驾驶设备的速度和/或加速度，所述处理单元用于：
    根据所述第一采样加速度，以及所述第一智能驾驶设备的速度和/或加速度确定第一子推演轨迹，所述第一子推演轨迹的终点处所述第一智能驾驶设备的速度为所述第一预设速度；
    根据所述第一预设速度确定第二子推演轨迹，所述第一子推演轨迹的终点为所述第二子推演轨迹的起点；
    根据所述第一子推演轨迹和所述第二子推演轨迹，确定所述第一推演轨迹。
  13. 如权利要求11或12所述的装置，其特征在于，所述处理单元用于：
    根据所述第一推演轨迹和所述第二推演轨迹确定第一博弈策略，所述第一博弈策略用于指示所述第一智能驾驶设备以所述第一预设速度抢行或让行所述第一物体；
    在所述第一博弈策略持续的时长大于第一时间阈值时，根据所述第一博弈策略、所述规划行驶路径和所述第二推演轨迹确定所述规划运动轨迹。
  14. 如权利要求13所述的装置，其特征在于，所述处理单元用于：
    根据所述规划行驶路径和所述第二推演轨迹确定所述第一物体的第一位置空间占据所述第一智能驾驶设备的第二位置空间时的时间段；
    基于所述第一博弈策略，根据所述第二位置空间和所述时间段确定所述规划运动轨迹，所述规划运动轨迹包括所述第一智能驾驶设备在所述第二位置空间以所述第一预设速度行驶的轨迹，或者，所述规划运动轨迹包括所述第一智能驾驶设备在所述时间段内以所述第一预设速度行驶的轨迹。
  15. 如权利要求13所述的装置，其特征在于，所述处理单元还用于：
    确定所述第一推演轨迹和所述第二推演轨迹组成的推演轨迹对的策略代价最小。
  16. 如权利要求11至15中任一项所述的装置，其特征在于，所述处理单元用于：
    根据所述第一智能驾驶设备的横向偏移、所述第一采样加速度和所述第一运动参数确定所述第一推演轨迹，所述横向偏移为垂直于所述第一智能驾驶设备行驶方向上的偏移。
  17. 如权利要求10至16中任一项所述的装置，其特征在于，所述处理单元还用于：
    确定所述第一物体与所述第一智能驾驶设备之间的距离小于或等于第一距离阈值，且大于或等于第二距离阈值。
  18. 如权利要求10至17中任一项所述的装置，其特征在于，所述第一预设速度大于或等于3公里每小时，且小于或等于15公里每小时。
  19. 一种运动轨迹规划的装置，其特征在于，包括：
    存储器，用于存储计算机程序；
    处理器，用于执行所述存储器中存储的计算机程序，以使得所述装置执行如权利要求1至9中任一项所述的方法。
  20. 一种智能驾驶设备，其特征在于，包括权利要求10至19中任一项所述的装置。
  21. 一种芯片，其特征在于，所述芯片包括处理器与数据接口，所述处理器通过所述数据接口读取存储器上存储的指令，以执行如权利要求1至9中任一项所述的方法。
PCT/CN2022/113426 2022-08-18 2022-08-18 运动轨迹规划的方法、装置以及智能驾驶设备 WO2024036580A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2022/113426 WO2024036580A1 (zh) 2022-08-18 2022-08-18 运动轨迹规划的方法、装置以及智能驾驶设备

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2022/113426 WO2024036580A1 (zh) 2022-08-18 2022-08-18 运动轨迹规划的方法、装置以及智能驾驶设备

Publications (1)

Publication Number Publication Date
WO2024036580A1 true WO2024036580A1 (zh) 2024-02-22

Family

ID=89940415

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/113426 WO2024036580A1 (zh) 2022-08-18 2022-08-18 运动轨迹规划的方法、装置以及智能驾驶设备

Country Status (1)

Country Link
WO (1) WO2024036580A1 (zh)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111775961A (zh) * 2020-06-29 2020-10-16 北京百度网讯科技有限公司 自动驾驶车辆规划方法、装置、电子设备及存储介质
CN111845774A (zh) * 2020-07-20 2020-10-30 上海大学 一种基于横纵向协调的自动驾驶汽车动态轨迹规划及跟踪方法
CN112002144A (zh) * 2020-09-02 2020-11-27 中国科学技术大学 无信号交叉路口行车风险态势评估方法及***
WO2021005392A1 (ja) * 2019-07-05 2021-01-14 日産自動車株式会社 運転制御方法及び運転制御装置
CN112373485A (zh) * 2020-11-03 2021-02-19 南京航空航天大学 一种考虑交互博弈的自动驾驶车辆决策规划方法
CN114475585A (zh) * 2022-03-11 2022-05-13 阿波罗智能技术(北京)有限公司 路口自动驾驶方法、装置、电子设备及自动驾驶车辆

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021005392A1 (ja) * 2019-07-05 2021-01-14 日産自動車株式会社 運転制御方法及び運転制御装置
CN111775961A (zh) * 2020-06-29 2020-10-16 北京百度网讯科技有限公司 自动驾驶车辆规划方法、装置、电子设备及存储介质
CN111845774A (zh) * 2020-07-20 2020-10-30 上海大学 一种基于横纵向协调的自动驾驶汽车动态轨迹规划及跟踪方法
CN112002144A (zh) * 2020-09-02 2020-11-27 中国科学技术大学 无信号交叉路口行车风险态势评估方法及***
CN112373485A (zh) * 2020-11-03 2021-02-19 南京航空航天大学 一种考虑交互博弈的自动驾驶车辆决策规划方法
CN114475585A (zh) * 2022-03-11 2022-05-13 阿波罗智能技术(北京)有限公司 路口自动驾驶方法、装置、电子设备及自动驾驶车辆

Similar Documents

Publication Publication Date Title
US11572099B2 (en) Merge behavior systems and methods for merging vehicles
EP3934955B1 (en) Autonomous vehicle maneuvering based upon risk associated with occluded regions
US11117584B2 (en) Merge behavior systems and methods for mainline vehicles
CN105151043B (zh) 一种无人驾驶汽车紧急避让的方法
WO2021030476A1 (en) Cost-based path determination
CN114270360A (zh) 让步行为建模和预测
JP7234103B2 (ja) 車両制御装置、および、車両制御システム
CN110473416B (zh) 车辆控制装置
US20200307564A1 (en) Collision behavior recognition and avoidance
EP3911548B1 (en) Vehicle control
US10522041B2 (en) Display device control method and display device
US11801864B1 (en) Cost-based action determination
WO2022125568A1 (en) Velocity-based relevance filter
US20200226927A1 (en) Travel control device, travel control method, and storage medium storing program
US20220204056A1 (en) Permeable Speed Constraints
CN114651286A (zh) 避障动作
JP6632581B2 (ja) 走行制御装置、走行制御方法およびプログラム
CN114516329A (zh) 车辆自适应巡航控制***、方法和计算机可读介质
CN115593429A (zh) 自动驾驶车辆对紧急车辆的响应
CN111376901B (zh) 车辆的控制方法、装置及车辆
WO2024036580A1 (zh) 运动轨迹规划的方法、装置以及智能驾驶设备
JP7446216B2 (ja) 車両制御装置
EP4244692A1 (en) Optimization of performance in automotive autonomous driving of recurrent low speed manoeuvres in digital road maps-free areas
CN114056328B (zh) 基于dqn实现车辆l3级自动驾驶的行人避让方法及***
CN114269618A (zh) 转弯前用于并入的车道处理

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22955360

Country of ref document: EP

Kind code of ref document: A1