WO2023273067A1 - 一种行驶规划方法、装置、计算机设备及存储介质 - Google Patents

一种行驶规划方法、装置、计算机设备及存储介质 Download PDF

Info

Publication number
WO2023273067A1
WO2023273067A1 PCT/CN2021/127434 CN2021127434W WO2023273067A1 WO 2023273067 A1 WO2023273067 A1 WO 2023273067A1 CN 2021127434 W CN2021127434 W CN 2021127434W WO 2023273067 A1 WO2023273067 A1 WO 2023273067A1
Authority
WO
WIPO (PCT)
Prior art keywords
motion state
longitudinal
lateral
relationship information
driving
Prior art date
Application number
PCT/CN2021/127434
Other languages
English (en)
French (fr)
Inventor
周程杨
万登科
Original Assignee
上海商汤临港智能科技有限公司
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 上海商汤临港智能科技有限公司 filed Critical 上海商汤临港智能科技有限公司
Publication of WO2023273067A1 publication Critical patent/WO2023273067A1/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
    • B60W60/00—Drive control systems specially adapted for autonomous road vehicles
    • B60W60/001—Planning or execution of driving tasks
    • 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
    • B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
    • B60W40/10—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to vehicle motion
    • 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
    • B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
    • 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
    • B60W60/00—Drive control systems specially adapted for autonomous road vehicles
    • B60W60/001—Planning or execution of driving tasks
    • B60W60/0011—Planning or execution of driving tasks involving control alternatives for a single driving scenario, e.g. planning several paths to avoid obstacles
    • 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
    • B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
    • B60W2050/0001—Details of the control system
    • B60W2050/0019—Control system elements or transfer functions
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Definitions

  • the present disclosure relates to the technical field of automatic driving, and in particular to a driving planning method, device, computer equipment and storage medium.
  • Embodiments of the present disclosure at least provide a travel planning method, device, computer equipment, and storage medium.
  • the embodiment of the present disclosure provides a driving planning method, including: based on the reference trajectory information of the target road and the vehicle parameters of the self-driving vehicle, determining the relationship between the motion state parameters of the self-driving vehicle in the current control cycle information; based on the motion state parameter relationship information, determine the lateral motion state parameter relationship information and the longitudinal motion state parameter relationship information of the self-driving vehicle; based on preset constraints, the lateral motion state parameter relationship information and the longitudinal motion
  • the state parameter relationship information determines the target motion state of the automatic driving vehicle in at least one control cycle in the future.
  • an embodiment of the present disclosure further provides a driving planning device, including: a first determination module configured to determine the current control of the automatic driving vehicle based on the reference trajectory information of the target road and the vehicle parameters of the automatic driving vehicle.
  • the second determination module is configured to determine the lateral motion state parameter relationship information and the longitudinal motion state parameter relationship information of the self-driving vehicle based on the motion state parameter relationship information;
  • the third determination module It is configured to determine the target motion state of the self-driving vehicle in at least one control cycle in the future based on preset constraint conditions, the lateral motion state parameter relationship information, and the longitudinal motion state parameter relationship information.
  • an optional implementation manner of the present disclosure further provides a computer device, including a processor and a memory, the memory stores machine-readable instructions executable by the processor, and the processor is configured to execute the Stored machine-readable instructions, when the machine-readable instructions are executed by the processor, when the machine-readable instructions are executed by the processor, the above-mentioned first aspect, or any possible one of the first aspect steps in the implementation.
  • an optional implementation manner of the present disclosure further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed, the above-mentioned first aspect, or any one of the first aspects in the first aspect, may be executed. Steps in one possible implementation.
  • an optional implementation manner of the present disclosure further provides a computer program product, where the computer program product includes computer-executable instructions. After the computer-executable instructions are executed, the above-mentioned first aspect, or the first aspect in the first aspect can be realized. A step in any possible implementation.
  • FIG. 1 shows a flowchart of a driving planning method provided by an embodiment of the present disclosure
  • FIG. 2 shows a schematic diagram of a Fleiner coordinate system provided by an embodiment of the present disclosure
  • FIG. 3 shows a schematic diagram of an S-T coordinate system determined by using the first change relationship information between the longitudinal displacement of an autonomous vehicle and time provided by an embodiment of the present disclosure
  • FIG. 4 shows a schematic diagram of a D-T coordinate system determined by using the first change relationship information between the longitudinal displacement of the autonomous vehicle and time provided by an embodiment of the present disclosure
  • Fig. 5 shows a schematic diagram of a travel planning device provided by an embodiment of the present disclosure
  • Fig. 6 shows a schematic diagram of a computer device provided by an embodiment of the present disclosure.
  • the quadratic programming method can only solve the planning of the primary safety of the self-driving vehicle (that is, complete basic operations such as avoiding vehicles), but cannot ensure stability in a timely manner (such as When avoiding a vehicle, in the control cycle, the time can only allow to determine the planned path to ensure the avoidance operation, but cannot avoid the occurrence of sharp turns and sudden advances), which makes the user less comfortable when riding in an automatic driving vehicle.
  • the embodiments of the present disclosure provide a driving planning method, which uses the reference trajectory information of the target road and the vehicle parameters of the self-driving vehicle to determine the relationship information of the motion parameters of the self-driving vehicle in the current control cycle, and thus obtains The motion state parameter relationship information corresponding to the horizontal and vertical directions of the self-driving vehicle, so as to determine the target motion state of the self-driving vehicle in at least one control cycle in the future based on the preset constraints, and use the motion state parameter relationship information that can be obtained more accurately Carry out actual driving planning, so as to be able to control the autonomous vehicle more accurately.
  • constraints that can ensure stability can be used to determine a more stable driving plan to improve the comfort of users when driving or riding in autonomous vehicles .
  • the execution subject of the driving planning method provided in the embodiment of the present disclosure is generally an automatic driving control device.
  • the automatic driving control device is installed in the automatic driving vehicle, and can plan the process of automatic driving based on the driving planning method provided by the embodiment of the present disclosure.
  • the driving planning method may be implemented by calling a computer-readable instruction stored in a memory by a processor.
  • FIG. 1 it is a flowchart of a driving planning method provided by an embodiment of the present disclosure.
  • the method includes:
  • S101 Based on the reference trajectory information of the target road and the vehicle parameters of the self-driving vehicle, determine the motion state parameter relationship information of the self-driving vehicle in the current control cycle;
  • S103 Determine a target motion state of the autonomous vehicle in at least one control period in the future based on preset constraint conditions, the lateral motion state parameter relationship information, and the longitudinal motion state parameter relationship information.
  • the embodiments of the present disclosure determine the motion state parameter relation information of the self-driving vehicle in the current control cycle based on the reference trajectory information of the target road and the vehicle parameters of the self-driving vehicle, and then according to the motion parameter relation information of the self-driving vehicle determined using The lateral motion state parameter relationship information, the longitudinal motion state parameter relationship information, and the preset constraint conditions determine the target motion state of the autonomous vehicle in at least one control cycle in the future.
  • this method can more accurately plan the actual driving of the self-driving vehicle, thereby enabling more accurate control of the self-driving vehicle.
  • it can also provide more stable driving planning to improve the comfort of users when riding in self-driving vehicles.
  • autonomous vehicles may include, for example, intelligent driving vehicles such as fully automatic unmanned vehicles and semi-automatic unmanned vehicles, or intelligent robots such as wheeled mobile robots and crawler mobile robots.
  • intelligent driving vehicles such as fully automatic unmanned vehicles and semi-automatic unmanned vehicles
  • intelligent robots such as wheeled mobile robots and crawler mobile robots.
  • the self-driving vehicles are different, and the target roads on which the self-driving vehicles are driving are also different.
  • the self-driving vehicle may include a fully automatic or semi-automatic unmanned vehicle.
  • the corresponding target road may include, for example, driving lanes in highways and bridge roads.
  • self-driving vehicles can include wheeled mobile robots or tracked robots.
  • the corresponding target road may include, for example, a robot driving lane reserved between shelves.
  • the self-driving vehicle is an intelligent driving vehicle, and it is described as an example when driving in a road driving scene.
  • the self-driving vehicle When the self-driving vehicle is traveling on the target road, it mainly conducts periodic detection of the target road where the self-driving vehicle is located in the control cycle, and then conducts driving planning based on the detection results to determine the position, speed, and acceleration at the next moment. and other motion state parameters. According to these motion state parameters, the corresponding moment, the torque output by the power output device of the self-driving vehicle, and the angle of driving rotation can be obtained by reverse solution, so as to control the progress of the self-driving vehicle.
  • the time of a control cycle can be set to 2s, or in order to ensure the safety of the self-driving vehicle, the time of a control cycle can be shortened to 1s or 1.5s or even more Small, the duration of the control period is not limited in this embodiment.
  • the reference trajectory information of the target road includes a plurality of position points on the centerline of the target road, and coordinate values of the plurality of position points in a Cartesian coordinate system.
  • the reference trajectory information can be obtained by using other devices mounted on the self-driving vehicle, such as a high-precision map module.
  • the high-precision map module may include a depth camera device, for example.
  • the high-precision map module can use at least one of the following solutions to collect images of the target road: structured light (Structured-light), binocular vision (Stereo), and optical time of flight (Time of Flight, TOF). After the high-precision map module collects the road image of the target road, it can also detect the road image to determine the centerline in the road image and determine multiple position points on the centerline.
  • the multiple position points on the center line may include, for example, position points corresponding to the two ends of the center line, so as to determine the position of the center line relatively simply; or, may include position points corresponding to the two ends of the center line, and The image shows at least one point between two endpoints of the centerline to more accurately characterize the location of the centerline.
  • the coordinate values of the multiple position points in the Cartesian coordinate system can also be obtained.
  • the determined multiple location points can be obtained by using the high-precision map module mounted on the self-driving vehicle, it is possible to first determine that the multiple location points are in the image coordinate system corresponding to the road image collected by the high-precision map module position information, and then use the coordinate system conversion relationship between the image coordinate system and the scene coordinate system corresponding to the target road to determine the position information of multiple position points in the scene coordinate system.
  • the scene coordinate system of the target road can be established based on the location of the self-driving vehicle as the coordinate origin.
  • the coordinate system is the Cartesian coordinate system determined for the target road. Then, using the established Cartesian coordinate system, the coordinate values corresponding to the multiple positions on the center line are determined respectively.
  • vehicle parameters of the autonomous vehicle can also be determined.
  • the vehicle parameters of the self-driving vehicle may include power performance parameters and vehicle body parameters.
  • the power performance parameters may include, for example, the full load of the vehicle, the maximum torque of the engine, the acceleration of starting and shifting, the braking performance and other parameters that can directly control the running of the vehicle.
  • the vehicle body parameters may include, for example, the length of the vehicle body, the width of the vehicle body, etc., which can determine the space size of the area occupied by the vehicle.
  • the motion state parameter relationship information of the self-driving vehicle in the current control cycle can be determined.
  • the motion state parameter relationship information of the self-driving vehicle in the current control cycle can be determined in the following manner: based on the reference trajectory information of the target road and the current position of the self-driving vehicle, the Fleiner coordinates are established Frame (Frenet-Serret frame); in the Frenet coordinate system, based on the vehicle parameters of the self-driving vehicle, determine the relationship information of the motion state parameters of the self-driving vehicle in the current control cycle.
  • the motion state parameter relationship information may be represented by a relationship formula between variable acceleration and position, speed, and acceleration, for example, expressed by a kinematic equation.
  • the driving planning of autonomous vehicles is a high-dimensional optimization problem with multiple nonlinear constraints, the amount of data to be processed is relatively large, and the scene is relatively complex. At the same time, for safety reasons, it is necessary to ensure real-time . Therefore, when planning the driving of the automatic driving vehicle, the Fleiner coordinate system is selected to adapt to the route planning of the automatic driving curve while adapting to the straight road planning. At the same time, using the Fleiner coordinate system can also reduce the dimensionality of map data, reduce the amount of calculation, and improve efficiency to meet the real-time requirements of driving planning.
  • the coordinate origin of the Fleiner coordinate system may be determined according to the current position of the autonomous vehicle, and then the Fleiner coordinate system may be determined according to the determined reference trajectory of the target road.
  • the determined coordinate origin can be used as the tangent point, and the tangent line can be determined along the direction of the reference line as the S-axis of the Freiner coordinate system, and the normal line perpendicular to the determined tangent line at the tangent point can be determined as the Freiner The D axis of the coordinate system.
  • the position of the self-driving vehicle can be determined simply by using the longitudinal distance (that is, the distance along the centerline) and the lateral distance (that is, the distance away from the centerline) .
  • the calculation of the speed, acceleration, and variable acceleration of the self-driving vehicle in the longitudinal and lateral directions is also simpler.
  • FIG. 2 it is a schematic diagram of a Fryner coordinate system provided by an embodiment of the present disclosure; it includes an automatic driving vehicle 21, a driving road 22, a centerline 23 in the driving road 22, and the corresponding Fryner coordinate system. Nano coordinate system 24 .
  • the motion state parameters can include lateral motion state parameters and longitudinal motion state parameter information.
  • the longitudinal motion state parameter includes: at least one of longitudinal position, longitudinal velocity, longitudinal acceleration, and longitudinal variable acceleration;
  • the lateral motion state parameter includes: at least one of lateral position, lateral angle, lateral angular velocity, and lateral angular acceleration.
  • the motion state parameter relationship information represents the relationship between each motion state parameter, the lateral motion state parameter relationship information represents the relationship between each lateral motion state parameter, and the longitudinal motion state parameter relationship information represents each longitudinal motion state parameter The relationship between.
  • the motion state parameter relation information of the automatic driving vehicle in the current control cycle can be determined according to at least one of the following models, for example: a bicycle model (Bicycle Model) and a four-wheel model.
  • a bicycle model (Bicycle Model) and a four-wheel model.
  • a single-vehicle model it can also be subdivided into a vehicle motion model with the rear axle as the origin, a vehicle kinematics model centered on the center of mass, and an Ackerman Turning Geometry model.
  • the core part is to plan the starting point and the position, speed, and acceleration of the ending point in a control cycle, so when using the model to determine the
  • the relationship information of the motion state parameters in the variable acceleration is used as the input state quantity
  • the relationship information corresponding to at least one of the output control quantity, including position, velocity and acceleration can be determined.
  • the obtained motion state parameter relationship information includes both the motion state parameters of the self-driving vehicle in the longitudinal direction
  • the relevant information also includes the relevant information of the motion state parameters of the self-driving vehicle in the longitudinal direction.
  • the motion state parameter relationship information can be decoupled horizontally and vertically to obtain the horizontal motion state parameter relationship information and/or the vertical motion state parameter relationship information.
  • the following method when decoupling the relationship information of motion state parameters horizontally and vertically, for example, the following method can be adopted: After setting the value of the horizontal state parameter to a preset value, based on the motion state parameter relationship information , to obtain the longitudinal motion state parameter relationship information; based on the longitudinal motion state parameter relationship information and the motion state parameter relationship information, the lateral motion state parameter relationship information is obtained.
  • a value that can characterize the lateral state such as lateral angular acceleration
  • a preset value such as 0, to eliminate the possible impact of lateral movement on the longitudinal direction.
  • the value 0 of the lateral angular acceleration is brought into the state parameter relationship information to obtain the longitudinal motion state parameter relationship information.
  • the relevant information of the motion state parameters in the longitudinal direction of the self-driving vehicle can be determined, and the obtained relative information of the motion state parameters in the longitudinal direction and the motion state parameter information can be used to determine the motion in the lateral direction.
  • the relevant information of the state parameters completes the decoupling of the state parameter information of the lateral and vertical motion of the autonomous vehicle.
  • the optimal longitudinal motion state can be obtained more conveniently and quickly; similarly, since part of the lateral motion state of the autonomous vehicle is related to the longitudinal motion state, the obtained optimal longitudinal motion state State, and lateral motion state parameter relationship information, can obtain better lateral motion state more conveniently and quickly.
  • the preset constraint conditions may include, for example, horizontal constraint conditions and/or vertical constraint conditions.
  • the longitudinal constraint conditions may include, for example, the first change relationship information between the longitudinal displacement of the self-driving vehicle and time; wherein, the first change relationship information is used to characterize the future Longitudinal displacement boundaries at multiple moments.
  • the longitudinal constraints further include at least one of the following: information about the relationship between the longitudinal displacement of the autonomous vehicle and time, the longitudinal speed threshold, the longitudinal displacement change threshold corresponding to the adjacent control cycle, the adjacent control cycle The corresponding longitudinal speed change threshold, the acceleration change threshold corresponding to the adjacent control period, and the longitudinal acceleration change threshold corresponding to the adjacent control period.
  • (1a) the first change relationship information between the longitudinal displacement of the autonomous vehicle and time.
  • the first change relationship information may be determined in the following manner: obtaining obstacle trajectory information of the obstacle in the Cartesian coordinate system; projecting the obstacle trajectory information into the Freiner coordinate system; based on the obtained The projection result of the obstacle trajectory information in the Fleiner coordinate system is used to determine the first change relationship information between the longitudinal displacement and time of the self-driving vehicle.
  • using the Fleiner coordinate system can easily represent the information of the target road, and can also reduce the dimensionality of the map data, reduce the amount of calculation, and improve efficiency to meet the real-time requirements of driving planning.
  • the predicted trajectory information of the obstacle may be, for example, the trajectory information output by the prediction module mounted on the automatic driving vehicle or other perception modules.
  • the prediction module may include, for example, a laser radar device capable of obtaining obstacle track information in real time.
  • the obtained predicted trajectory of the obstacle is similar to the coordinate values in the Cartesian coordinate system of the multiple position points determined in S101 above, and they are all coordinate values in the Cartesian coordinate system. Then, using the coordinate system conversion relationship between the image coordinate system used in the above S101 and the scene coordinate system corresponding to the target road, the obstacle trajectory information can be projected in the Freiner coordinate system, and based on the obstacle trajectory The projection result of the information in the Fleiner coordinate system determines the first change relationship information between the longitudinal displacement and time of the self-driving vehicle.
  • FIG. 3 it is a schematic diagram of an S-T coordinate system determined by using the first change relationship information between the longitudinal displacement of an autonomous vehicle and time provided by an embodiment of the present disclosure.
  • the abscissa 31 of the S-T coordinate system represents time t;
  • the ordinate 32 of the S-T coordinate system represents the longitudinal displacement distance s of the self-driving vehicle.
  • the trajectory 33 of vehicle A limits the maximum distance that the self-driving vehicle can travel forward at any time during this period to s1. If the forward distance of the self-driving vehicle exceeds s1 , there is a greater possibility of colliding with vehicle A.
  • the driving trajectory 33 of vehicle A changes to the driving trajectory 34, and there is a driving behavior of accelerating away from the self-driving vehicle. Therefore, the maximum distance that the self-driving vehicle can travel forward, with the rapid Driving away, the maximum travelable distance of the self-driving vehicle from t1 to t2 changes to s2.
  • vehicle B overtakes the self-driving vehicle.
  • the forward driving distance of the self-driving vehicle at time t2 should not be less than s3, otherwise it will collide with vehicle B.
  • the driving distance between vehicle A and the self-driving vehicle is far away, it does not pose a safety threat.
  • the route of the self-driving vehicle since in the longitudinal direction, the route of the self-driving vehicle is generally forward, the projected trajectory of the vehicle only appears on the positive semi-axis of the s-axis.
  • the driving route of the self-driving vehicle is in the backward direction, such as in the scene of reversing, when there is an obstacle approaching the direction of the vehicle, emergency braking is required, so it is not included in this example.
  • the longitudinal speed threshold may only include the upper limit of the speed, for example, the maximum speed shall not exceed 80 km/h; or only the lower limit of the speed, for example, the minimum speed shall not be lower than 5 km/h; Include lower bounds for speed, such as a speed limit between 5 km/h and 80 km/h. Specifically, it can be formulated according to the actual road speed limit rules.
  • the prescribed speed limit is 30 km/h
  • the longitudinal speed threshold can be set to include an upper limit of 30 km/h
  • the minimum speed shall not be lower than 60 km/h and shall not exceed 120 km/h
  • the longitudinal speed threshold can be set to include an upper limit of 120 km/h and a lower limit of 60 km/h.
  • the longitudinal speed threshold when setting the longitudinal speed threshold, it can also be set based on a preset driving strategy under the premise of satisfying safety, so as to improve the comfort when using the self-driving vehicle.
  • a preset driving strategy under the premise of satisfying safety, so as to improve the comfort when using the self-driving vehicle.
  • the vehicle when driving on a high-speed road section, since the process of acceleration is required when entering the high-speed road section, and after a certain distance, such as 500 meters, the vehicle can enter the state of high-speed driving, so it can be driven at no less than Within the specified speed limit of 60 km/h and no more than 120 km/h, set a driving strategy of a smaller speed limit of no less than 80 km/h and no more than 100 km/h, so that the automatic driving vehicle Under the premise of ensuring safe driving, driving at a relatively uniform speed can keep the self-driving vehicle stable while driving, and there are fewer behaviors of the human body rushing or leaning back due to excessive speed changes, which helps to
  • the threshold value of the longitudinal displacement change corresponding to the adjacent control cycle it can be set that the self-driving vehicle maintains a relatively constant speed for a long period of time, and it can also ensure that when there is a vehicle that needs to avoid or overtake, the automatic driving The vehicle achieves a smoother gear change.
  • the longitudinal displacement variation corresponding to adjacent control cycles should remain unchanged; when the speed changes only to a small extent, the longitudinal displacement variation corresponding to adjacent control cycles There will be a small degree of change, so the threshold value of the longitudinal displacement change corresponding to the adjacent control cycle can be set according to the actual situation, for example, not exceeding 100 meters, or not exceeding 200 meters, so as to ensure that the self-driving vehicle can move at a relatively small speed. Drive under changing conditions to reduce sudden changes in speed.
  • the lateral constraint conditions may include, for example, the second change relationship information between the lateral displacement of the autonomous vehicle and time; wherein, the second transformation relationship information is used to characterize the Lateral displacement boundaries at multiple moments.
  • the lateral constraints also include at least one of the following: information about the second change relationship between the lateral displacement of the autonomous vehicle and time, the lateral displacement change threshold corresponding to adjacent control cycles, the lateral angle corresponding to adjacent control cycles The variation threshold, the lateral angular velocity variation threshold corresponding to the adjacent control period, and the lateral angular acceleration variation threshold corresponding to the adjacent control period.
  • the second change relationship information may be determined in the following manner: obtaining obstacle trajectory information of the obstacle in the Cartesian coordinate system; projecting the obstacle trajectory information into the Freiner coordinate system; based on the obtained The projection result of the obstacle trajectory information in the Fleiner coordinate system is used to determine the second change relationship information between the lateral displacement and time of the self-driving vehicle.
  • the manner of determining the second change relationship information is similar to the manner of determining the first change relationship information in (1a) above, and will not be repeated here.
  • FIG. 4 it is a schematic diagram of a D-T coordinate system determined by using the first change relationship information between the longitudinal displacement of the autonomous vehicle and time provided by an embodiment of the present disclosure.
  • the abscissa 41 of the D-t coordinate system represents the time t
  • the ordinate 42 of the D-T coordinate system represents the lateral displacement distance d of the self-driving vehicle.
  • the negative semi-axis of the D-axis represents the distance between the autonomous vehicle and the road edge on the right
  • the negative semi-axis of the D-axis represents the distance between the autonomous vehicle and the road edge on the left.
  • the self-driving vehicle is driving along the centerline of the target road. Therefore, in order to avoid vehicles that merge into the centerline in the left and right lanes, the self-driving vehicle can deviate to both sides along the centerline .
  • the distances from the road sidelines on the left and right sides of the self-driving vehicle to the center line are s1 and s2, respectively.
  • the vehicle C drives out, and the distance that the self-driving vehicle can drive to the right returns to the maximum distance s2.
  • the distance that can travel to the left is the maximum distance to the left s1.
  • the determination method of the lateral displacement change threshold corresponding to the adjacent control period is similar to the method of determining the longitudinal displacement change threshold corresponding to the adjacent control period in (1c) above, and will not be described again.
  • the relationship information of the lateral motion state parameters and the relationship information of the longitudinal motion state parameters may be used The following method: based on the longitudinal constraints, the longitudinal motion state parameter relationship information, and the longitudinal driving strategy, determine the longitudinal motion state of the self-driving vehicle in the future at least one control cycle; based on the lateral constraints, The lateral motion state parameter relationship information, the lateral driving strategy, and the lateral motion state of the self-driving vehicle in the at least one control cycle in the future, determine the lateral motion state of the self-driving vehicle in the at least one control cycle in the future .
  • the target motion states respectively corresponding to the automatic driving vehicle in at least one future control cycle include: the lateral motion state and the longitudinal motion state of the automatic driving vehicle in the at least one future control cycle.
  • the longitudinal variable acceleration and lateral variable acceleration respectively corresponding to at least one control cycle can be used as an operating parameter characterizing the lateral motion state and the longitudinal motion state of the self-driving vehicle in the at least one future control cycle, for example.
  • the longitudinal motion state parameter relationship information, and the driving strategy for example, the following may be adopted: Way: based on the longitudinal constraints, preset optimization step size, optimization time domain, road speed limit, vehicle performance parameters, and the longitudinal driving strategy, generate a first objective function; the first objective function is based on each The longitudinal motion state at the optimization time point and the first distance from the target longitudinal state are the optimization goals; based on the longitudinal motion state parameter relationship information and the longitudinal constraint conditions, at each optimization time point in the optimization time domain , optimize the optimal variable sequence for the first objective function, and obtain the first variable sequence composed of longitudinal vehicle state variables at multiple optimization time points; based on the first variable sequence, determine the automatic driving vehicle in the The longitudinal motion state for at least one control period in the future.
  • the preset optimal step size can be determined, for example, according to the optional step size of the automatic driving system, or according to the actual driving demand, specifically, it can include 0.1 second or 0.2 second, so as to ensure that the automatic driving vehicle can quickly respond to the changing driving environment. reaction.
  • an optimization time domain can also be set, and in a driving plan, the motion states corresponding to multiple optimization time points in the optimization time domain can be obtained. With the driving of the self-driving vehicle, the motion state can be optimized in real time according to the control cycle to improve the control accuracy.
  • the corresponding road speed limit is determined, so the road speed limit at the corresponding time can be determined correspondingly according to the actual road section; and for vehicle performance parameters Different self-driving vehicles have their corresponding vehicle parameters.
  • vehicle parameters of the self-driving vehicle in the embodiment corresponding to S101 above, which will not be repeated here.
  • the longitudinal form strategy may include, for example, the behavior strategy of the autonomous vehicle when driving longitudinally, for example, it may include behavior strategies such as overtaking and avoiding in the longitudinal direction.
  • the first objective function can be determined by using the above-mentioned relevant information corresponding to the vertical direction.
  • the first objective function can include the optimization time point as an independent variable, and use the above-mentioned relevant information corresponding to the vertical direction as the mapping relationship information, and the target vertical state The first distance of is used as the dependent variable, which is the optimization objective.
  • the first objective when optimizing the optimal variable sequence for the first objective function, for example, the first objective can be used.
  • the function establishes a longitudinal optimization model, and then uses the longitudinal constraints as the constraints of the longitudinal optimization model to determine the optimal one or more first variable sequences respectively corresponding to at least one control cycle in the future. Then, according to the first variable sequence, determine the longitudinal motion state of the automatic driving vehicle in the at least one control period in the future.
  • a penalty method can also be introduced. For example, by determining the amount of longitudinal displacement change in the longitudinal motion state corresponding to adjacent control cycles, the amount of longitudinal velocity change corresponding to adjacent control cycles, the amount of acceleration change corresponding to adjacent control cycles, and the longitudinal variable acceleration corresponding to adjacent control cycles Amount of variation, set penalty.
  • each change amount may be directly used as the penalty.
  • the training model outputs the longitudinal motion state in a direction with a small change amount.
  • the lateral motion state of the autonomous vehicle in at least one control cycle in the future since the lateral motion state of the autonomous vehicle will be affected by the longitudinal motion state, for example, when the longitudinal driving process is fast, if there is and When vehicles appear on the road, it is necessary to move laterally earlier to ensure safe avoidance and deflection in a more moderate manner. Therefore, when specifically determining the lateral motion state of the autonomous vehicle in at least one control cycle in the future, it is also necessary to consider the longitudinal motion state of the autonomous vehicle in at least one control cycle in the future, for example, the determined longitudinal motion state As a constraint when determining the lateral motion state.
  • the following method may be used: based on the lateral motion state, the lateral driving strategy, the predicted Set the optimization step size, optimization time domain, and the lateral constraints to generate a second objective function, the second objective function is optimized with the lateral motion state at each optimization time point and the second distance from the target lateral state Objective: Based on the relationship information of the lateral motion state parameters and the lateral constraint conditions, at each optimization time point in the optimization time domain, optimize the optimal variable sequence for the second objective function, and obtain a plurality of A second variable sequence composed of lateral vehicle state variables at the optimization time point; based on the second variable sequence, determine the lateral motion state of the autonomous vehicle in at least one control cycle in the future.
  • the horizontal and vertical state of the autonomous vehicle can be decoupled, and the horizontal and vertical planning can be more targeted, making the planned driving strategy more detailed, thereby ensuring a better driving plan for the autonomous vehicle.
  • the manner of determining the lateral motion state corresponding to the autonomous vehicle in at least one control cycle in the future is similar to the above-mentioned method of determining the longitudinal motion state corresponding to the autonomous vehicle in at least one control cycle in the future, and will not be repeated here.
  • the longitudinal movement state of the optimization time point can also be calculated according to multiple optimization time points , and the lateral motion state at the optimization time point are combined to generate an optimized trajectory; the optimized trajectory includes: a lateral motion trajectory and a longitudinal speed planning curve.
  • the horizontal motion state and the vertical motion state corresponding to each optimization time point have been determined, the horizontal motion state and the vertical motion state can be combined according to the same optimization time point, so as to obtain a line containing the horizontal and vertical positions, and Three-dimensionally optimized trajectory with safety and comfort in mind.
  • the self-driving vehicle after determining the respective longitudinal motion states of the self-driving vehicle in at least one control period in the future and determining the lateral movement states of the self-driving vehicle in at least one control period in the future, it further includes: using The target motion states respectively corresponding to at least part of the target control periods in at least one control period in the future determine the power parameters output by the self-driving vehicle in the target control period.
  • the power parameters output by the self-driving vehicle in the target control period can be obtained through calculation, so that the power parameters can be used to more directly control the driving of the self-driving vehicle.
  • the target control cycle may include, for example, at least one control cycle that can determine the longitudinal motion state and the lateral motion state.
  • the longitudinal motion state and the lateral motion state can be combined at the same time point in the control cycle in at least one control cycle in the future, so as to complete the driving planning of the automatic driving vehicle.
  • the motion state of the self-driving vehicle at any time in at least one control cycle in the future can be determined as the target motion state, and according to the dynamic parameters and driving behavior of the self-driving vehicle Corresponding relationship, determine the power parameters output in the target control cycle.
  • the writing order of each step does not mean a strict execution order and constitutes any limitation on the implementation process.
  • the specific execution order of each step should be based on its function and possible
  • the inner logic is OK.
  • the embodiment of the present disclosure also provides a driving planning device corresponding to the driving planning method. Since the problem-solving principle of the device in the embodiment of the present disclosure is similar to the above-mentioned driving planning method in the embodiment of the disclosure, the implementation of the device Reference can be made to the implementation of the method, and repeated descriptions will not be repeated.
  • FIG. 5 it is a schematic diagram of a travel planning device provided by an embodiment of the present disclosure.
  • the device includes: a first determination module 51 , a second determination module 52 and a third determination module 53 ; wherein,
  • the first determination module 51 is configured to determine the motion state parameter relationship information of the automatic driving vehicle in the current control cycle based on the reference trajectory information of the target road and the vehicle parameters of the automatic driving vehicle;
  • the second determination module 52 is configured to Based on the motion state parameter relationship information, determine the lateral motion state parameter relationship information and the longitudinal motion state parameter relationship information of the self-driving vehicle;
  • the third determination module 53 is configured to be based on pre-constraint conditions, the lateral motion state parameter relationship information and the longitudinal motion state parameter relationship information to determine the target motion state of the self-driving vehicle in at least one control cycle in the future.
  • the preset constraints include: horizontal constraints and/or vertical constraints.
  • the longitudinal constraint conditions include: first change relationship information between the longitudinal displacement of the self-driving vehicle and time; wherein the first change relationship information is used to characterize the self-driving vehicle The longitudinal displacement boundary at multiple times in the future.
  • the third determination module 53 is configured to determine the first change relationship information between the longitudinal displacement of the self-driving vehicle and time by using the following steps: Obtaining the information of the relationship between the obstacle in the Cartesian coordinate system Obstacle track information; projecting the obstacle track information to the Freiner coordinate system; determining the longitudinal displacement and time of the self-driving vehicle based on the projection result of the obstacle track information in the Freiner coordinate system The first change relation information of .
  • the lateral constraint conditions include: second variation relationship information between the lateral displacement of the autonomous vehicle and time; wherein the second transformation relationship information is used to characterize the automatic driving vehicle Lateral displacement boundaries at multiple times in the future.
  • the third determination module 53 is configured to determine the second change relationship information between the lateral displacement and time of the self-driving vehicle by adopting the following steps: obtaining the Obstacle track information; projecting the obstacle track information into the Freiner coordinate system; determining the lateral displacement and time of the autonomous vehicle based on the projection result of the obstacle track information in the Freiner coordinate system The second change relationship information of .
  • the longitudinal constraint conditions further include at least one of the following: a longitudinal velocity threshold, a longitudinal displacement change threshold corresponding to adjacent control periods, a longitudinal velocity change threshold corresponding to adjacent control periods, The acceleration change threshold corresponding to the adjacent control cycle, the longitudinal acceleration change threshold corresponding to the adjacent control cycle; the lateral constraint conditions also include at least one of the following: the lateral displacement change threshold of the adjacent control cycle, the adjacent The lateral angle variation threshold of the control cycle, the lateral angular velocity variation threshold of the adjacent control cycle, and the lateral angular acceleration variation threshold of the adjacent control cycle.
  • the target motion state of the autonomous vehicle in at least one control cycle in the future includes: the lateral motion state and the longitudinal motion state of the autonomous vehicle in the at least one control cycle in the future;
  • the third determination module 53 is configured to determine the longitudinal motion state of the self-driving vehicle in the future at least one control cycle based on the longitudinal constraint conditions, the longitudinal motion state parameter relationship information, and the longitudinal driving strategy;
  • the lateral constraints, the lateral motion state parameter relationship information, the lateral driving strategy, and the lateral motion state of the self-driving vehicle in the at least one control cycle in the future determine that the automatic driving vehicle will Controls the lateral motion state of the cycle.
  • the third determination module 53 is configured to: Strategy, generating a first objective function; the first objective function takes the longitudinal motion state at each optimization time point and the first distance from the target longitudinal state as the optimization goal; based on the longitudinal motion state parameter relationship information, and the longitudinal constraint conditions, at each optimization time point in the optimization time domain, optimize the optimal variable sequence for the first objective function, and obtain the first variable sequence composed of longitudinal vehicle state variables at multiple optimization time points; Based on the first sequence of variables, a longitudinal motion state of the autonomous vehicle in the at least one future control period is determined.
  • the third determination module 53 is configured to Conditions, generate a second objective function, the second objective function takes the lateral motion state at each optimization time point and the second distance from the target lateral state as the optimization target; based on the lateral motion state parameter relationship information, and the Transverse constraint conditions, at each optimization time point in the optimization time domain, optimize the optimal variable sequence for the second objective function to obtain a second variable sequence composed of lateral vehicle state variables at multiple optimization time points; Based on the second sequence of variables, a lateral motion state of the autonomous vehicle in the at least one future control cycle is determined.
  • the travel planning device further includes a trajectory generation module 54 configured to, according to a plurality of optimization time points, perform a longitudinal motion state at the optimization time point and a lateral motion state at the optimization time point combined to generate an optimized trajectory; the optimized trajectory includes: a lateral motion trajectory and a longitudinal speed planning curve.
  • the motion state parameters include: the lateral motion state parameters and the longitudinal motion state parameters;
  • the longitudinal motion state parameters include: at least one of: longitudinal position, longitudinal velocity, longitudinal acceleration, and longitudinal variable acceleration
  • the state parameters of lateral motion include: at least one of lateral position, lateral angle, lateral angular velocity, and lateral angular acceleration;
  • the relational information of the state parameters of motion represents the relationship between each state parameter of motion, and the state parameters of lateral motion
  • the relationship information characterizes the relationship between the various lateral motion state parameters
  • the longitudinal motion state parameter relationship information represents the relationship between the various longitudinal motion state parameters.
  • the reference track information includes: a plurality of position points on the centerline of the target road, and coordinate values of the plurality of position points in a Cartesian coordinate system.
  • the first determination module 51 is configured to establish a Fleiner coordinate system based on the reference trajectory information of the target road and the current position of the self-driving vehicle; In the Leiner coordinate system, based on the vehicle parameters of the automatic driving vehicle, the motion state parameter relationship information of the automatic driving vehicle in the current control cycle is determined.
  • the second determination module 52 is configured to decouple the motion state parameter relationship information horizontally and vertically to obtain the horizontal motion state parameter relationship information and/or the vertical motion state Parameter relationship information.
  • the second determination module 52 is configured to obtain the longitudinal motion state parameter based on the motion state parameter relationship information after setting the value of the lateral state parameter to a preset value.
  • Relational information obtaining the lateral motion state parameter relational information based on the longitudinal motion state parameter relational information and the motional state parameter relational information.
  • the target motion state of the at least one control period in the future includes: the longitudinal acceleration and the lateral acceleration of the at least one control period.
  • the travel planning device further includes a fourth determining module 55 configured to determine the automatic driving The power parameters output by the vehicle in the target control cycle.
  • FIG. 6 is a schematic structural diagram of the computer device provided by the embodiment of the present disclosure, including:
  • Processor 10 and memory 20 stores machine-readable instructions executable by the processor 10, the processor 10 is used to execute the machine-readable instructions stored in the memory 20, and the machine-readable instructions are executed by the processor 10 When executing, the processor 10 executes the steps of the aforementioned travel planning method in the embodiment of the present disclosure.
  • memory 20 comprises memory 210 and external memory 220;
  • Memory 210 here is also called internal memory, is used for temporarily storing the operation data in processor 10, and the data exchanged with external memory 220 such as hard disk, processor 10 communicates with memory 210 through memory 210.
  • the external memory 220 performs data exchange.
  • Embodiments of the present disclosure further provide a computer-readable storage medium, on which a computer program is stored, and when the computer program is run by a processor, the steps of the driving planning method described in the foregoing method embodiments are executed.
  • the storage medium may be a volatile or non-volatile computer-readable storage medium.
  • Embodiments of the present disclosure also provide a computer program product, the computer program product includes computer executable instructions, and after the computer executable instructions are executed, the steps of the driving planning method described in the above method embodiments can be implemented, specifically Refer to the foregoing method embodiments, and details are not repeated here.
  • the above-mentioned computer program product may be specifically implemented by means of hardware, software or a combination thereof.
  • the computer program product is embodied as a computer storage medium, and in another optional embodiment, the computer program product is embodied as a software product, such as a software development kit (Software Development Kit, SDK) etc. Wait.
  • the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Part 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 disclosure may be integrated into one processing unit, each unit may exist separately physically, or two or more units may be integrated into one unit.
  • the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium executable by a processor.
  • the technical solution of the present disclosure is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including Several instructions are used to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in various embodiments of the present disclosure.
  • the aforementioned storage media include: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic disk or optical disc and other media that can store program codes. .

Landscapes

  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
  • Traffic Control Systems (AREA)

Abstract

一种行驶规划方法、装置、计算机设备及存储介质，其中，该方法包括：基于目标道路的参考轨迹信息、以及自动驾驶车辆的车辆参数，确定自动驾驶车辆在当前控制周期中的运动状态参量关系信息（S101）；基于运动状态参量关系信息，确定自动驾驶车辆的横向运动状态参量关系信息、以及纵向运动状态参量关系信息（S102）；基于预设约束条件、横向运动状态参量关系信息以及纵向运动状态参量关系信息，确定自动驾驶车辆在未来至少一个控制周期的目标运动状态（S103）。这种方法可以更为准确地对自动驾驶车辆进行实际行驶规划，从而能够对自动驾驶车辆进行更为准确的控制。

Description

一种行驶规划方法、装置、计算机设备及存储介质
相关申请的交叉引用
本公开基于申请号为202110739069.6、申请日为2021年06月30日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此以引入方式并入本公开。
技术领域
本公开涉及自动驾驶技术领域，具体涉及一种行驶规划方法、装置、计算机设备及存储介质。
背景技术
随着人工智能(Artificial Intelligence，AI)的发展，自动驾驶汽车(Autonomous Vehicles)应运而生。自动驾驶汽车在实际生活中使用时，规划(Planning)和控制(Control)是自动驾驶最底层的部分，决定自动驾驶车辆在道路上如何行驶。
当前，在对自动驾驶车辆的行驶进行规划时，效率较低。
发明内容
本公开实施例至少提供一种行驶规划方法、装置、计算机设备及存储介质。
第一方面，本公开实施例提供了一种行驶规划方法，包括：基于目标道路的参考轨迹信息、以及自动驾驶车辆的车辆参数，确定所述自动驾驶车辆在当前控制周期中的运动状态参量关系信息；基于运动状态参量关系信息，确定所述自动驾驶车辆的横向运动状态参量关系信息、以及纵向运动状态参量关系信息；基于预设约束条件、所述横向运动状态参量关系信息以及所述纵向运动状态参量关系信息，确定所述自动驾驶车辆在未来至少一个控制周期的目标运动状态。
这样，可以更为准确地对自动驾驶车辆进行实际行驶规划，从而能够对自动驾驶车辆进行更为准确的控制。
第二方面，本公开实施例还提供一种行驶规划装置，包括：第一确定模块，配置为基于目标道路的参考轨迹信息、以及自动驾驶车辆的车辆参数，确定所述自动驾驶车辆在当前控制周期中的运动状态参量关系信息；第二确定模块，配置为基于运动状态参量关系信息，确定所述自动驾驶车辆的横向运动状态参量关系信息、以及纵向运动状态参 量关系信息；第三确定模块，配置为基于预设的约束条件、所述横向运动状态参量关系信息以及所述纵向运动状态参量关系信息，确定所述自动驾驶车辆在未来至少一个控制周期的目标运动状态。
第三方面，本公开可选实现方式还提供一种计算机设备，包括处理器和存储器，所述存储器存储有所述处理器可执行的机器可读指令，所述处理器用于执行所述存储器中存储的机器可读指令，所述机器可读指令被所述处理器执行时，所述机器可读指令被所述处理器执行时执行上述第一方面，或第一方面中任一种可能的实施方式中的步骤。
第四方面，本公开可选实现方式还提供一种计算机可读存储介质，该计算机可读存储介质上存储有计算机程序，该计算机程序被运行时执行上述第一方面，或第一方面中任一种可能的实施方式中的步骤。
第五方面，本公开可选实现方式还提供一种计算机程序产品，所述计算机程序产品包括计算机可执行指令，该计算机可执行指令被执行后，能够实现上述第一方面，或第一方面中任一种可能的实施方式中的步骤。
关于上述行驶规划装置、计算机设备、及计算机可读存储介质的效果描述参见上述行驶规划方法的说明，这里不再赘述。
为使本公开的上述目的、特征和优点能更明显易懂，下文特举实施例，并配合所附附图，作详细说明如下。
附图说明
为了更清楚地说明本公开实施例的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，此处的附图被并入说明书中并构成本说明书中的一部分，这些附图示出了符合本公开的实施例，并与说明书一起用于说明本公开的技术方案。应当理解，以下附图仅示出了本公开的某些实施例，因此不应被看作是对范围的限定，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他相关的附图。
图1示出了本公开实施例所提供的一种行驶规划方法的流程图；
图2示出了本公开实施例所提供的一种弗莱纳坐标系的示意图；
图3示出了本公开实施例所提供的一种利用自动驾驶车辆的纵向位移与时间的第一变化关系信息，确定的S-T坐标系的示意图；
图4示出了本公开实施例所提供的一种利用自动驾驶车辆的纵向位移与时间的第一变化关系信息，确定的D-T坐标系的示意图；
图5示出了本公开实施例所提供的一种行驶规划装置的示意图；
图6示出了本公开实施例所提供的一种计算机设备的示意图。
具体实施方式
为使本公开实施例的目的、技术方案和优点更加清楚，下面将结合本公开实施例中附图，对本公开实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本公开一部分实施例，而不是全部的实施例。通常在此处描述和示出的本公开实施例的组件可以以各种不同的配置来布置和设计。因此，以下对本公开的实施例的详细描述并非旨在限制要求保护的本公开的范围，而是仅仅表示本公开的选定实施例。基于本公开的实施例，本领域技术人员在没有做出创造性劳动的前提下所获得的所有其他实施例，都属于本公开保护的范围。
经研究发现，在控制自动驾驶车辆行驶时，通常通过对自动驾驶车辆所处的道路进行周期性的图像检测，然后依据检测得到的图像进行行驶的规划。通常，在进行行驶规划时，会采用二次规划(Quadratic Programming，QP)的方式进行。但由于二次规划方法在处理自动驾驶车辆的行驶规划的动态问题时，通常是通过拟合的方式确定横向和纵向分别对应的运动参量，而确定的运动参量并不能很好的适应于实际的运动学模型，导致利用这样的运动参量对自动驾驶车辆不能进行准确地控制，也即容易出现行驶规划的准确性较大的问题。
另外，由于在时间较短的检测周期中，二次规划方法仅能解决自动驾驶车辆在首要的安全性上的规划(也即完成避让车辆等基本操作)，但不能适时的保证稳定性(例如在避让车辆时，在控制周期中，时间仅能允许确定规划路径以保证避让的操作，而不能避免急转、突进的情况发生)，使得用户在乘坐自动驾驶车辆时，舒适性较低。
基于上述研究，本公开实施例提供了一种行驶规划方法，利用目标道路的参考轨迹信息以及自动驾驶车辆的车辆参数，确定自动驾驶车辆在当前控制周期中的运动参量关系信息，并由此得到自动驾驶车辆在横向和纵向分别对应的运动状态参量关系信息，从而基于预设的约束条件，确定自动驾驶车辆在未来至少一个控制周期的目标运动状态，利用可以较为准确获取的运动状态参量关系信息进行实际地行驶规划，从而能够对自动驾驶车辆进行更为准确地控制。
另外，由于在行驶规划时的效率较高，数据量较少，因此可以利用能够保证稳定性的约束条件，确定更为稳定的行驶规划，以提高用户在驾驶或者乘坐自动驾驶车辆时的舒适性。
针对以上方案所存在的缺陷，均是发明人在经过实践并仔细研究后得出的结果，因此，上述问题的发现过程以及下文中本公开针对上述问题所提出的解决方案，都应该是发明人在本公开过程中对本公开做出的贡献。
应注意到：相似的标号和字母在下面的附图中表示类似项，因此，一旦某一项在一个附图中被定义，则在随后的附图中不需要对其进行进一步定义和解释。
为便于对本实施例进行理解，首先对本公开实施例所公开的一种行驶规划方法进行 详细介绍，本公开实施例所提供的行驶规划方法的执行主体一般为自动驾驶控制设备。自动驾驶控制设备安装在自动驾驶车辆中，能够基于本公开实施例提供的行驶规划方法对自动驾驶的过程进行规划。在一些可能的实现方式中，该行驶规划方法可以通过处理器调用存储器中存储的计算机可读指令的方式来实现。
下面对本公开实施例提供的行驶规划方法加以说明。
参见图1所示，为本公开实施例提供的行驶规划方法的流程图，所述方法包括：
S101：基于目标道路的参考轨迹信息、以及自动驾驶车辆的车辆参数，确定所述自动驾驶车辆在当前控制周期中的运动状态参量关系信息；
S102：基于运动状态参量关系信息，确定所述自动驾驶车辆的横向运动状态参量关系信息、以及纵向运动状态参量关系信息；
S103：基于预设约束条件、所述横向运动状态参量关系信息以及所述纵向运动状态参量关系信息，确定所述自动驾驶车辆在未来至少一个控制周期的目标运动状态。
本公开实施例通过基于目标道路的参考轨迹信息、以及自动驾驶车辆的车辆参数，确定自动驾驶车辆在当前控制周期中的运动状态参量关系信息，然后依据利用运动参量关系信息确定的自动驾驶车辆的横向运动状态参量关系信息、纵向运动状态参量关系信息、以及预设的约束条件，确定自动驾驶车辆在未来至少一个控制周期的目标运动状态。这种方式与现有的方式相比，可以更为准确地对自动驾驶车辆进行实际行驶规划，从而能够对自动驾驶车辆进行更为准确的控制。此外，还可以在保证安全性的基础上，提供更稳定的行驶规划，以提高用户乘坐自动驾驶车辆时的舒适性。
下面对上述S101～S103加以详细说明。
针对上述S101，自动驾驶车辆(Autonomous vehicles)例如可以包括全自动无人驾驶汽车、半自动无人驾驶汽车等智能驾驶汽车，或者轮式移动机器人、以及履带式移动机器人等智能机器人。在不同的行驶场景中，自动驾驶车辆不同，自动驾驶车辆行驶时所处的目标道路也不同。
示例性的，在道路行驶场景中，自动驾驶车辆可以包括全自动、或半自动无人驾驶汽车。自动驾驶车辆在道路行驶场景中行驶时，对应的目标道路例如可以包括公路、桥梁道路中的行驶车道。在智能仓储场景中，自动驾驶车辆可以包括轮式移动机器人、或者履带式机器人。自动驾驶车辆在仓库中行驶时，对应的目标道路例如可以包括在在货架间预留的机器人行驶车道。
本实施例以自动驾驶车辆为智能驾驶汽车，且在道路行驶场景中行驶为例进行说明。
自动驾驶车辆在目标道路上行进时，主要通过在控制周期对自动驾驶车辆所处的目标道路进行周期性的检测，然后基于检测的结果进行行驶规划，确定在下一时刻到达的位置、速度、加速度等运动状态参量。根据这些运动状态参量，可以反解得到相应时刻、 自动驾驶车辆的动力输出装置输出的力矩、以及行驶转动的角度，从而控制自动驾驶车辆行进。
其中，为了保证自动驾驶车辆在行驶时的实时性，例如可以将一个控制周期的时间设置为2s，或者为了保证自动驾驶车辆的安全性，将一个控制周期的时间缩短至1s或者1.5s甚至更小，本实施例中对控制周期的时长不做限定。
在一些可选实施例中，目标道路的参考轨迹信息包括目标道路的中心线上的多个位置点，以及多个位置点在笛卡尔坐标系下的坐标值。
其中，参考轨迹信息可以利用搭载在自动驾驶车辆上的其他装置获得，例如高精地图模块。示例性的，高精地图模块，例如可以包括深度相机设备。高精地图模块可以利用下述至少一种方案对目标道路进行图像采集：结构光(Structured-light)、双目视觉(Stereo)、以及光飞行时间法(Time of Flight，TOF)。在高精地图模块采集到目标道路的道路图像后，还可以通过对道路图像进行检测，确定道路图像中的中心线，并确定中心线上的多个位置点。其中，中心线上的多个位置点例如可以包括中心线两端分别对应的位置点，以较为简单地确定中心线的位置；或者，可以包括中心线两端分别对应的位置点、以及在道路图像显示出的中心线中两个端点之间的至少一个点，以更加准确地表征中心线的位置。
在确定目标道路的中心线上的多个位置点后，还可以获取多个位置点在笛卡尔坐标系下的坐标值。其中，由于确定的多个位置点可以是利用搭载在自动驾驶车辆上的高精地图模块得到的，因此可以先确定多个位置点在高精地图模块采集得到的道路图像对应的图像坐标系中的位置信息，然后再利用图像坐标系、与目标道路对应的场景坐标系之间的坐标系转换关系，确定多个位置点在场景坐标系下的位置信息。由于在确定目标道路时，可以较为容易地根据全球定位***(Global Positioning System，GPS)确定二维位置信息，因此可以依据自动驾驶车辆的所在位置作为坐标原点建立目标道路的场景坐标系，该场景坐标系即为为目标道路确定的笛卡尔坐标系。然后，利用建立得到的该笛卡尔坐标系，确定中心线上的多个位置分别对应的坐标值。
另外，还可以确定自动驾驶车辆的车辆参数。示例性的，自动驾驶车辆的车辆参数可以包括动力性能参数、以及车身参数。其中，动力性能参数例如可以包括汽车满载、发动机最大扭矩、起步换挡加速度、制动性能等可以直接控制车辆行驶的参数。车身参数例如可以包括车体长度、车体宽度等可以确定车辆的行驶占据区域空间大小的参数。
在确定目标道路的参考轨迹信息、以及自动驾驶车辆的车辆参数后，即可以确定自动驾驶车辆在当前控制周期中的运动状态参量关系信息。
在一些可选实施例中，可以根据下述方式确定自动驾驶车辆在当前控制周期中的运动状态参量关系信息：基于目标道路的参考轨迹信息、以及自动驾驶车辆的当前位置，建立弗莱纳坐标系(Frenet-Serret frame)；在弗莱纳坐标系下，基于自动驾驶车辆的车 辆参数，确定自动驾驶车辆在当前控制周期中的运动状态参量关系信息。
示例性的，运动状态参量关系信息可以利用变加速度与位置、速度、以及加速度之间的相互关系公式，例如用运动学方程来表示。
具体地，由于对自动驾驶车辆的行驶规划是一个具有多个非线性约束的高维优化问题，进行处理的数据量较大，且场景较为复杂，同时出于安全性的考虑，需要保证实时性。因此在进行自动驾驶车辆的行驶规划时，选用弗莱纳坐标系，以在适应直行道路规划的同时，适应自动驾驶弯道的路线规划。同时，利用弗莱纳坐标系还可以将地图数据降维处理，计算量减小，提高效率，以满足行驶规划的实时性要求。
在具体实施中，可以根据自动驾驶车辆的当前位置确定弗莱纳坐标系的坐标原点，然后根据确定的目标道路的参考轨迹确定弗莱纳坐标系。具体地，可以以确定的坐标原点作为切点，沿着参考线的方向确定切线，作为弗莱纳坐标系的S轴，将在切点处与确定的切线垂直的法线确定为弗莱纳坐标系的D轴。
在建立弗莱纳坐标系后，基于参考线的位置，即可以简单的使用纵向距离(即沿着中心线方向的距离)、以及横向距离(即偏离中心线的距离)确定自动驾驶车辆的位置。自动驾驶车辆在纵向以及横向两个方向上的速度、加速度、变加速度等的计算也更为简单。
参见图2所示，为本公开实施例提供的一种弗莱纳坐标系的示意图；其中，包括自动驾驶车辆21、行驶道路22，在行驶道路22中包含中心线23，以及对应的弗莱纳坐标系24。
在确定运动状态参量关系信息时，由于对于自动驾驶的实际行驶场景而言，较少的存在翻滚、漂移等具有危险性和专业性的行驶需求，更多在保证交通行驶安全的前提下正常行驶，因此自动驾驶车辆的行驶动作可以简单的拆解为横向、以及纵向的不同行驶策略，也即运动状态参量可以包括横向运动状态参量以及纵向运动状态参量信息。其中，纵向运动状态参量，包括：纵向位置，纵向速度，纵向加速度及纵向变加速度中至少一种；横向运动状态参量包括：横向位置，横向角度，横向角速度，横向角加速度中至少一种。所述运动状态参量关系信息表征各个运动状态参量之间的关系，所述横向运动状态参量关系信息表征各个横向运动状态参量之间的关系，所述纵向运动状态参量关系信息表征各个纵向运动状态参量之间的关系。
这样，由于不同的运动状态参量可以表征不同的运动情况，因此设置较多种类的运动状态参量，可以更好的对自动驾驶车辆的行驶做出规划。
其中，自动驾驶车辆在当前控制周期中的运动状态参量关系信息，例如可以根据下述至少一种模型确定：单车模型(Bicycle Model)以及四轮模型确定。具体的，在使用单车模型时，还可以细分为以后轴为原点的车辆运动模型、以质心为中心的车辆运动学模型、阿克曼转向几何模型(Ackerman Turning Geometry)。
此外，由于在利用此关系信息进行行驶规划时，最核心的部分是规划在一个控制周期中的起始点、以及结束点的位置、速度、加速度，因此在利用模型确定自动驾驶车辆在当前控制周期中的运动状态参量关系信息时，以变加速度为输入的状态量，可以确定与输出的控制量，包括位置、速度、加速度中至少一种分别对应的关系信息。此时，由于是直接利用自动驾驶车辆的变加速度作为输入确定表征其运动状态的运动状态参量关系信息的，因此得到的运动状态参量关系信息中既包含了自动驾驶车辆在纵向上的运动状态参量的相关信息，也包含了自动驾驶车辆在纵向上的运动状态参量的相关信息。
针对上述S102，在确定运动状态参量信息后，可以对运动状态参量关系信息进行横向与纵向的解耦，得到横向运动状态参量关系信息和/或纵向运动状态参量关系信息。
在一些可选实施例中，在对运动状态参量关系信息进行横向与纵向的解耦时，例如可以采用下述方式：将横向状态参量的数值设置为预设数值后，基于运动状态参量关系信息，得到纵向运动状态参量关系信息；基于纵向运动状态参量关系信息、以及运动状态参量关系信息，得到横向运动状态参量关系信息。
具体地，在确定纵向运动状态参量关系信息时，可以将可以表征横向状态的数值，例如横向角加速度，设置为预设数值，例如为0，以消除横向的移动对纵向可能造成的影响。然后，在状态参量关系信息中带入横向角加速度的数值0，即可得到纵向运动状态参量关系信息。此时，即可以确定自动驾驶车辆的在纵向上的运动状态参量的相关信息，利用得到的此在纵向上的运动状态参量的相关信息、以及运动状态参量信息，即可以确定在横向上的运动状态参量的相关信息，完成对自动驾驶车辆横向、纵向上运动状态参量信息的解耦。
这样，通过纵向运动状态参量关系信息，可以较为便利快捷地获取较优的纵向运动状态；同样的，由于自动驾驶车辆的部分横向运动状态与纵向运动状态相关，因此利用得到的较优的纵向运动状态、以及横向运动状态参量关系信息，可以较为便利快捷地获取较优的横向运动状态。
针对上述S103，预设约束条件例如可以包括横向约束条件和/或纵向约束条件。在一些可选实施例中，纵向约束条件例如可以包括所述自动驾驶车辆的纵向位移与时间的第一变化关系信息；其中，所述第一变化关系信息用于表征所述自动驾驶车辆在未来多个时刻的纵向位移边界。
可选地，纵向约束条件还包括下述至少一种：自动驾驶车辆的纵向位移与时间的第一变化关系信息、纵向速度阈值、相邻控制周期对应的纵向位移变化量阈值、相邻控制周期对应的纵向速度变化量阈值、相邻控制周期对应的加速度变化量阈值、相邻控制周期对应的纵向变加速度变化量阈值。
示例性的，纵向约束条件的说明，如下述(1a)～(1d)：
(1a)、自动驾驶车辆的纵向位移与时间的第一变化关系信息。
可选地，例如可以利用下述方式确定第一变化关系信息：获取障碍物在笛卡尔坐标系下的障碍物轨迹信息；将所述障碍物轨迹信息投影至弗莱纳坐标系下；基于所述障碍物轨迹信息在弗莱纳坐标系下的投影结果，确定所述自动驾驶车辆的纵向位移与时间的第一变化关系信息。
这样，利用弗莱纳坐标系可以较为容易的表征目标道路的信息，还可以将地图数据降维处理，计算量减小，提高效率，以满足行驶规划的实时性要求。
其中，障碍物预测轨迹信息例如可以是在自动驾驶车辆上搭载的预测模块、或者其他感知模块输出的轨迹信息。示例性的，预测模块例如可以包括能够实时获取障碍物轨迹信息的激光雷达设备。
此时，得到的障碍物预测轨迹，与上述S101中确定的多个位置点分别在笛卡尔坐标系下的坐标值相似，均为在笛卡尔坐标系下的坐标值。然后，利用与上述S101中使用的图像坐标系、与目标道路对应的场景坐标系之间的坐标系转换关系，即可以将障碍物轨迹信息投影在弗莱纳坐标系下，并基于障碍物轨迹信息在弗莱纳坐标系下的投影结果，确定自动驾驶车辆的纵向位移与时间的第一变化关系信息。
参见图3所示，为本公开实施例提供的一种利用自动驾驶车辆的纵向位移与时间的第一变化关系信息，确定的S-T坐标系的示意图。其中，S-T坐标系的横坐标31，表征时间t；S-T坐标系的纵坐标32，表征自动驾驶车辆的纵向位移距离s。
在t0～t1时间段，自动驾驶车辆的前方存在车辆A，车辆A的轨迹33限制自动驾驶车辆在此时段的任意时刻向前行驶的最大距离为s1，若自动驾驶车辆向前行驶距离超过s1，则有较大的可能性与车辆A相撞。在t1～t2时间段，车辆A的行驶轨迹33发生变化，变化为行驶轨迹34，存在加速远离自动驾驶车辆的行驶行为，因此自动驾驶车辆可以向前行驶的最大距离，随着车辆A的快速驶离，自动驾驶车辆在从t1～t2最大能够行驶的距离变化至s2。
在t2时刻，车辆B对自动驾驶车辆进行超车，为了可以避免与该车辆B碰撞，自动驾驶车辆在t2时刻的向前行驶距离不应低于s3，否则将与车辆B相撞。在t2～t3时间段，由于车辆A与自动驾驶车辆行驶距离较远，不构成安全性威胁，但是由于车辆自身速度限制、以及道路限速规则的限制，存在可向前行驶的最大距离s4。
此外，由于在纵向，自动驾驶车辆的行进路线一般为前进向，因此车辆的投影轨迹仅在s轴的正半轴出现。当自动驾驶车辆的行进路线为后退方向时，例如倒车的场景，存在向车辆方向靠近的障碍物时，需要紧急制动，因此不在该示例中。
(1b)、纵向速度阈值。
纵向速度阈值，例如可以仅包括速度的上界，例如速度最高不得超过80公里/小时；或者仅包括速度的下界，例如速度最低不得低于5公里/小时；或者既包括速度的上界，又包括速度的下界，例如速度限制在5公里/小时与80公里/小时之间。具体地，可以根 据实际的道路限速规则制定，例如在学校附近区域行驶时，规定限速30公里/小时，因此可以设置纵向速度阈值包括速度的上界30公里/小时；在高速路段行驶时，规定最低速度不得低于60公里/小时，且不得超过120公里/小时，因此可以设置纵向速度阈值包括速度的上界120公里/小时，以及速度的下界60公里/小时。
此外，在设置纵向速度阈值时，还可以在满足安全性的前提下，基于预设的驾驶策略进行设置，以提高在使用该自动驾驶车辆时的舒适度。示例性的，在高速路段行驶时，由于在刚进入高速路段时，需要有加速的过程，而在行驶一段距离，例如500米后，车辆即可以进入高速行驶的状态，因此可以在不得低于60公里/小时，且不得超过120公里/小时的规定限速内，设置不得低于80公里/小时，且不得超过100公里/小时的更小范围的限速的驾驶策略，以使自动驾驶车辆能够在保证安全行驶的前提下，较为匀速的行驶，可以使自动驾驶车辆在行驶时保持稳定，较少的出现由于速度变化过大，导致的人体猛冲或者后仰的行为，有助于提高用户在乘坐自动驾驶汽车时的舒适度。
(1c)、相邻控制周期对应的纵向位移变化量阈值。
其中，通过相邻控制周期对应的纵向位移变化量阈值，可以设置在较长一段时间内自动驾驶车辆保持较为匀速的行驶，并且，还可以保证在存在需要避让或者超车的车辆出现时，自动驾驶车辆实现较为平缓的变速。
示例性的，由于在速度保持不变时，相邻控制周期对应的纵向位移变化量应当保持不变；在速度仅有较小程度的变化时，相邻控制周期对应的纵向位移变化量对应的会有较小程度的变化，因此可以根据实际情况设置相邻控制周期对应的纵向位移变化量阈值例如包括不超过100米，或者不超过200米，以保证自动驾驶车辆可以在较小幅度的速度变化下行驶，减少速度突变的情况。
(1d)、相邻控制周期对应的纵向速度变化量阈值、相邻控制周期对应的加速度变化量阈值、相邻控制周期对应的纵向变加速度变化量阈值。
其中，由于(1d)中的任一纵向变化量阈值的设置方式与(1c)中相邻控制周期对应的纵向位移变化量阈值的设置的方式相似，因此此处不再进行赘述。
在一些可选实施例中，横向约束条件例如可以包括所述自动驾驶车辆的横向位移与时间的第二变化关系信息；其中，所述第二变换关系信息用于表征所述自动驾驶车辆在未来多个时刻的横向位移边界。
可选地，横向约束条件还包括下述至少一种：自动驾驶车辆的横向位移与时间的第二变化关系信息、相邻控制周期对应的横向位移变化量阈值、相邻控制周期对应的横向角度变化量阈值、相邻控制周期对应的横向角速度变化量阈值、相邻控制周期对应的横向角加速度变化量阈值。
示例性的，横向约束条件的说明，如下述(2a)～(2c)：
(2a)、自动驾驶车辆的横向位移与时间的第二变化关系信息。
可选地，例如可以利用下述方式确定第二变化关系信息：获取障碍物在笛卡尔坐标系下的障碍物轨迹信息；将所述障碍物轨迹信息投影至弗莱纳坐标系下；基于所述障碍物轨迹信息在弗莱纳坐标系下的投影结果，确定所述自动驾驶车辆的横向位移与时间的第二变化关系信息。
其中，确定第二变化关系信息的方式，与上述(1a)中确定第一变化关系信息的方式相似，在此不再赘述。
参见图4所示，为本公开实施例提供的一种利用自动驾驶车辆的纵向位移与时间的第一变化关系信息，确定的D-T坐标系的示意图。其中，D-t坐标系的横坐标41，表征时间t；D-T坐标系的纵坐标42，表征自动驾驶车辆的横向位移距离d。图中，以D轴的负半轴表征自动驾驶车辆与右侧的道路边线之间的距离；以D轴的负半轴表征自动驾驶车辆与左侧的道路边线之间的距离。
此处，为了便于理解，认为该自动驾驶车辆沿目标道路的中心线为基准行驶，因此，为了躲避左右两个车道上并入中心线的车辆，该自动驾驶车辆可以沿中心线向两侧偏离。并且，自动驾驶车辆左侧、右侧的道路边线，与中心线的距离分别为s1、s2。
在t0～t1时间段，自动驾驶车辆右侧存在车辆C，车辆C的行驶轨迹43使得自动驾驶车辆向右行驶受限，且向右行驶的距离根据车辆C的行驶轨迹43限制在s3内。在t1～t2时间段，车辆C驶出，自动驾驶车辆可向右行驶的距离恢复至最大距离s2。同时，在t0～t2时间段，由于自动驾驶车辆的左侧没有并入的车辆，因此可以向左行驶的距离为最大的向左行驶距离s1。
在t2～t3时间段，自动驾驶车辆右侧存在车辆D，车辆D的行驶轨迹44使得自动驾驶车辆向右行驶受限，且向右行驶的距离根据车辆D的行驶轨迹44限制在s4内。同样的，在t3时刻，车辆D驶离自动驾驶车辆，自动驾驶车辆可向左行驶的距离恢复至最大距离s1。同时，在t2时刻以后，由于自动驾驶车辆的右侧没有并入的车辆，因此可以向右行驶的距离为最大的向右行驶距离s2。
(2b)、相邻控制周期对应的横向位移变化量阈值。
此处，相邻控制周期对应的横向位移变化量阈值的确定方式，与上述(1c)中确定相邻控制周期对应的纵向位移变化量阈值的方式相似，再次不再赘述。
(2c)、相邻控制周期对应的横向角度变化量阈值、相邻控制周期对应的横向角速度变化量阈值、相邻控制周期对应的横向角加速度变化量阈值。
其中，由于(2c)中的任一横向的变化量阈值的设置方式与(2b)中相邻控制周期对应的横向位移变化量阈值的设置的方式相似，因此此处不再进行赘述。
这样，通过设置多个不同的约束条件，可以在保证自动驾驶车辆安全性的同时，有针对性的对自动驾驶车辆行驶的稳定性提出更高的要求，使得自动驾驶车辆不会发生位移、速度、加速度等的突变，减少急刹车、猛冲等行为的发生，从而保证乘坐自动驾驶 车辆的用户，可以在安全行驶的前提下，有更舒适的乘车体验。
在一些可选实施例中，在基于预设约束条件、横向运动状态参量关系信息以及纵向运动状态参量关系信息，确定自动驾驶车辆在未来至少一个控制周期分别对应的目标运动状态时，例如可以采用下述方式：基于纵向约束条件、以及所述纵向运动状态参量关系信息、以及纵向行驶策略，确定所述自动驾驶车辆在所述未来至少一个控制周期的纵向运动状态；基于所述横向约束条件、所述横向运动状态参量关系信息、横向行驶策略、以及所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态，确定所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态。
其中，自动驾驶车辆在未来至少一个控制周期分别对应的目标运动状态包括：所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态、以及纵向运动状态。可选地，至少一个控制周期分别对应的纵向变加速度以及横向变加速度例如可以作为表征所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态、以及纵向运动状态的一个运行参量。
在具体实施中，在基于纵向约束条件、以及所述纵向运动状态参量关系信息、以及行驶策略，确定所述自动驾驶车辆在所述未来至少一个控制周期的纵向运动状态时，例如可以采用下述方式：基于所述纵向约束条件、预设的优化步长、优化时间域、道路限速、车辆性能参数、以及所述纵向行驶策略，生成第一目标函数；所述第一目标函数以每个优化时间点的纵向运动状态、与目标纵向状态的第一距离为优化目标；基于所述纵向运动状态参量关系信息、以及所述纵向约束条件，在所述优化时间域内的每个优化时间点上，对第一目标函数进行最优变量序列寻优，得到由多个优化时间点的纵向车辆状态变量构成的第一变量序列；基于所述第一变量序列，确定所述自动驾驶车辆在所述未来至少一个控制周期的纵向运动状态。
其中，纵向约束条件可以参见上述对纵向约束条件的说明，这里不再赘述。预设的优化步长例如可以根据自动驾驶***的可选步长、或者根据实际的驾驶需求确定，具体可以包括0.1秒或0.2秒，以保证自动驾驶车辆能够迅速对不断变化的行驶环境做出反应。另外，还可以设置优化时间域，在一次行驶规划中，能够得到该优化时间域内多个优化时间点分别对应的运动状态。随着自动驾驶车辆的驾驶，可以按照控制周期，实时进行运动状态的优化，以提升控制精度。
对于道路限速、车辆性能参数而言，由于对应于不同的道路，其对应的道路限速是确定的，因此可以根据实际行驶的路段相应的确定对应时刻的道路限速；而对于车辆性能参数，不同的自动驾驶车辆由与其相对应的车辆参数，具体可以参见上述S101对应的实施例中对自动驾驶车辆的车辆参数的说明，在此不再赘述。
另外，纵向形式策略例如可以包括自动驾驶车辆在纵向行驶时的行为策略，例如可以包括在纵向上的超车、避让等的行为策略。
利用上述说明的纵向对应的相关信息可以确定第一目标函数，第一目标函数例如可以包括以优化时间点为自变量，并利用上述说明的纵向对应的相关信息作为映射关系信息，将目标纵向状态的第一距离作为因变量，也即优化目标。
在利用纵向运动状态参量关系信息、以及所述纵向约束条件，在所述优化时间域内的每个优化时间点上，对第一目标函数进行最优变量序列寻优时，例如可以利用第一目标函数建立纵向优化模型，然后，将纵向约束条件作为纵向优化模型的约束条件，确定未来至少一个控制周期分别对应的最优的一个或者多个第一变量序列。然后，在根据所述第一变量序列，确定所述自动驾驶车辆在所述未来至少一个控制周期的纵向运动状态。
另外，在利用纵向优化模型确定纵向运动状态时，为了保证得到的纵向运动状态在应用于实际的控制周期时保证稳定性，以使乘坐的用户能更加舒适的行车环境，还可以引入惩罚方法，例如可以通过确定相邻控制周期对应的纵向运动状态中的纵向位移变化量、相邻控制周期对应的纵向速度变化量、相邻控制周期对应的加速度变化量、相邻控制周期对应的纵向变加速度变化量，设定惩罚。
具体地，在设定惩罚时，例如可以直接依据各个变化量作为惩罚，在变化量较大时，训练模型向变化量较小的方向输出纵向运动状态。此外，在希望获得更为稳定的纵向运动状态时，还可以将对变化量进行平方、三次方等运算处理，并将运算处理后的结果作为惩罚，以加大对纵向优化模型的调整力度，输出在相邻控制周期之间变化量更小的纵向运动状态，以使自动驾驶车辆在行驶过程中在保证安全性的同时，更加稳定舒适。
同样的，在确定自动驾驶车辆在未来至少一个控制周期分别对应的横向运动状态时，由于自动驾驶车辆的横向运动状态会受到纵向运动状态的影响，例如在纵向行驶过程较快时，若有并道的车辆出现，需要更早的进行横向移动，以保证可以安全的避让，并且以较为缓和的方式偏转行进。因此，在具体确定自动驾驶车辆在未来至少一个控制周期分别对应的横向运动状态的方式时，还需要考虑自动驾驶车辆在未来至少一个控制周期分别对应的纵向运动状态，例如将确定的纵向运动状态作为确定横向运动状态时的约束条件。
在一些可选实施例中，在确定自动驾驶车辆在未来至少一个控制周期分别对应的横向运动状态时，例如可以采用下述方法：基于所述横向运动状态、所述横向行驶策略、所述预设的优化步长、优化时间域、以及所述横向约束条件，生成第二目标函数，所述第二目标函数以每个优化时间点的横向运动状态、与目标横向状态的第二距离为优化目标；基于所述横向运动状态参量关系信息、以及所述横向约束条件，在所述优化时间域内的每个优化时间点上，对第二目标函数进行最优变量序列寻优，得到由多个优化时间点的横向车辆状态变量构成的第二变量序列；基于所述第二变量序列，确定所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态。
这样，可以对自动驾驶车辆状态的横向、纵向进行解耦，更有针对性的进行横向、纵向的规划，使得规划的行驶策略更加细致，从而保证获得较优的自动驾驶车辆的行驶规划。
此处，确定自动驾驶车辆在未来至少一个控制周期分别对应的横向运动状态的方式，与上述确定自动驾驶车辆在未来至少一个控制周期分别对应的纵向运动状态的方式相似，在此不再赘述。
在一些可选实施例中，在确定未来至少一个控制周期的横向运动状态、以及未来至少一个控制周期的纵向运动状态后，还可以按照多个优化时间点，将该优化时间点的纵向运动状态、和该优化时间点的横向运动状态进行合并，生成优化轨迹；所述优化轨迹包括：横向运动轨迹、以及纵向速度规划曲线。
其中，由于各优化时间点对应的横向运动状态、以及纵向运动状态均已经确定，因此可以按照相同的优化时间点对横向运动状态以及纵向运动状态进行合并，从而得到一条包含横、纵向位置，并且考虑到安全性和舒适性的三维优化轨迹。
在本公开另一实施例中，在确定自动驾驶车辆在未来至少一个控制周期分别对应的纵向运动状态、以及确定自动驾驶车辆在未来至少一个控制周期分别对应的横向运动状态后，还包括：利用未来至少一个控制周期中的至少部分目标控制周期分别对应的目标运动状态，确定自动驾驶车辆在目标控制周期输出的动力参数。
这样，利用确定的目标运动状态，可以解算得到自动驾驶车辆在目标控制周期输出的动力参数，从而可以利用该动力参数更直接的对自动驾驶车辆的行驶进行控制。
可选地，目标控制周期例如可以包括可以确定纵向运动状态、以及横向运动状态的至少一个控制周期。
在确定自动驾驶车辆在未来至少一个控制周期分别对应的纵向运动状态、以及确定自动驾驶车辆在未来至少一个控制周期分别对应的横向运动状态后，由于对于未来至少一个控制周期，均确定了对应的纵向运动状态以及横向运动状态，因此，可以在未来至少一个控制周期将纵向运动状态以及横向运动状态按照控制周期中相同的时间点进行合并，以完成对自动驾驶车辆的行驶规划。在确定自动驾驶车辆的行驶规划后，利用规划的结果，可以确定自动驾驶车辆在未来至少一个控制周期中任一时刻的运动状态，作为目标运动状态，并根据自动驾驶车辆的动力参数与行驶行为的对应关系，确定在目标控制周期输出的动力参数。
本领域技术人员可以理解，在具体实施方式的上述方法中，各步骤的撰写顺序并不意味着严格的执行顺序而对实施过程构成任何限定，各步骤的具体执行顺序应当以其功能和可能的内在逻辑确定。
基于同一发明构思，本公开实施例中还提供了与行驶规划方法对应的行驶规划装置，由于本公开实施例中的装置解决问题的原理与本公开实施例上述行驶规划方法相 似，因此装置的实施可以参见方法的实施，重复之处不再赘述。
参照图5所示，为本公开实施例提供的一种行驶规划装置的示意图，所述装置包括：第一确定模块51、第二确定模块52和第三确定模块53；其中，
第一确定模块51，配置为基于目标道路的参考轨迹信息、以及自动驾驶车辆的车辆参数，确定所述自动驾驶车辆在当前控制周期中的运动状态参量关系信息；第二确定模块52，配置为基于运动状态参量关系信息，确定所述自动驾驶车辆的横向运动状态参量关系信息、以及纵向运动状态参量关系信息；第三确定模块53，配置为基于预约束条件、所述横向运动状态参量关系信息以及所述纵向运动状态参量关系信息，确定所述自动驾驶车辆在未来至少一个控制周期的目标运动状态。
一种可选的实施方式中，所述预设约束条件包括：横向约束条件和/或纵向约束条件。
一种可选的实施方式中，所述纵向约束条件包括：所述自动驾驶车辆的纵向位移与时间的第一变化关系信息；其中，所述第一变化关系信息用于表征所述自动驾驶车辆在未来多个时刻的纵向位移边界。
一种可选的实施方式中，所述第三确定模块53，配置为采用以下步骤确定所述自动驾驶车辆的纵向位移与时间的第一变化关系信息：获取障碍物在笛卡尔坐标系下的障碍物轨迹信息；将所述障碍物轨迹信息投影至弗莱纳坐标系下；基于所述障碍物轨迹信息在弗莱纳坐标系下的投影结果，确定所述自动驾驶车辆的纵向位移与时间的第一变化关系信息。
一种可选的实施方式中，所述横向约束条件包括：所述自动驾驶车辆的横向位移与时间的第二变化关系信息；其中，所述第二变换关系信息用于表征所述自动驾驶车辆在未来多个时刻的横向位移边界。
一种可选的实施方式中，所述第三确定模块53，配置为采用以下步骤确定所述自动驾驶车辆的横向位移与时间的第二变化关系信息：获取障碍物在笛卡尔坐标系下的障碍物轨迹信息；将所述障碍物轨迹信息投影至弗莱纳坐标系下；基于所述障碍物轨迹信息在弗莱纳坐标系下的投影结果，确定所述自动驾驶车辆的横向位移与时间的第二变化关系信息。
一种可选的实施方式中，所述纵向约束条件还包括下述至少一种：纵向速度阈值、相邻控制周期对应的纵向位移变化量阈值、相邻控制周期对应的纵向速度变化量阈值、相邻控制周期对应的加速度变化量阈值、相邻控制周期对应的纵向变加速度变化量阈值；所述横向约束条件还包括下述至少一种：相邻控制周期的横向位移变化量阈值、相邻控制周期的横向角度变化量阈值、相邻控制周期的横向角速度变化量阈值、相邻控制周期的横向角加速度变化量阈值。
一种可选的实施方式中，所述自动驾驶车辆在未来至少一个控制周期的目标运动状态，包括：所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态、以及纵向 运动状态；所述第三确定模块53，配置为基于纵向约束条件、以及所述纵向运动状态参量关系信息、以及纵向行驶策略，确定所述自动驾驶车辆在所述未来至少一个控制周期的纵向运动状态；基于所述横向约束条件、所述横向运动状态参量关系信息、横向行驶策略、以及所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态，确定所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态。
一种可选的实施方式中，所述第三确定模块53，配置为基于所述纵向约束条件、预设的优化步长、优化时间域、道路限速、车辆性能参数、以及所述纵向行驶策略，生成第一目标函数；所述第一目标函数以每个优化时间点的纵向运动状态、与目标纵向状态的第一距离为优化目标；基于所述纵向运动状态参量关系信息、以及所述纵向约束条件，在所述优化时间域内的每个优化时间点上，对第一目标函数进行最优变量序列寻优，得到由多个优化时间点的纵向车辆状态变量构成的第一变量序列；基于所述第一变量序列，确定所述自动驾驶车辆在所述未来至少一个控制周期的纵向运动状态。
一种可选的实施方式中，所述第三确定模块53，配置为基于所述横向运动状态、所述横向行驶策略、所述预设的优化步长、优化时间域、以及所述横向约束条件，生成第二目标函数，所述第二目标函数以每个优化时间点的横向运动状态、与目标横向状态的第二距离为优化目标；基于所述横向运动状态参量关系信息、以及所述横向约束条件，在所述优化时间域内的每个优化时间点上，对第二目标函数进行最优变量序列寻优，得到由多个优化时间点的横向车辆状态变量构成的第二变量序列；基于所述第二变量序列，确定所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态。
一种可选的实施方式中，所述行驶规划装置还包括轨迹生成模块54，配置为按照多个优化时间点，将该优化时间点的纵向运动状态、和该优化时间点的横向运动状态进行合并，生成优化轨迹；所述优化轨迹包括：横向运动轨迹、以及纵向速度规划曲线。
一种可选的实施方式中，运动状态参量包括：所述横向运动状态参量以及纵向运动状态参量；所述纵向运动状态参量，包括：纵向位置，纵向速度，纵向加速度及纵向变加速度中至少一种；所述横向运动状态参量包括：横向位置，横向角度，横向角速度，横向角加速度中至少一种；所述运动状态参量关系信息表征各个运动状态参量之间的关系，所述横向运动状态参量关系信息表征各个横向运动状态参量之间的关系，所述纵向运动状态参量关系信息表征各个纵向运动状态参量之间的关系。
一种可选的实施方式中，所述参考轨迹信息包括：所述目标道路的中心线上的多个位置点，以及所述多个位置点分别在笛卡尔坐标系下的坐标值。
一种可选的实施方式中，所述第一确定模块51，配置为基于所述目标道路的参考轨迹信息、以及所述自动驾驶车辆的当前位置，建立弗莱纳坐标系；在所述弗莱纳坐标系下，基于所述自动驾驶车辆的车辆参数，确定所述自动驾驶车辆在当前控制周期中的运动状态参量关系信息。
一种可选的实施方式中，所述第二确定模块52，配置为对所述运动状态参量关系信息进行横向与纵向的解耦，得到所述横向运动状态参量关系信息和/或纵向运动状态参量关系信息。
一种可选的实施方式中，所述第二确定模块52，配置为将所述横向状态参量的数值设置为预设数值后，基于所述运动状态参量关系信息，得到所述纵向运动状态参量关系信息；基于所述纵向运动状态参量关系信息、以及所述运动状态参量关系信息，得到所述横向运动状态参量关系信息。
一种可选的实施方式中，所述未来至少一个控制周期的目标运动状态包括：所述至少一个控制周期的纵向变加速度以及横向变加速度。
一种可选的实施方式中，所述行驶规划装置还包括，第四确定模块55，配置为利用所述未来至少一个控制周期中的至少部分目标控制周期的目标运动状态，确定所述自动驾驶车辆在目标控制周期输出的动力参数。
关于装置中的各模块的处理流程、以及各模块之间的交互流程的描述可以参照上述方法实施例中的相关说明，这里不再详述。
本公开实施例还提供了一种计算机设备，如图6所示，为本公开实施例提供的计算机设备结构示意图，包括：
处理器10和存储器20；所述存储器20存储有处理器10可执行的机器可读指令，处理器10用于执行存储器20中存储的机器可读指令，所述机器可读指令被处理器10执行时，处理器10执行本公开实施例前述行驶规划方法的步骤。
上述存储器20包括内存210和外部存储器220；这里的内存210也称内存储器，用于暂时存放处理器10中的运算数据，以及与硬盘等外部存储器220交换的数据，处理器10通过内存210与外部存储器220进行数据交换。
上述指令的具体执行过程可以参考本公开实施例中所述的行驶规划方法的步骤，此处不再赘述。
本公开实施例还提供一种计算机可读存储介质，该计算机可读存储介质上存储有计算机程序，该计算机程序被处理器运行时执行上述方法实施例中所述的行驶规划方法的步骤。其中，该存储介质可以是易失性或非易失的计算机可读取存储介质。
本公开实施例还提供一种计算机程序产品，所述计算机程序产品包括计算机可执行指令，该计算机可执行指令被执行后，能够实现上述方法实施例中所述的行驶规划方法的步骤，具体可参见上述方法实施例，在此不再赘述。
其中，上述计算机程序产品可以具体通过硬件、软件或其结合的方式实现。在一个可选实施例中，所述计算机程序产品具体体现为计算机存储介质，在另一个可选实施例中，计算机程序产品具体体现为软件产品，例如软件开发包(Software Development Kit，SDK)等等。
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***和装置的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。在本公开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。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，又例如，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通信接口，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，在本公开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，可以存储在一个处理器可执行的非易失的计算机可读取存储介质中。基于这样的理解，本公开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)执行本公开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、只读存储器(Read-Only Memory，ROM)、随机存取存储器(Random Access Memory，RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
最后应说明的是：以上所述实施例，仅为本公开的具体实施方式，用以说明本公开的技术方案，而非对其限制，本公开的保护范围并不局限于此，尽管参照前述实施例对本公开进行了详细的说明，本领域的普通技术人员应当理解：任何熟悉本技术领域的技术人员在本公开揭露的技术范围内，其依然可以对前述实施例所记载的技术方案进行修改或可轻易想到变化，或者对其中部分技术特征进行等同替换；而这些修改、变化或者替换，并不使相应技术方案的本质脱离本公开实施例技术方案的精神和范围，都应涵盖在本公开的保护范围之内。因此，本公开的保护范围应所述以权利要求的保护范围为准。

Claims (21)

  1. 一种行驶规划方法，包括：
    基于目标道路的参考轨迹信息、以及自动驾驶车辆的车辆参数，确定所述自动驾驶车辆在当前控制周期中的运动状态参量关系信息；
    基于运动状态参量关系信息，确定所述自动驾驶车辆的横向运动状态参量关系信息、以及纵向运动状态参量关系信息；
    基于预设约束条件、所述横向运动状态参量关系信息以及所述纵向运动状态参量关系信息，确定所述自动驾驶车辆在未来至少一个控制周期的目标运动状态。
  2. 根据权利要求1所述的行驶规划方法，其中，所述预设约束条件包括：横向约束条件和/或纵向约束条件。
  3. 根据权利要求2所述的行驶规划方法，其中，所述纵向约束条件包括：所述自动驾驶车辆的纵向位移与时间的第一变化关系信息；
    其中，所述第一变化关系信息用于表征所述自动驾驶车辆在未来多个时刻的纵向位移边界。
  4. 根据权利要求3所述的行驶规划方法，其中，采用以下步骤确定所述自动驾驶车辆的纵向位移与时间的第一变化关系信息：
    获取障碍物在笛卡尔坐标系下的障碍物轨迹信息；
    将所述障碍物轨迹信息投影至弗莱纳坐标系下；
    基于所述障碍物轨迹信息在弗莱纳坐标系下的投影结果，确定所述自动驾驶车辆的纵向位移与时间的第一变化关系信息。
  5. 根据权利要求2-4任一项所述的行驶规划方法，其中，所述横向约束条件包括：所述自动驾驶车辆的横向位移与时间的第二变化关系信息；
    其中，所述第二变换关系信息用于表征所述自动驾驶车辆在未来多个时刻的横向位移边界。
  6. 根据权利要求5所述的行驶规划方法，其中，采用以下步骤确定所述自动驾驶车辆的横向位移与时间的第二变化关系信息：
    获取障碍物在笛卡尔坐标系下的障碍物轨迹信息；
    将所述障碍物轨迹信息投影至弗莱纳坐标系下；
    基于所述障碍物轨迹信息在弗莱纳坐标系下的投影结果，确定所述自动驾驶车辆的横向位移与时间的第二变化关系信息。
  7. 根据权利要求2-6任一项所述的行驶规划方法，其中，所述纵向约束条件还包括下述至少一种：
    纵向速度阈值、相邻控制周期对应的纵向位移变化量阈值、相邻控制周期对应的纵 向速度变化量阈值、相邻控制周期对应的加速度变化量阈值、相邻控制周期对应的纵向变加速度变化量阈值；
    所述横向约束条件还包括下述至少一种：
    相邻控制周期的横向位移变化量阈值、相邻控制周期的横向角度变化量阈值、相邻控制周期的横向角速度变化量阈值、相邻控制周期的横向角加速度变化量阈值。
  8. 根据权利要求1-7任一项所述的行驶规划方法，其中，所述自动驾驶车辆在未来至少一个控制周期的目标运动状态，包括：所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态、以及纵向运动状态；
    所述基于预设约束条件、所述横向运动状态参量关系信息以及所述纵向运动状态参量关系信息，确定所述自动驾驶车辆在未来至少一个控制周期的目标运动状态，包括：
    基于纵向约束条件、以及所述纵向运动状态参量关系信息、以及纵向行驶策略，确定所述自动驾驶车辆在所述未来至少一个控制周期的纵向运动状态；
    基于所述横向约束条件、所述横向运动状态参量关系信息、横向行驶策略、以及所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态，确定所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态。
  9. 根据权利要求8所述的行驶规划方法，其中，所述基于纵向约束条件、以及所述纵向运动状态参量关系信息、以及纵向行驶策略，确定所述自动驾驶车辆在所述未来至少一个控制周期的纵向运动状态，包括：
    基于所述纵向约束条件、预设的优化步长、优化时间域、道路限速、车辆性能参数、以及所述纵向行驶策略，生成第一目标函数；所述第一目标函数以每个优化时间点的纵向运动状态、与目标纵向状态的第一距离为优化目标；
    基于所述纵向运动状态参量关系信息、以及所述纵向约束条件，在所述优化时间域内的每个优化时间点上，对第一目标函数进行最优变量序列寻优，得到由多个优化时间点的纵向车辆状态变量构成的第一变量序列；
    基于所述第一变量序列，确定所述自动驾驶车辆在所述未来至少一个控制周期的纵向运动状态。
  10. 根据权利要求8或9所述的行驶规划方法，其中，基于所述横向约束条件、所述横向运动状态参量关系信息、横向行驶策略、以及所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态，确定所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态，包括：
    基于所述横向运动状态、所述横向行驶策略、所述预设的优化步长、优化时间域、以及所述横向约束条件，生成第二目标函数，所述第二目标函数以每个优化时间点的横向运动状态、与目标横向状态的第二距离为优化目标；
    基于所述横向运动状态参量关系信息、以及所述横向约束条件，在所述优化时间域 内的每个优化时间点上，对第二目标函数进行最优变量序列寻优，得到由多个优化时间点的横向车辆状态变量构成的第二变量序列；
    基于所述第二变量序列，确定所述自动驾驶车辆在所述未来至少一个控制周期的横向运动状态。
  11. 根据权利要求8-10任一项所述的行驶规划方法，其中，所述方法还包括：按照多个优化时间点，将该优化时间点的纵向运动状态、和该优化时间点的横向运动状态进行合并，生成优化轨迹；
    所述优化轨迹包括：横向运动轨迹、以及纵向速度规划曲线。
  12. 根据权利要求1-11任一项所述的行驶规划方法，其中，运动状态参量包括：所述横向运动状态参量以及纵向运动状态参量；
    所述纵向运动状态参量，包括：纵向位置，纵向速度，纵向加速度及纵向变加速度中至少一种；
    所述横向运动状态参量包括：横向位置，横向角度，横向角速度，横向角加速度中至少一种；
    所述运动状态参量关系信息表征各个运动状态参量之间的关系，所述横向运动状态参量关系信息表征各个横向运动状态参量之间的关系，所述纵向运动状态参量关系信息表征各个纵向运动状态参量之间的关系。
  13. 根据权利要求1-12任一项所述的行驶规划方法，其中，所述参考轨迹信息包括：所述目标道路的中心线上的多个位置点，以及所述多个位置点分别在笛卡尔坐标系下的坐标值。
  14. 根据权利要求1-13任一项所述的行驶规划方法，其中，所述基于目标道路的参考轨迹、以及自动驾驶车辆的车辆参数，确定所述自动驾驶车辆在当前控制周期中的运动状态参量关系信息，包括：
    基于所述目标道路的参考轨迹信息、以及所述自动驾驶车辆的当前位置，建立弗莱纳坐标系；
    在所述弗莱纳坐标系下，基于所述自动驾驶车辆的车辆参数，确定所述自动驾驶车辆在当前控制周期中的运动状态参量关系信息。
  15. 根据权利要求1-14任一项所述的行驶规划方法，其中，所述基于运动状态参量关系信息，确定所述自动驾驶车辆的横向运动状态参量关系信息、以及纵向运动状态参量关系信息，包括：
    对所述运动状态参量关系信息进行横向与纵向的解耦，得到所述横向运动状态参量关系信息和/或纵向运动状态参量关系信息。
  16. 根据权利要求15所述的行驶规划方法，其中，所述对所述运动状态参量关系信息进行横向与纵向的解耦，包括：
    将所述横向状态参量的数值设置为预设数值后，基于所述运动状态参量关系信息，得到所述纵向运动状态参量关系信息；
    基于所述纵向运动状态参量关系信息、以及所述运动状态参量关系信息，得到所述横向运动状态参量关系信息。
  17. 根据权利要求1-16任一项所述的行驶规划方法，其中，所述方法还包括：利用所述未来至少一个控制周期中的至少部分目标控制周期的目标运动状态，确定所述自动驾驶车辆在目标控制周期输出的动力参数。
  18. 一种行驶规划装置，包括：
    第一确定模块，配置为基于目标道路的参考轨迹信息、以及自动驾驶车辆的车辆参数，确定所述自动驾驶车辆在当前控制周期中的运动状态参量关系信息；
    第二确定模块，配置为基于运动状态参量关系信息，确定所述自动驾驶车辆的横向运动状态参量关系信息、以及纵向运动状态参量关系信息；
    第三确定模块，配置为基于预设约束条件、所述横向运动状态参量关系信息以及所述纵向运动状态参量关系信息，确定所述自动驾驶车辆在未来至少一个控制周期的目标运动状态。
  19. 一种计算机设备，包括：处理器和存储器，所述存储器存储有所述处理器可执行的机器可读指令，所述处理器用于执行所述存储器中存储的机器可读指令，所述机器可读指令被所述处理器执行时，所述处理器执行如权利要求1至17任一项所述的行驶规划方法的步骤。
  20. 一种计算机可读存储介质，所述计算机可读存储介质上存储有计算机程序，所述计算机程序被计算机设备运行时，所述计算机设备执行如权利要求1至17任一项所述的行驶规划方法的步骤。
  21. 一种计算机程序产品，所述计算机程序产品包括计算机可执行指令，该计算机可执行指令被执行后，能够实现权利要求1至17任一项所述的行驶规划方法的步骤。
PCT/CN2021/127434 2021-06-30 2021-10-29 一种行驶规划方法、装置、计算机设备及存储介质 WO2023273067A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110739069.6 2021-06-30
CN202110739069.6A CN113247023B (zh) 2021-06-30 2021-06-30 一种行驶规划方法、装置、计算机设备及存储介质

Publications (1)

Publication Number Publication Date
WO2023273067A1 true WO2023273067A1 (zh) 2023-01-05

Family

ID=77190499

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/127434 WO2023273067A1 (zh) 2021-06-30 2021-10-29 一种行驶规划方法、装置、计算机设备及存储介质

Country Status (2)

Country Link
CN (1) CN113247023B (zh)
WO (1) WO2023273067A1 (zh)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113247023B (zh) * 2021-06-30 2023-03-24 上海商汤临港智能科技有限公司 一种行驶规划方法、装置、计算机设备及存储介质
CN113734183B (zh) * 2021-09-17 2023-05-02 京东鲲鹏(江苏)科技有限公司 基于转向迟滞的车辆控制方法、装置、设备及存储介质
CN114475584B (zh) * 2022-02-28 2023-11-28 中国第一汽车股份有限公司 一种车辆的避障控制方法、装置、汽车及存储介质
CN114537442B (zh) * 2022-03-21 2022-09-23 中维建通信技术服务有限公司 移动智能车辆管理方法及***

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190004517A1 (en) * 2017-07-03 2019-01-03 Baidu Online Network Technology (Beijing) Co., Ltd. Method and apparatus for generating automatic driving strategy
CN111591306A (zh) * 2020-03-30 2020-08-28 浙江吉利汽车研究院有限公司 自动驾驶车辆的行驶轨迹规划方法、相关设备及存储介质
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
CN112389427A (zh) * 2021-01-19 2021-02-23 腾讯科技（深圳）有限公司 车辆轨迹优化方法、装置、电子设备和存储介质
CN113247023A (zh) * 2021-06-30 2021-08-13 上海商汤临港智能科技有限公司 一种行驶规划方法、装置、计算机设备及存储介质

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102017205040A1 (de) * 2017-03-24 2018-09-27 Bayerische Motoren Werke Aktiengesellschaft Verfahren zur Planung von kombinierter Längs- und Querbewegung für Fahrzeuge
CN110341711B (zh) * 2019-07-06 2021-10-22 深圳数翔科技有限公司 一种基于码头环境的行驶轨迹生成***及方法
CN110861650B (zh) * 2019-11-21 2021-04-16 驭势科技（北京）有限公司 车辆的路径规划方法、装置，车载设备和存储介质
CN111750887B (zh) * 2020-06-11 2023-11-21 上海交通大学 降低事故严重程度的无人驾驶车辆轨迹规划方法及***
CN111857152B (zh) * 2020-07-31 2024-07-19 北京京东乾石科技有限公司 用于生成车辆控制信息的方法和装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190004517A1 (en) * 2017-07-03 2019-01-03 Baidu Online Network Technology (Beijing) Co., Ltd. Method and apparatus for generating automatic driving strategy
CN111591306A (zh) * 2020-03-30 2020-08-28 浙江吉利汽车研究院有限公司 自动驾驶车辆的行驶轨迹规划方法、相关设备及存储介质
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
CN112389427A (zh) * 2021-01-19 2021-02-23 腾讯科技（深圳）有限公司 车辆轨迹优化方法、装置、电子设备和存储介质
CN113247023A (zh) * 2021-06-30 2021-08-13 上海商汤临港智能科技有限公司 一种行驶规划方法、装置、计算机设备及存储介质

Also Published As

Publication number Publication date
CN113247023A (zh) 2021-08-13
CN113247023B (zh) 2023-03-24

Similar Documents

Publication Publication Date Title
WO2023273067A1 (zh) 一种行驶规划方法、装置、计算机设备及存储介质
CN109375632B (zh) 自动驾驶车辆实时轨迹规划方法
Sun et al. Courteous autonomous cars
Andersen et al. Trajectory optimization for autonomous overtaking with visibility maximization
WO2018176593A1 (zh) 一种面向无人自行车的局部避障路径规划方法
CN112046484B (zh) 一种基于q学习的车辆变道超车路径规划方法
WO2019042295A1 (zh) 一种无人驾驶路径规划方法、***和装置
GB2570887A (en) A system for a vehicle
CN106371439B (zh) 一种统一的自动驾驶横向规划方法与***
Pérez et al. Autonomous driving manoeuvres in urban road traffic environment: a study on roundabouts
Bautista-Montesano et al. Autonomous navigation at unsignalized intersections: A coupled reinforcement learning and model predictive control approach
Vaskov et al. Guaranteed safe reachability-based trajectory design for a high-fidelity model of an autonomous passenger vehicle
CN112906542B (zh) 一种基于强化学习的无人车避障方法及装置
Guo et al. Toward human-like behavior generation in urban environment based on Markov decision process with hybrid potential maps
Li et al. Real-time local path planning for intelligent vehicle combining tentacle algorithm and B-spline curve
Zhang et al. An intelligent driver model with trajectory planning
Gong et al. Game theory-based decision-making and iterative predictive lateral control for cooperative obstacle avoidance of guided vehicle platoon
Kim et al. Trajectory planning for autonomous highway driving using the adaptive potential field
Liu et al. A model for safe lane changing of connected vehicles based on quintic polynomial Trajectory planning
Zhao et al. Intelligent Vehicle Path Planning Based on Q-Learning Algorithm with Consideration of Smoothness
Li et al. Decision making for autonomous vehicles
Han et al. Moving Horizon path planning for intelligent vehicle oriented to dynamic obstacle avoidance
CN113674529A (zh) 一种自主超车方法及***
Yoon et al. Social force aggregation control for autonomous driving with connected preview
Probst et al. Automated driving in complex real-world scenarios using a scalable risk-based behavior generation framework

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: 21947994

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE