WO2022052856A1 - 基于交通工具的数据处理方法、装置、计算机及存储介质 - Google Patents

基于交通工具的数据处理方法、装置、计算机及存储介质 Download PDF

Info

Publication number
WO2022052856A1
WO2022052856A1 PCT/CN2021/116193 CN2021116193W WO2022052856A1 WO 2022052856 A1 WO2022052856 A1 WO 2022052856A1 CN 2021116193 W CN2021116193 W CN 2021116193W WO 2022052856 A1 WO2022052856 A1 WO 2022052856A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
predicted
lane change
benefit
offset
Prior art date
Application number
PCT/CN2021/116193
Other languages
English (en)
French (fr)
Inventor
由长喜
Original Assignee
腾讯科技（深圳）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 腾讯科技（深圳）有限公司 filed Critical 腾讯科技（深圳）有限公司
Priority to JP2022565903A priority Critical patent/JP7520444B2/ja
Priority to EP21865910.0A priority patent/EP4119412A4/en
Publication of WO2022052856A1 publication Critical patent/WO2022052856A1/zh
Priority to US17/971,495 priority patent/US20230053459A1/en

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
    • B60W60/0027—Planning or execution of driving tasks using trajectory prediction for other traffic participants
    • 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
    • B60W30/095—Predicting travel path or likelihood of collision
    • B60W30/0956—Predicting travel path or likelihood of collision the prediction being responsive to traffic or environmental parameters
    • 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
    • B60W30/18—Propelling the vehicle
    • B60W30/18009—Propelling the vehicle related to particular drive situations
    • B60W30/18163—Lane change; Overtaking manoeuvres
    • 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
    • B60W40/02—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 ambient conditions
    • B60W40/04—Traffic conditions
    • 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
    • B60W50/0097—Predicting future conditions
    • 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
    • B60W60/00—Drive control systems specially adapted for autonomous road vehicles
    • B60W60/001—Planning or execution of driving tasks
    • B60W60/0015—Planning or execution of driving tasks specially adapted for safety
    • 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
    • B60W60/0027—Planning or execution of driving tasks using trajectory prediction for other traffic participants
    • B60W60/00274—Planning or execution of driving tasks using trajectory prediction for other traffic participants considering possible movement changes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/04—Inference or reasoning models
    • G06N5/045—Explanation of inference; Explainable artificial intelligence [XAI]; Interpretable artificial intelligence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G06V10/765—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects using rules for classification or partitioning the feature space
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
    • G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
    • 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
    • B60W2554/00—Input parameters relating to objects
    • B60W2554/40—Dynamic objects, e.g. animals, windblown objects
    • B60W2554/404—Characteristics
    • B60W2554/4041—Position
    • 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
    • B60W2554/00—Input parameters relating to objects
    • B60W2554/40—Dynamic objects, e.g. animals, windblown objects
    • B60W2554/404—Characteristics
    • B60W2554/4045—Intention, e.g. lane change or imminent movement
    • 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
    • B60W2554/00—Input parameters relating to objects
    • B60W2554/40—Dynamic objects, e.g. animals, windblown objects
    • B60W2554/404—Characteristics
    • B60W2554/4046—Behavior, e.g. aggressive or erratic
    • 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
    • B60W2554/00—Input parameters relating to objects
    • B60W2554/40—Dynamic objects, e.g. animals, windblown objects
    • B60W2554/406—Traffic density
    • 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
    • B60W2554/00—Input parameters relating to objects
    • B60W2554/80—Spatial relation or speed relative to objects
    • B60W2554/801—Lateral distance
    • 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
    • B60W2554/00—Input parameters relating to objects
    • B60W2554/80—Spatial relation or speed relative to objects
    • B60W2554/802—Longitudinal distance
    • 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
    • B60W2556/00—Input parameters relating to data
    • B60W2556/10—Historical data

Definitions

  • the present application relates to the field of computer technology, and in particular, to a vehicle-based data processing method, device, computer, and readable storage medium.
  • An autonomous vehicle also known as an unmanned vehicle or a computer-driven vehicle, is an intelligent vehicle that realizes unmanned driving through a computer system.
  • unmanned driving is generally divided into Level 0 (Level 0, L0) to Level 5 (Level 5, L5), that is, there is no automation to full automation.
  • the existing autonomous vehicle technology is generally based on the CT6 autopilot system of Cadillac or the Autopilot system of Tesla.
  • the self-vehicle unmanned vehicle
  • the self-vehicle needs to change lanes, to a certain extent, it needs the cooperation of the environmental vehicle, yielding, etc., the self-vehicle can obtain enough lane-changing space to change lanes.
  • the self-vehicle changes lanes.
  • the key is whether the environmental vehicle gives way to the own vehicle, that is, the own vehicle does not have the right of way. In actual driving, not all vehicles will give way to the self-vehicle based on the turn signal signal of the self-vehicle. Therefore, the self-vehicle is usually manually triggered by the driver to change lanes, or wait for the driver in the lane that needs to enter. The environmental vehicle gives way to the ego vehicle to realize the lane change of the ego vehicle.
  • the embodiments of the present application provide a vehicle-based data processing method, device, computer and readable storage medium, which can improve the efficiency of determining the lane change of the current vehicle.
  • One aspect of the embodiments of the present application provides a method for processing data based on a vehicle, the method comprising:
  • the second vehicle is a reference for the first vehicle when changing lanes means of transportation;
  • the first driving state and the second driving state determine the first lane change benefit of each predicted offset when the second vehicle is in the yield prediction state, and determine when the second vehicle is in the non-yield prediction state , the second lane change benefit of each predicted offset;
  • the predicted offset of the target lane change benefit of is determined as the target predicted offset; the target predicted offset is used to represent the predicted lateral lane change travel distance for the first vehicle.
  • An aspect of the embodiments of the present application provides a vehicle-based data processing device, the device comprising:
  • a state acquisition module for determining at least two predicted offsets of the first vehicle, a first driving state of the first vehicle, and a second driving state of the second vehicle; the second vehicle is the first vehicle in Vehicles referred to when changing lanes;
  • the benefit obtaining module is configured to determine, according to the first driving state and the second driving state, the first lane change benefit of each predicted offset when the second vehicle is in the yield prediction state, and determine the first lane change benefit of each predicted offset when the second vehicle is in the yield prediction state. In the non-yield prediction state, the second lane change benefit of each prediction offset;
  • the offset selection module is used to determine the predicted yield probability of the second vehicle, and generate the target of each predicted offset according to the predicted yield probability, the first lane change benefit and the second lane change benefit of each predicted offset For the lane change benefit, the predicted offset with the greatest target lane change benefit is determined as the target predicted offset; the target predicted offset is used to represent the predicted lateral lane change travel distance for the first vehicle.
  • An aspect of the embodiments of the present application provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and the computer program includes program instructions, and when the program instructions are executed by a processor, the program instructions in one aspect of the embodiments of the present application are executed.
  • Vehicle-based data processing methods
  • embodiments of the present application provide a computer program product or computer program, where the computer program product or computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium.
  • the processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the methods provided in the various implementations in the aspect of the embodiments of the present application.
  • FIG. 1 is a network architecture diagram of a vehicle-based data processing provided by an embodiment of the present application
  • FIGS. 2A to 2C are schematic diagrams of an application scenario provided by an embodiment of the present application.
  • FIG. 3 is a schematic diagram of a determination scenario for guiding a vehicle provided by an embodiment of the present application.
  • FIG. 4 is a flowchart of a method for data processing based on a vehicle provided by an embodiment of the present application
  • FIG. 5 is a schematic diagram of a driving decision-making scenario provided by an embodiment of the present application.
  • FIG. 6 is a schematic diagram of a scenario for determining a yield distance provided by an embodiment of the present application.
  • FIG. 7 is a flowchart of a specific method for data processing based on a vehicle provided by an embodiment of the present application.
  • FIG. 8 is a schematic diagram of a tree structure of a decision tree provided by an embodiment of the present application.
  • FIG. 9 is a schematic diagram of a vehicle-based data processing device provided by an embodiment of the present application.
  • FIG. 10 is a schematic structural diagram of a computer device provided by an embodiment of the present application.
  • the embodiments of the present application may be implemented by an automatic driving system in a vehicle.
  • the automatic driving system may include, but is not limited to, an algorithm end, a client end, and a cloud.
  • the algorithm side includes related algorithms for sensing, perception and decision-making
  • the client includes the robot operating system and hardware platform
  • the cloud can perform data storage, simulation, high-precision map drawing, and deep learning model training or prediction.
  • An autonomous driving system can be implemented by autonomous driving technology.
  • Autonomous driving technology usually includes high-precision maps, environmental perception, behavioral decision-making, path planning, motion control and other technologies.
  • Autonomous driving technology has broad application prospects.
  • the algorithm side is used to extract effective information from the raw data collected by the sensor to obtain the surrounding environment information of the ego car, and make decisions based on the surrounding environment information (such as what route to follow, what speed to drive or how to avoid obstacles, etc.).
  • Sensors used in existing autonomous driving systems generally include Global Positioning System (GPS)/Inertial Measurement Unit (IMU), Lidar (Light Detection and Ranging, LIDAR), cameras, radar and Sonar etc.
  • GPS Global Positioning System
  • IMU Inertial Measurement Unit
  • LIDAR Light Detection and Ranging
  • the algorithm side generally obtains relevant information such as the leading car (Leading car), the logical leading car (Putative Leader, PL) or the logical follower (Putative Follower, PF) of the own vehicle through sensors, etc.
  • a guide vehicle refers to a vehicle that appears or is about to appear in front of the vehicle and is close to the vehicle during the driving process of the vehicle. It can be used as a reference for the vehicle to drive.
  • the driving state between the ego vehicle and the logical follower PF can be coordinated.
  • the self-vehicle can obtain the vehicles existing in the roadway that the self-vehicle will enter through the sensor.
  • the self-vehicle obtains the vehicles A and B existing in the roadway to be entered, wherein the self-vehicle is to be inserted between the vehicles A and B to change lanes, and the vehicle A is located in front of the vehicle B , then the vehicle A is the logical lead vehicle PL of the own vehicle, and the vehicle B is the logical follower PF of the own vehicle. "Ahead" is based on the direction of travel of each vehicle.
  • the sensing part can acquire valid data from the sensing part, and perform positioning, object recognition, and object tracking of the second vehicle according to the valid data.
  • the decision-making part may include behavior prediction (eg, prediction of surrounding environment, prediction of subsequent operations of the first vehicle, etc.), path planning and obstacle avoidance mechanism of the first vehicle, and the like.
  • the embodiments of the present application include improvements to the decision-making part.
  • FIG. 1 is a network architecture diagram of a vehicle-based data processing provided by an embodiment of the present application.
  • the functions implemented by the embodiments of the present application can be applied to any vehicle with an automatic driving system, and the vehicle is recorded as the first vehicle.
  • the first vehicle may change lanes through the functions implemented in the embodiments of the present application.
  • the automatic driving system of the first vehicle (ie, self-vehicle) 101 may include a perception module, a prediction module, a decision module, etc.; the perception module here is used to implement the functions of the above-mentioned sensing part and sensing part.
  • the automatic driving system of the first vehicle 101 may also include a sensing module, a perception module, a prediction module, a decision-making module, and the like.
  • the first vehicle 101 detects other vehicles, such as the vehicle 102a, the vehicle 102b, or the vehicle 102c, through the sensing module.
  • the first vehicle 101 When the first vehicle 101 is about to change lanes, the first vehicle 101 obtains the vehicle in the roadway to be entered through the perception module, determines the second vehicle (ie, the logical follower PF), The vehicle 101 makes a decision to obtain at least two predicted offsets for the first vehicle 101 . Each predicted offset corresponds to a decision for the first vehicle 101 . By obtaining the benefit value of each predicted offset, the predicted offset corresponding to the largest benefit value can be selected as the target predicted offset, that is, the optimal decision of the first vehicle 101 is determined to determine the first transportation
  • the driving route of the tool 101 at the next moment realizes actively occupying the driving space of the second vehicle, forcing the second vehicle to give way to the vehicle, so that the vehicle can have the right of way to a certain extent when changing lanes. Improve the lane-changing efficiency of the first vehicle.
  • Artificial Intelligence is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.
  • artificial intelligence is a comprehensive technique of computer science that attempts to understand the essence of intelligence and produce a new kind of intelligent machine that can respond in a similar way to human intelligence.
  • Artificial intelligence is to study the design principles and implementation methods of various intelligent machines, so that the machines have the functions of perception, reasoning and decision-making.
  • Artificial intelligence technology is a comprehensive discipline, involving a wide range of fields, including both hardware-level technology and software-level technology.
  • the basic technologies of artificial intelligence generally include technologies such as sensors, special artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation/interaction systems, and mechatronics.
  • Artificial intelligence software technology mainly includes computer vision technology, speech processing technology, natural language processing technology, and machine learning/deep learning.
  • artificial intelligence technology has been researched and applied in many fields, such as common smart homes, smart wearable devices, virtual assistants, smart speakers, smart marketing, unmanned driving, autonomous driving, drones , robots, intelligent medical care, intelligent customer service, etc. It is believed that with the development of technology, artificial intelligence technology will be applied in more fields and play more and more important value. Among them, the embodiment of this application is artificial intelligence in automatic applications in the field of driving.
  • FIG. 2A to FIG. 2C are schematic diagrams of an application scenario provided by an embodiment of the present application.
  • the first vehicle ie, the vehicle
  • the roadway to be entered by the first vehicle 2011 is determined.
  • the vehicles that exist in 210 specifically determine the logical guide vehicle 2013 and the logical follower 2012 when the first vehicle 2011 changes lanes, wherein the first vehicle 2011 is to enter the logical guide vehicle 2013 and the logical follower 2012 On the roadway 210.
  • the automatic driving system in the first vehicle 2011 can determine the target predicted offset based on the embodiment of the present application, and drive based on the target predicted offset, so as to squeeze the driving space of the logical follower 2012, occupy a certain active road right, and improve the Lane change efficiency of the first vehicle 2011.
  • the ramp is an indispensable part of the interchange, and it is the road for the upper and lower intersections.
  • the main road that intersects is usually defined as the main line
  • the intersecting secondary road is defined as the lead
  • the line connecting the lead and the main line is called the ramp.
  • the first vehicle ie, the own vehicle
  • determine the logical leading vehicle 2023 and the logical follower 2022 of the first vehicle 2021 and adopt this
  • the method implemented by the application embodiment determines the target predicted offset, and drives based on the target predicted offset, so as to occupy the driving space of the logical follower 2022, occupy a certain active right of way, and improve the lane-changing efficiency of the first vehicle 2021 .
  • the vehicle that is located on the same roadway 203 as the first vehicle 2021 and that travels in front of the first vehicle 2021 is the guide vehicle 2024 of the first vehicle 2021 .
  • a guided vehicle may refer to a vehicle that is located on the same roadway as the self-vehicle and driving in front of the self-vehicle, and a logical guided vehicle may refer to a vehicle that travels in front of the self-vehicle in the roadway to be entered by the self-vehicle.
  • the logical guide vehicle 2033 and the logical follower 2032 of the first vehicle 2031 are located, the logical guide vehicle 2033 and the logical follower 2032 are inserted into the roadway.
  • the planning of the driving route of the first vehicle 2031 can be realized through the embodiments of the present application, and the lane changing efficiency of the first vehicle 2031 can be improved.
  • 2A to 2C merely enumerate several possible application scenarios applicable to the present application, and other lane changing scenarios may also apply the solutions implemented by the embodiments of the present application, which are not limited herein.
  • FIG. 3 is a schematic diagram of a determination scenario for guiding a vehicle according to an embodiment of the present application.
  • the traffic road where the first vehicle 3011 is located includes the first left, the second left, the third left and the fourth left, the roadway where the first vehicle 3011 is located is the second left, and the first traffic
  • the tool 3011 needs to change lanes from the second lane to the third lane to be inserted between the logical lead car 3012 and the logical follower 3013 .
  • This logical follower 3013 is recorded as the second vehicle.
  • the autonomous driving system in the first vehicle 3011 can make a decision on the first vehicle 3011 to determine at least two predicted offsets for the first vehicle 3011, the predicted offsets being used to indicate the predicted first vehicle 3011 The offset distance to the left three lanes (the lane to be changed to) at the next moment.
  • the at least two prediction offsets include n prediction offsets, such as prediction offsets 1, . . . and prediction offsets n, where n is a positive integer.
  • the automatic driving system determines the target lane change benefit of each predicted offset based on the driving states of the first vehicle 3011 and the second vehicle 3013, and obtains the target lane change benefit 1 of the predicted offset 1, ... and the predicted offset
  • the automatic driving system determines the predicted offset 3 corresponding to the target lane change benefit 3 as the target predicted offset.
  • the displacement determines the predicted offset trajectory 302 of the first vehicle 3011 , and controls the first vehicle 3011 to drive along the predicted offset trajectory 302 to increase the second vehicle 3013 , on the basis of the predicted offset trajectory 302 , for the first traffic
  • the probability of the tool 3011 yielding lanes increases the probability of the first vehicle 3011 successfully changing lanes to the third lane on the left, which improves the lane changing efficiency of the first vehicle 3011 .
  • FIG. 4 is a flowchart of a method for data processing based on a vehicle provided by an embodiment of the present application.
  • the vehicle-based data processing process includes the following steps:
  • Step S401 determining at least two predicted offsets of the first vehicle, a first driving state of the first vehicle, and a second driving state of the second vehicle.
  • the automatic driving system in the first vehicle determines the traffic information in the roadway to be entered by the self-vehicle to change lanes, and based on the Based on the determined traffic information, it is detected whether there is a logical follower PF of the first vehicle in the roadway.
  • the logical follower PF can be regarded as a vehicle that affects the lane change of the first vehicle, and the logical follower PF is recorded as the second vehicle. In other words, if the second vehicle does not yield to the first vehicle, the first vehicle cannot enter the roadway where the second vehicle is located.
  • the automatic driving system in the first vehicle After detecting the second vehicle, the automatic driving system in the first vehicle makes a decision on the offset when the first vehicle travels to the roadway where the second vehicle is located, and obtains at least two values of the first vehicle. a predicted offset, and a first driving state of the first vehicle and a second driving state of the second vehicle are determined. If the automatic driving system detects that there is no logical follower PF and logical lead vehicle PL of the first vehicle in the roadway to be entered by the vehicle, it controls the first vehicle to change lanes directly.
  • the automatic driving system detects that there is no logical follower PF of the first vehicle in the roadway to be entered by the self-vehicle, but there is a logical guide vehicle PL of the first vehicle, it controls the first vehicle to adjust and the logical guide vehicle After the distance between PLs in the longitudinal direction is adjusted, the first vehicle is controlled to change lanes.
  • the second vehicle is the vehicle that the first vehicle refers to when changing lanes.
  • the autonomous driving system can determine the lane width of the roadway and the number of decisions.
  • the number of decisions may be predetermined, and in this case, the automatic driving system may obtain the predetermined number of decisions from memory, for example.
  • the automatic driving system may also determine the lateral distance between the first vehicle and the lane line of the first roadway where the first vehicle is located, and determine at least two of the first vehicle based on the lane width, the lateral distance, and the number of decisions. Prediction offset.
  • the number of the at least two prediction offsets may be equal to the number of decisions.
  • FIG. 5 is a schematic diagram of a driving decision scenario provided by an embodiment of the present application. As shown in FIG.
  • the autonomous driving system of the first vehicle 501 determines a logical lead vehicle PL and a logical follower PF.
  • the logical follower PF is denoted as the second vehicle 502
  • the logical lead vehicle PL is denoted as the third vehicle 503
  • the lane width is denoted as lane_width
  • the number of decisions is assumed to be n.
  • the automated driving system determines a lateral distance 5041 between the first vehicle 501 and the lane line 511 of the first roadway 504 where the first vehicle 501 is located.
  • the lane line 511 is the common edge of the first lane 504 and the second lane where the second vehicle 502 is located.
  • the automatic driving system may determine n predicted offsets based on the lane width lane_width and the lateral distance 5041 .
  • the minimum of the n prediction offsets may be 0, and the maximum may be the lane width lane_width.
  • the autonomous driving system can determine n predicted offsets between 0 and the lane width lane_width.
  • the at least two predicted offsets may be divided into original lane offsets and lane change offsets, and the like.
  • the original lane offset refers to the predicted offset less than or equal to the lateral distance 5041
  • the lane change offset refers to the predicted offset greater than the lateral distance 5041 and less than or equal to the lane width lane_width.
  • the automatic driving system can determine the number n1 of predicted offsets included in the original lane offset and the number n2 of predicted offsets included in the lane change offset, based on the lateral distance 5041 and the original lane.
  • the number n1 of predicted offsets included in the offset determine n1 predicted offsets, and determine n2 predicted offsets based on the lateral distance 5041, the lane width lane_width, and the number n2 of predicted offsets included in the lane change offset. Shift amount, where n1 and n2 are both positive integers, and the sum of n1 and n2 is n.
  • the automatic driving system can determine the first roadway where the first vehicle is located, take the center line of the first roadway as the vertical axis of coordinates, and take the point where the first vehicle is mapped on the vertical axis of coordinates as the origin of coordinates. , take the normal corresponding to the vertical axis of the coordinates as the horizontal axis of the coordinates, and establish the road coordinate system according to the origin of the coordinates, the horizontal axis of the coordinates and the vertical axis of the coordinates.
  • the shortest distance between the first vehicle and the vertical axis of coordinates may be determined, and the point corresponding to the shortest distance on the vertical axis of coordinates may be used as the origin of coordinates, or,
  • the first vehicle is mapped on the vertical axis of coordinates, the mapping route from the first vehicle to the vertical axis of coordinates is perpendicular to the vertical axis of coordinates, and the point on the vertical axis of coordinates is mapped to the first vehicle, which is determined as the origin of coordinates.
  • the first driving state of the first vehicle may include, but is not limited to, first location information, a first driving speed, a first driving direction, and the like of the first vehicle.
  • the second driving state of the second vehicle may include, but is not limited to, the second position information, the second driving speed, the second driving direction, and the like of the second vehicle.
  • the first driving state of the first vehicle and the second driving state of the second vehicle may be determined by a perception module in the automatic driving system.
  • the automatic driving system can determine the first position information of the first vehicle in the road coordinate system, and determine the first driving state of the first vehicle according to the first position information; determine that the second vehicle is in the road coordinate system
  • the second position information of the second vehicle is determined, and the second driving state of the second vehicle is determined according to the second position information.
  • the first travel speed, the first travel direction, the second travel speed, the second travel direction, and the like may also be determined based on a road coordinate system.
  • the automatic driving system determines the coordinate origin O, the coordinate vertical axis S, and the coordinate horizontal axis D.
  • the coordinate origin O, the coordinate vertical axis S, and the coordinate horizontal axis D form a road coordinate system 505 .
  • the direction of the coordinate longitudinal axis S may be the traveling direction of the first vehicle 501 .
  • the first position information of the first vehicle 501 is (x1, y1). If the first vehicle 501 is at the coordinate origin O, then x1 is 0 and y1 is 0.
  • the first traveling speed may be directly determined by acquiring the speed of the first vehicle 501 (eg, the speed displayed on the instrument panel).
  • the first driving direction may be represented by a first unit vector.
  • the first unit vector is (0, 1)
  • it indicates that the first traveling direction is the traveling direction along the coordinate vertical axis S.
  • the first position information, the first traveling speed, and the first traveling direction constitute the first traveling state of the first vehicle 501 .
  • the second position information of the second vehicle 502 is (x2, y2), and the second traveling direction can be represented by a second unit vector.
  • the second position information, the second travel direction, the second travel speed, etc. are determined based on the road coordinate system 505 , wherein the second position information, the second travel direction, the second travel speed, etc. constitute the second vehicle 502 the second driving state.
  • Step S402 according to the first driving state and the second driving state, determine the first lane change benefit (payoff) of each predicted offset when the second vehicle is in the yield prediction state, and determine when the second vehicle is in the yield prediction state.
  • the first lane change benefit of each predicted offset determined by the automatic driving system in the first vehicle, and the second lane change benefit of each predicted offset can be represented by Table 1:
  • the table 1 includes at least two predicted offsets of the ego vehicle EV (ie, the first vehicle) and two predicted states of the second vehicle (ie, the logical follower PF).
  • the at least two prediction offsets include n prediction offsets, assuming n is 5.
  • the at least two prediction offsets can be denoted as represents the ith prediction offset.
  • the two predicted states of the second vehicle can be written as represents the j-th predicted state of the second vehicle.
  • the two prediction states of the second vehicle include a yield prediction state (Yield) and a non-yield prediction state (Not Yield).
  • a i,j can represent the j-th lane-changing benefit of the i-th predicted offset when the second vehicle is in the j-th predicted state, where i and j are positive integers, i is less than or equal to n, n
  • the number of prediction offsets included for the at least two prediction offsets, j is less than or equal to 2.
  • a 1,1 is used to represent the first lane change benefit of the first predicted offset (ie, predicted offset 1) when the second vehicle is in the yield prediction state.
  • S EV is used to represent the first driving state of the first vehicle
  • S PF is used to represent the second driving state of the second vehicle.
  • the first traveling state and the second traveling state are, for example, the first traveling state and the second traveling state at the current time.
  • the autonomous driving system may acquire a yield probability model.
  • the yield probability model may be predetermined. Input the first driving state, the second driving state and the i-th predicted offset into the yield probability model, and obtain the probability of the yield predicted state corresponding to the i-th predicted offset and the probability of the non-yield predicted state; i is a positive integer, i is less than or equal to the number of at least two prediction offsets.
  • the probability of the yield predicted state is determined as the first lane change benefit of the ith predicted offset when the second vehicle is in the yield predicted state.
  • the probability of the non-yield prediction state is determined as the second lane change benefit of the ith predicted offset when the second vehicle is in the non-yield prediction state.
  • Step S403 determine the predicted yield probability of the second vehicle, and generate the target lane change benefit of each predicted offset according to the predicted yield probability, the first lane change benefit and the second lane change benefit of each predicted offset, The predicted offset with the greatest target lane change benefit is determined as the target predicted offset.
  • the automatic driving system can obtain the historical longitudinal collision time and the historical collision distance corresponding to the historical longitudinal collision time, and determine the average collision time and the standard deviation of the collision time according to the historical longitudinal collision time and historical collision distance.
  • the historical longitudinal collision time and the historical collision distance corresponding to the historical longitudinal collision time may be longitudinal collision time samples of one or more vehicles and collision distance samples corresponding to the longitudinal collision time samples collected in advance in the testing process. For example, count the distribution law of the longitudinal collision time when vehicle A turns on the turn signal to change lanes many times, and the vehicle behind it gives way to vehicle A, and the corresponding longitudinal collision distance.
  • the historical longitudinal collision distance can be obtained by testing and data collection respectively in different collision distance intervals.
  • the historical collision distance can be used to divide the historical longitudinal collision time. For example, the statistical distance range is obtained, the historical collision distance within the statistical distance range is obtained, and the average collision time and the standard deviation of the collision time are determined according to the historical longitudinal collision time corresponding to the historical collision distance within the statistical distance range. If the collision distance between the vehicle 1 and the vehicle 2 falls within the statistical distance range, it means that the vehicle 1 and the vehicle 2 may be inserted between the two (ie, the vehicle 1 and the vehicle 2) to change lanes. lane change of vehicles.
  • Time to Collision also known as the estimated time to collision
  • TTC time to Collision
  • Crash time is also sometimes referred to as longitudinal crash time.
  • the automatic driving system can determine the longitudinal collision time between the second vehicle and the first vehicle, and map the longitudinal collision time to the first probability density function generated by the mean value of the collision time and the standard deviation of the collision time, so as to determine the collision time of the second vehicle.
  • Initial yield probability The initial yield probability can be expressed by formula 1:
  • p(Yield; ttc(EV, PF)) represents the initial yield probability of the second vehicle when the longitudinal collision time between the first vehicle and the second vehicle is given, and ttc(EV, PF) is used for represents the longitudinal collision time between the first vehicle and the second vehicle, ⁇ means proportional to, is the first probability density function.
  • the automatic driving system can obtain the historical traffic lateral distance, determine the distance mean and distance standard deviation according to the historical traffic lateral distance, record the mean distance as ⁇ dy , and record the distance standard deviation as ⁇ dy . Similar to the historical longitudinal collision time and the historical collision distance, the historical traffic lateral distance refers to the historical traffic lateral distance samples of one or more vehicles that are collected or acquired in advance.
  • the automatic driving system can also obtain the traffic lateral distance between the second vehicle and the first vehicle, and map the traffic lateral distance to the second probability density function generated by the mean distance and the distance standard deviation, so as to determine the second vehicle.
  • the driving hold probability may be used to represent the probability that the second vehicle will not yield to the first vehicle (ie, the probability that the second vehicle is in the second predicted state).
  • the driving hold probability can be expressed by formula 2:
  • ⁇ means proportional to, is the second probability density function.
  • the variance corresponding to each standard deviation can also be obtained.
  • the corresponding collision time variance, or the distance variance corresponding to the distance standard deviation where the standard deviation is the arithmetic square root of the corresponding variance. Whether the variance or the standard deviation is obtained, it does not affect the implementation of the embodiments of the present application, or in other words, does not affect the realization of the formula 1 and the formula 2, which will not be further described here.
  • the automatic driving system can determine the predicted yield probability of the second vehicle according to the initial yield probability and the driving hold probability.
  • the predicted yield probability can be expressed by formula 3:
  • Formula 3 is a formula for generating the predicted yield probability when the initial yield probability and the driving hold probability are independent events. If the initial yield probability and the driving hold probability are non-independent events, the formula for generating the predicted yield probability according to the initial yield probability and the driving hold probability can be determined based on the relationship between the initial yield probability and the driving hold probability.
  • the automatic driving system can have the largest target lane change benefit.
  • the predicted offset of is determined as a target predicted offset, wherein the target predicted offset is used to represent the predicted lateral lane change travel distance for the first vehicle.
  • step S402 when determining the first lane change benefit and the second lane change benefit of the first vehicle, the specific process is as follows:
  • each lane change benefit in Table 1 may be determined by multiple benefit parameters. Assuming that there are m benefit parameters, each lane change benefit can be determined by any one of the m benefit parameters or any h benefit parameters, where m is a positive integer and h is a positive integer less than or equal to m. For example, if m is 3, that is, there are 3 benefit parameters, each lane change benefit can be determined by 1 benefit parameter, or can be determined by any two benefit parameters, or can be determined by 3 benefit parameters It is determined by the parameters and is not limited here.
  • Can use Indicates the jth lane change benefit of the ith predicted offset when the second vehicle is determined to be in the jth predicted state according to the pth benefit parameter, where p is a positive integer, and p is less than or equal to m.
  • the automatic driving system may determine an offset distance of the first vehicle according to the first position information, where the offset distance is the first roadway where the first vehicle and the first vehicle are located The distance between the centerlines of the The first lane change benefit for each predicted offset, and the second lane change benefit for the i-th predicted offset when the second vehicle is in a non-yielding predicted state; i is a positive integer, i is less than or equal to The number of at least two prediction offsets.
  • the benefit parameter determined according to the difference between the ith predicted offset and the offset distance can be recorded as the first benefit parameter of each predicted offset when the second vehicle is in different predicted states.
  • the first benefit parameter of each predicted offset is determined as the first lane change benefit and the second lane change benefit of each predicted offset.
  • the first benefit parameter is used to represent the self-vehicle comfort, and the self-vehicle comfort can be expressed by formula 4:
  • the automatic driving system can obtain the first longitudinal coordinate value in the first position information and the second longitudinal coordinate value in the second position information, and according to the first longitudinal coordinate value, the first driving speed , the second longitudinal coordinate value, and the second traveling speed to determine the longitudinal collision time between the first vehicle and the second vehicle.
  • the automatic driving system can obtain the first lateral coordinate value in the first position information and the second lateral coordinate value in the second position information, and obtain the i-th predicted offset, according to the first lateral coordinate value and the second lateral coordinate value.
  • the coordinate value and the ith predicted offset determine the traffic lateral distance corresponding to the ith predicted offset; i is a positive integer, and i is less than or equal to the number of at least two predicted offsets.
  • the automatic driving system can also determine the first change of the i-th predicted offset when the second vehicle is in the yield prediction state according to the longitudinal collision time and the traffic lateral distance corresponding to the i-th predicted offset. lane benefit, and determining the second lane change benefit for the ith predicted offset when the second vehicle is in a non-yield predicted state.
  • the automatic driving system can also determine multiple benefit parameters, such as the second benefit parameter "lateral safety" and the third benefit parameter "collision" according to the longitudinal collision time and the traffic lateral distance corresponding to the i-th predicted offset.
  • first lane change benefit and the second lane change benefit can be determined directly according to one of the benefit parameters, or each benefit parameter can be randomly combined, according to any A set of benefit parameters determines the first lane change benefit and the second lane change benefit.
  • the second benefit parameter is used to represent lateral safety, which can be expressed by formula 5:
  • the second benefit parameter which can represent the lateral safety of the first vehicle at the i-th predicted offset and the second vehicle in the j-th predicted state, where ttc(EV, PF) is used to represent the longitudinal collision time between the first vehicle and the second vehicle, then it means that at the given i-th prediction offset In the case of , the traffic lateral spacing corresponding to the ith predicted offset.
  • Formula 5 can indicate that when the longitudinal collision time is less than 2 seconds and the traffic lateral distance of the ith predicted offset is less than 1 meter, the second benefit parameter of the ith predicted offset is determined to be -1; In this case (that is, otherwise), it can be determined that the second benefit parameter of the ith prediction offset is 0. If the first lane change benefit and the second lane change benefit are directly determined according to the second benefit parameter, they can be recorded as
  • the first time threshold for comparison with the longitudinal collision time and the first distance threshold for comparison with the lateral distance of traffic can be modified as required.
  • the first time threshold "2 seconds" and the first distance threshold "1 meter” in formula 5 are an empirical value, and may also be other values, which are not limited here.
  • the determined second benefit parameter of the ith predicted offset may also be performed on an as-needed basis. Change.
  • ttc(EV, PF) can be expressed by formula 6:
  • ttc(EV,PF) (l EV -l PF )/(v PF -v EV ), v PF >v EV 6
  • l EV represents the first longitudinal coordinate value of the first vehicle in the road coordinate system
  • l PF is used to represent the second longitudinal coordinate value of the second vehicle in the road coordinate system
  • v PF represents the second traveling speed of the second vehicle
  • v EV represents the first traveling speed of the first vehicle.
  • the third benefit parameter is used to represent the collision safety, which can be expressed by the formula 7:
  • formula 7 can indicate that when the longitudinal collision time is less than 2 seconds and the traffic lateral distance of the ith predicted offset is less than 0 meters, the third benefit parameter of the ith predicted offset is determined to be -10; In other cases, the third benefit parameter of the ith prediction offset can be determined to be 0. If the first lane change benefit and the second lane change benefit are determined directly according to the third benefit parameter, they can be recorded as
  • the second time threshold compared with the longitudinal collision time and the second distance threshold compared with the traffic lateral distance can be modified as needed, such as the second time threshold in formula 7 "2 seconds"
  • the second distance threshold "0 meters” is an empirical value, and may also be other values, which are not limited here.
  • the determined third benefit parameter of the ith predicted offset may also be performed on an as-needed basis. Change.
  • the fourth benefit parameter is used to represent the lane-changing reward
  • the lane-changing reward can be expressed by formula 8:
  • the fourth benefit parameter which can represent the lane-changing reward of the first vehicle at the i-th predicted offset and the second vehicle in the j-th predicted state, where ttc(EV, PF) and For the meaning of , please refer to the relevant description in formula 5.
  • Formula 8 can indicate that when the longitudinal collision time is greater than 3 seconds and the traffic lateral distance of the ith predicted offset is less than 0 meters, the fourth benefit parameter of the ith predicted offset is determined to be 5; in other cases , then it can be determined that the fourth benefit parameter of the ith prediction offset is 0. If the first lane change benefit and the second lane change benefit are determined directly according to the fourth benefit parameter, they can be recorded as
  • the third time threshold compared with the longitudinal collision time and the third distance threshold compared with the traffic lateral distance can be modified as required.
  • the third time threshold "3 seconds" and the third distance threshold "0 meters” in formula 8 are an empirical value, and can also be other values, which are not limited here.
  • the determined fourth benefit parameter of the ith predicted offset may also be based on Changes are required.
  • the automatic driving system may determine the third location information of the third vehicle, and determine the distance between the second vehicle and the third vehicle according to the second location information and the third location information. Guide the longitudinal distance.
  • the second vehicle and the third vehicle are on the same roadway, the third vehicle and the second vehicle are traveling in the same direction, and the third vehicle is in front of the second vehicle, which is in front of the second vehicle. based on the direction of travel.
  • the third vehicle can be considered as the logical lead vehicle PL of the first vehicle.
  • the automatic driving system may then acquire the third travel speed of the third vehicle.
  • the automatic driving system can determine the ith yield distance of the second vehicle, and based on the second driving speed, the third driving speed and the ith yield distance, determine the ith yield distance.
  • the first lane change benefit of the ith predicted offset when the second vehicle is in the yield prediction state, and the second lane change benefit of the ith predicted offset determined when the second vehicle is in the non-yield prediction state Lane change benefit; i is a positive integer, i is less than or equal to the number of at least two predicted offsets.
  • the i-th yield distance refers to the distance between the second vehicle and the third vehicle that is increased based on the guided longitudinal distance when the i-th predicted offset is in the yield prediction state.
  • FIG. 6 is a schematic diagram of a scene for determining a yield distance provided by an embodiment of the present application.
  • the automatic driving system of the first vehicle 6011 acquires the second longitudinal coordinate value of the second vehicle 6012 , the third position information of the third vehicle 6013 , and the third longitudinal coordinate in the third position information value, and according to the second longitudinal coordinate value and the third longitudinal coordinate value, the guiding longitudinal distance 602 between the second vehicle 6012 and the third vehicle 6013 is determined.
  • the automatic driving system obtains the ith predicted offset 603 , and determines the ith yield distance of the second vehicle based on the ith predicted offset 603 and the guidance longitudinal distance 602 .
  • the i-th yield distance refers to the guided longitudinal distance 602 between the second vehicle 6012 and the third vehicle 6013 when the i-th predicted offset 603 and the second vehicle 6012 is in the yield prediction state.
  • the increased distance 604 refers to the guided longitudinal distance 602 between the second vehicle 6012 and the third vehicle 6013 when the i-th predicted offset 60
  • the automatic driving system may determine a fifth benefit parameter based on the second travel speed, the third travel speed and the i-th yield distance, where the fifth benefit parameter is used to represent the traffic block cost.
  • the traffic congestion cost can be expressed by formula 9:
  • the first driving state S EV is used to represent the first driving state S EV , the ith predicted offset for a given first vehicle and the jth predicted state
  • the first driving state, the second driving state, the i-th predicted offset, the j-th predicted state, etc. may also be input into an Intelligent Driver Model (IDM), and based on the intelligent A driver model that determines the deceleration of the second vehicle.
  • IDM Intelligent Driver Model
  • Formula 9 indicates that when the deceleration of the second vehicle is less than -1, the fifth benefit parameter of the ith prediction offset is determined to be -5; in other cases (ie otherwise), the ith prediction can be determined
  • the fifth benefit parameter of the offset is 0. If the first lane change benefit and the second lane change benefit are determined directly according to the fifth benefit parameter, they can be recorded as
  • the first benefit parameter to the fifth benefit parameter above are examples of several benefit parameters (m is 5 at this time), and other benefit parameters can also be added as needed, which is not limited here.
  • the first lane change benefit and the second lane change benefit may be determined based on any one of m benefit parameters or any h benefit parameters. If the first lane change benefit and the second lane change benefit are determined based on the h benefit parameters, the weighted summation of the h benefit parameters under the i-th predicted offset and the j-th predicted state can be obtained to obtain the The ith predicted offset and the jth lane change benefit in the jth predicted state. For example, the first lane change benefit and the second lane change benefit are determined according to m benefit parameters, then the first lane change benefit and the second lane change benefit can be expressed by formula 10:
  • ⁇ p can be determined according to the importance of each benefit parameter. For example, if the importance of each benefit parameter is the same, no matter what the value of p is, ⁇ p is 1.
  • At least two predicted offsets of the first vehicle, the first driving state of the first vehicle, and the second driving state of the second vehicle are determined; according to the first driving state and the second driving state, determining a first lane change benefit for each predicted offset when the second vehicle is in a yield predicted state, and determining a second lane change benefit for each predicted offset when the second vehicle is in a non-yield predicted state Benefit; determine the predicted yield probability of the second vehicle, generate the target lane change benefit of each predicted offset according to the predicted yield probability, the first lane change benefit and the second lane change benefit of each predicted offset, The predicted offset with the greatest target lane change benefit is determined as the target predicted offset; the target predicted offset is used to represent the predicted lateral lane change travel distance for the first vehicle.
  • a decision is made on the first vehicle (that is, the self-vehicle), and the decision is used to represent the possible offset distance of the first vehicle (that is, at least two predicted offsets), and by obtaining the benefit value of each decision,
  • the driving of the first vehicle can be controlled based on the decision (target prediction offset), so that the first vehicle can have the right of way to a certain extent, and can be actively based on Decide to change lanes to improve the efficiency of lane changing of the first vehicle.
  • the automatic driving system can build a decision tree; the decision edge in the decision tree includes at least two prediction offsets, a yield prediction state and a non-yield prediction state; the tree nodes in the decision tree include the first vehicle and second means of transport.
  • the weighted sum of the first lane change benefit and the second lane change benefit of each predicted offset is performed layer by layer according to the predicted yield probability and the decision edge, until the decision tree is obtained.
  • Tree benefit value for each prediction offset in the root node The tree-shaped benefit value of each predicted offset in the root node is determined as the target lane change benefit of each predicted offset.
  • the data processing process may include the following steps:
  • Step S701 Determine the first driving state of the first vehicle and the second driving state of the second vehicle in the decision tree and each layer of the decision tree.
  • the automatic driving system may determine a decision tree. It is assumed that the decision tree includes T decision layers, and each decision layer includes a structure in which the first vehicle and the second vehicle alternate as tree nodes. That is, the parent node of the first vehicle may be the second vehicle, and the parent node of the second vehicle may be the first vehicle. Each decision layer includes at least two predicted offsets of the first vehicle, and a decision edge composed of the predicted state and non-yield state of the second vehicle. Among them, T is a positive integer, and T is the number of decision layers included in the decision tree.
  • the first vehicle corresponds to n decision edges, and each decision edge corresponds to a predicted offset of the first vehicle; the second vehicle corresponds to 2 decision edges, corresponding to the yield prediction state and the non-yield prediction state respectively.
  • the automatic driving system can acquire the first driving state of the first vehicle and the second driving state of the second vehicle in each layer of the decision tree.
  • the first driving state includes a first actual driving state and (T-1) first predicted driving states, wherein each first driving state corresponds to a decision level.
  • the second driving state includes a second actual driving state and (T-1) second predicted driving states, wherein each second driving state corresponds to a decision level.
  • the automatic driving system can obtain the first actual driving state and the second actual driving state, predict the first actual driving state, obtain the first predicted driving state of the second decision-making level, and predict the second actual driving state , obtain the second predicted driving state of the second decision-making layer; predict the first predicted driving state of the second decision-making layer to obtain the first predicted driving state of the third decision-making layer; 2. Predict the driving state to obtain the second predicted driving state of the third decision-making layer; ...; Predict the first predicted driving state of the (T-1)th decision-making layer to obtain the first predicted driving state of the T-th decision-making layer.
  • the driving state is predicted, and the second predicted driving state of the (T-1)th decision-making layer is predicted to obtain the second predicted driving state of the T-th decision-making layer.
  • Step S702 according to the first driving state and the second driving state of the kth floor, determine the first lane change benefit and the second lane change benefit of each predicted offset at the kth floor.
  • the automatic driving system may determine the first lane change benefit and the second lane change benefit of each predicted offset at the kth layer according to the first predicted driving state and the second predicted driving state of the kth layer , where k is a positive integer, and k is less than or equal to T. Among them, according to the first predicted driving state and the second predicted driving state of the kth layer, the automatic driving system determines the process of the first lane change benefit and the second lane change benefit of each predicted offset at the kth layer, as shown in Fig. The specific description shown in step S402 in step 4 will not be repeated here.
  • Step S703 determine the predicted yield probability of the second vehicle at the k-th floor, based on the predicted yield probability of the k-th floor and the parameter values of the (k+1)-th floor. Lane change benefit and second lane change benefit, generate the tree-shaped lane change benefit at the kth level for each predicted offset.
  • the automatic driving system may determine the predicted yield probability of the second vehicle at the kth layer.
  • the predicted yield probability of the second vehicle on the kth floor For the determination process of the predicted yield probability of the second vehicle on the kth floor, reference may be made to the specific description shown in step S403 in FIG. 4 .
  • ttc(EV, PF) in formula 1 represents the determined longitudinal collision time between the first vehicle and the second vehicle according to the first predicted driving state and the second predicted driving state of the kth layer.
  • the predicted yield probability of the second vehicle on the kth floor can be determined according to the first predicted driving state and the second predicted driving state of the kth floor.
  • the tree-shaped benefit value of each prediction offset in the kth layer can be shown in formula (1):
  • P(Yield) represents the predicted yield probability of the second vehicle at the kth floor
  • a i,1 is the first lane change of the i-th predicted offset when the second vehicle is in the yield prediction state
  • the benefit a i,2 is the second lane change benefit of the i-th predicted offset when the second vehicle is in the non-yield predicted state.
  • Tree benefit value used to represent the ith prediction offset in the kth layer.
  • the automatic driving system determines whether k is 1, and if k is 1, step S706 is performed, and if k is not 1, step S705 is performed.
  • Step S705 Determine the parameter value of the kth layer according to the tree-shaped lane change benefit of each predicted offset at the kth layer.
  • the automatic driving system can determine the parameter value of the kth layer according to the tree-shaped lane change benefit of each predicted offset at the kth layer, and the method of determining the parameter value of the kth layer can refer to the formula ( 2) as shown:
  • the preset parameter value is 0.
  • the parameter value of the k-th layer is the largest tree-shaped lane-change benefit among the tree-shaped lane-change benefits of the k-th layer for each prediction offset.
  • Step S706 determining the target lane change benefit of each predicted offset.
  • the automatic driving system can determine the tree-shaped benefit value of each predicted offset in the root node as the target variable benefit of each predicted offset, and the predicted offset with the largest target lane change benefit
  • the target prediction offset is determined as the target prediction offset, and the determination method of the target prediction offset can be shown in formula (3):
  • argmax(f(x)) is a variable point x (or a set of x) that maximizes f(x).
  • the predicted offset with the largest target lane change benefit and the largest value can be determined as the target prediction. Offset.
  • the automatic driving system can determine the predicted offset trajectory according to the target predicted offset and the second roadway where the second vehicle is located; the lateral travel distance corresponding to the predicted offset trajectory in the roadway is the target predicted offset.
  • the first vehicle is controlled to travel along the predicted offset trajectory.
  • a decision tree includes two decision layers, that is, at least two decision layers in the decision tree include a decision layer k1 and a decision layer k2, and the decision layer k1 includes a root node.
  • the first running state includes a first actual running state and a first predicted running state
  • the second running state includes a second actual running state and a second predicted running state.
  • the first lane change benefit includes the first lane change benefit of the decision layer k1 and the first lane change benefit of the decision layer k2
  • the second lane change benefit includes the second lane change benefit of the decision layer k1 and the second lane change benefit of the decision layer k2. benefit.
  • FIG. 8 FIG.
  • the root node of the decision tree is the second vehicle a1.
  • the root node a1 includes two decision edges, which represent the yield prediction state and the non-yield prediction state (no yield).
  • the two decision edges connect the first vehicle b21 and the first vehicle b22 respectively.
  • the first vehicle b21 corresponds to n decision edges, which represent the predicted offset 1, . . . and the predicted offset n, respectively.
  • n decision edges corresponding to the first vehicle b21 connect the second vehicle a31, . Vehicle a32.
  • the first vehicle b22 corresponds to n decision edges, which represent the predicted offsets 1, . Vehicle a34; ... until the leaf nodes of the decision tree are obtained, such as the second vehicle a51, the second vehicle a52, ... and the second vehicle a66, etc.
  • T is greater than 2
  • the automatic driving system can determine, according to the first actual driving state and the second actual driving state, the first lane change benefit of each predicted offset in the decision-making layer k1 when the second vehicle is in the yield prediction state, and determining the second lane change benefit of each predicted offset in the decision layer k1 when the second vehicle is in a non-yielding predicted state.
  • the first predicted traveling state of the first vehicle is predicted based on the first actual traveling state
  • the second predicted traveling state of the second vehicle is predicted based on the second actual traveling state.
  • the first lane change benefit of each predicted offset in the decision layer k2 is determined, and the second vehicle is determined In the non-yield prediction state, the second lane change benefit of each prediction offset in the decision layer k2.
  • the automatic driving system can obtain the predicted yield probability of the second vehicle in the decision layer k1 and the predicted yield probability in the decision layer k2. According to the predicted yield probability in the decision layer k2, the first lane change benefit and the second lane change benefit of each predicted offset in the decision layer k2 are weighted and summed, and the result of each predicted offset in the decision layer k2 is obtained. Tree benefit value. The maximum tree-shaped benefit value of each predicted offset in the tree-shaped benefit value of the decision-making layer k2 is determined as the parameter value of the decision-making layer k2.
  • the predicted yield probability in the decision layer k1 the parameter value of the decision layer k2, the first lane change benefit and the second lane change benefit of each predicted offset in the decision layer k1 are weighted and summed, and each prediction is obtained.
  • the predicted offset with the largest tree-shaped benefit value is determined as the target predicted offset
  • the predicted offset trajectory of the first vehicle is determined based on the target predicted offset
  • the first vehicle is controlled along the Predicting the offset trajectory driving to actively occupy the driving space of the second vehicle, so that the first vehicle can deviate as much as possible to the second roadway where the second vehicle is located, and is close to the second vehicle.
  • the lane-changing operation to the second carriageway is realized, and the lane-changing efficiency of the first vehicle is improved.
  • the first vehicle does not need to wait for the second vehicle to actively give way and then change lanes, but to offset as much as possible to the second lane where the second vehicle is located,
  • the driving space of the second vehicle is squeezed, so as to express the willingness to change lanes to the second vehicle more clearly, thereby improving the lane-changing efficiency of the first vehicle.
  • FIG. 9 is a schematic diagram of a vehicle-based data processing apparatus provided by an embodiment of the present application.
  • the vehicle-based data processing apparatus may be a computer program (including program code) running in a computer device, for example, the vehicle-based data processing apparatus is an application software; the apparatus may be used to execute the embodiments provided in the present application. corresponding steps in the method.
  • the vehicle-based data processing apparatus 900 can be used for the computer equipment in the embodiment corresponding to FIG. 4 .
  • the apparatus may include: a state acquisition module 11 , a benefit acquisition module 12 and an offset selection module 13 .
  • a state acquisition module 11 configured to determine at least two predicted offsets of the first vehicle, a first driving state of the first vehicle, and a second driving state of the second vehicle; the second vehicle is the first vehicle Vehicles referred to when changing lanes;
  • the benefit obtaining module 12 is configured to, according to the first driving state and the second driving state, determine the first lane change benefit of each predicted offset when the second vehicle is in the yield prediction state, and determine the first lane change benefit of the second vehicle The second lane change benefit of each forecast offset when in a non-yielding forecast state;
  • the offset selection module 13 is used to determine the predicted yield probability of the second vehicle, and generates the predicted yield of each predicted offset according to the predicted yield probability, the first lane change benefit and the second lane change benefit of each predicted offset. For the target lane change benefit, the predicted offset with the largest target lane change benefit is determined as the target predicted offset; the target predicted offset is used to represent the predicted lateral lane change travel distance for the first vehicle.
  • the state obtaining module 11 includes:
  • a decision obtaining unit 111 used to determine the lane width of the roadway and the number of decisions
  • the decision generating unit 112 is configured to determine the lateral distance between the first vehicle and the lane line of the first roadway where the first vehicle is located, and determine at least two of the first vehicle based on the lane width, the lateral distance and the number of decisions. prediction offsets; the number of at least two prediction offsets is the number of decisions.
  • the device 900 further includes:
  • the coordinate establishment module 14 is used to determine the first roadway where the first vehicle is located, and the center line of the first roadway is used as the coordinate vertical axis, and the point on which the first vehicle is mapped to the coordinate vertical axis is used as the coordinate origin , take the normal corresponding to the vertical axis of the coordinates as the horizontal axis of the coordinates, and establish the road coordinate system according to the origin of the coordinates, the horizontal axis of the coordinates and the vertical axis of the coordinates;
  • the state acquisition module 11 includes:
  • a first obtaining unit 113 configured to determine the first position information of the first vehicle in the road coordinate system, and determine the first driving state of the first vehicle according to the first position information
  • the second obtaining unit 114 is configured to determine the second position information of the second vehicle in the road coordinate system, and determine the second driving state of the second vehicle according to the second position information.
  • the first driving state includes first position information
  • the second driving state includes second position information
  • the benefit acquisition module 12 includes:
  • the distance determining unit 121 is configured to determine the offset distance of the first vehicle according to the first position information; the offset distance is the distance between the first vehicle and the center line of the first roadway where the first vehicle is located;
  • the benefit obtaining unit 122 is configured to obtain the ith predicted offset, and determine, according to the difference between the ith predicted offset and the offset distance, when the second vehicle is in the yield prediction state, the predicting the first lane change benefit of the offset, and determining the second lane change benefit of the ith predicted offset when the second vehicle is in a non-yield prediction state; i is a positive integer, i is less than or equal to at least The number of two prediction offsets.
  • the first driving state includes the first position information and the first driving speed
  • the second driving state includes the second position information and the second driving speed
  • the benefit acquisition module 12 includes:
  • the time acquisition unit 123 is configured to acquire the first longitudinal coordinate value in the first position information and the second longitudinal coordinate value in the second position information, according to the first longitudinal coordinate value, the first driving speed, the second longitudinal coordinate value and the second travel speed, to determine the longitudinal collision time between the first vehicle and the second vehicle;
  • the distance obtaining unit 124 is used to obtain the first horizontal coordinate value in the first position information and the second horizontal coordinate value in the second position information, and obtain the i-th predicted offset, according to the first horizontal coordinate value, the 2.
  • the horizontal coordinate value and the ith predicted offset determine the traffic lateral distance corresponding to the ith predicted offset; i is a positive integer, and i is less than or equal to the number of at least two predicted offsets;
  • the benefit obtaining unit 122 is further configured to determine, according to the longitudinal collision time and the traffic lateral distance corresponding to the i-th predicted offset, the ith predicted offset of the i-th predicted offset when the second vehicle is in the yield prediction state. a lane change benefit, and determining a second lane change benefit for the ith predicted offset when the second vehicle is in a non-yield predicted state.
  • the first driving state includes the first position information and the first driving speed
  • the second driving state includes the second position information and the second driving speed
  • the benefit acquisition module 12 includes:
  • the vehicle acquisition unit 125 is configured to determine the third position information of the third vehicle, and determine the guiding longitudinal distance between the second vehicle and the third vehicle according to the second position information and the third position information; the second vehicle The third vehicle is on the same roadway as the third vehicle, and the third vehicle and the second vehicle travel in the same direction;
  • a speed obtaining unit 126 configured to determine a third travel speed of the third vehicle
  • the benefit obtaining unit 122 is further configured to determine the ith yield distance of the second vehicle according to the ith predicted offset and the guiding longitudinal distance, based on the second travel speed, the third travel speed and the ith yield distance Line distance, determining the first lane change benefit of the i-th predicted offset when the second vehicle is in a yield prediction state, and determining the i-th prediction offset when the second vehicle is in a non-yield prediction state
  • the second lane change benefit of the shift amount; i is a positive integer, i is less than or equal to the number of at least two predicted offsets; the i-th yield distance refers to the i-th predicted offset and the second vehicle The increased distance between the second vehicle and the third vehicle based on the guided longitudinal distance when in the yield prediction state.
  • the benefit acquisition module 12 includes:
  • the module obtaining unit 127 is used to obtain the yield probability model
  • the probability prediction unit 128 is configured to input the first driving state, the second driving state and the i-th predicted offset into the yield probability model, and obtain the probability and non-existence of the yield predicted state corresponding to the i-th predicted offset.
  • the probability of yielding the predicted state; i is a positive integer, i is less than or equal to the number of at least two predicted offsets;
  • the benefit obtaining unit 122 is further configured to determine the probability of the yield prediction state as the first lane change benefit of the i-th predicted offset when the second vehicle is in the yield prediction state;
  • the benefit obtaining unit 122 is further configured to determine the probability of the non-yield predicted state as the second lane change benefit of the i-th predicted offset when the second vehicle is in the non-yield predicted state.
  • the offset selection module 13 includes:
  • the collision obtaining unit 131 is used to obtain the historical longitudinal collision time and the historical collision distance corresponding to the historical longitudinal collision time, and determine the mean value of the collision time and the standard deviation of the collision time according to the historical longitudinal collision time and the historical collision distance;
  • the first probability obtaining unit 132 is configured to determine the longitudinal collision time between the second vehicle and the first vehicle, and map the longitudinal collision time to the first probability density function generated by the collision time mean value and the collision time standard deviation to determine The initial yield probability of the second means of transport;
  • the historical distance obtaining unit 133 is used to obtain the historical traffic lateral distance, and determine the distance mean value and the distance standard deviation according to the historical traffic lateral distance;
  • the second probability obtaining unit 134 is configured to obtain the traffic lateral distance between the second vehicle and the first vehicle, and map the traffic lateral distance to the second probability density function generated by the mean distance and the distance standard deviation, so that the The second probability density function determines the travel retention probability of the second vehicle;
  • the yield probability determination unit 135 is configured to determine the predicted yield probability of the second vehicle according to the initial yield probability and the driving hold probability.
  • the offset selection module 13 includes:
  • the tree obtaining unit 136 is used to construct a decision tree; the decision edge in the decision tree includes at least two prediction offsets, a yield prediction state and a non-yield prediction state; the tree nodes in the decision tree include the first vehicle and the first vehicle 2. means of transportation;
  • the benefit iteration unit 137 is configured to, in at least two decision layers of the decision tree, perform a weighted calculation of the first lane change benefit and the second lane change benefit of each predicted offset according to the predicted yield probability and the decision edge layer by layer. and, until the tree benefit value of each prediction offset in the root node of the decision tree is obtained;
  • the benefit determination unit 138 is configured to determine the tree-shaped benefit value of each predicted offset in the root node as the target lane change benefit of each predicted offset.
  • At least two decision layers in the decision tree include decision layer k1 and decision layer k2, and decision layer k1 includes the root node;
  • the first driving state includes the first actual driving state and the first predicted driving state, and the second driving state includes the first driving state.
  • the first lane-changing benefit includes the first lane-changing benefit in the decision-making level k1 and the first lane-changing benefit in the decision-making level k2, and the second lane-changing benefit includes the first lane-changing benefit in the decision-making level k1
  • the benefit acquisition module 12 includes:
  • the actual processing unit 129a determines the first lane change benefit of each predicted offset in the decision-making layer k1 when the second vehicle is in the yield prediction state, and determines When the second vehicle is in the non-yield prediction state, the second lane change benefit of each prediction offset in the decision-making layer k1;
  • a state prediction unit 129b configured to predict a first predicted travel state of the first vehicle according to the first actual travel state, and predict a second predicted travel state of the second vehicle according to the second actual travel state;
  • the prediction processing unit 129c is configured to determine, according to the first predicted travel state and the second predicted travel state, the first lane change benefit of each predicted offset in the decision layer k2 when the second vehicle is in the yield predicted state, and Determine the second lane change benefit of each predicted offset in decision layer k2 when the second vehicle is in a non-yield predicted state.
  • the predicted yield probability includes the predicted yield probability in the decision layer k1 and the predicted yield probability in the decision layer k2;
  • the benefit iteration unit 137 includes:
  • the layer benefit obtaining subunit 1371 is used to perform a weighted sum of the first lane change benefit and the second lane change benefit of each predicted offset in the decision layer k2 according to the predicted yield probability in the decision layer k2, to obtain Tree-shaped benefit value of each prediction offset at decision level k2;
  • the parameter determination subunit 1372 is used to determine the maximum tree-shaped benefit value of each predicted offset in the tree-shaped benefit value of the decision-making layer k2 as the parameter value of the decision-making layer k2;
  • the benefit acquisition subunit 1371 of this layer is further configured to, according to the predicted yield probability in the decision layer k1, determine the parameter value of the decision layer k2 and the first lane change benefit and the second lane change benefit of each predicted offset in the decision layer k1.
  • the lane-changing benefits are weighted and summed to obtain the tree-shaped benefit value of each predicted offset in the decision-making layer k1; the tree-shaped benefit value of each predicted offset in the decision-making layer k1 is the predicted offset in the root node of the decision tree.
  • the amount of tree benefit value is further configured to, according to the predicted yield probability in the decision layer k1, determine the parameter value of the decision layer k2 and the first lane change benefit and the second lane change benefit of each predicted offset in the decision layer k1.
  • the lane-changing benefits are weighted and summed to obtain the tree-shaped benefit value of each predicted offset in the decision-making layer k1; the tree-shaped benefit value of each predicted offset in the decision-making layer k1
  • the device 900 further includes:
  • the trajectory determination module 15 is used to determine the predicted offset trajectory according to the target predicted offset amount; the lateral travel distance corresponding to the predicted offset trajectory in the roadway is the target predicted offset amount;
  • the travel control module 16 is configured to control the first vehicle to travel along the predicted deviation trajectory.
  • An embodiment of the present application provides a vehicle-based data processing device, the device obtains the first driving state of the first vehicle by acquiring at least two predicted offsets of the first vehicle, and obtains the the second driving state; the second vehicle is the vehicle that the first vehicle refers to when changing lanes; according to the first driving state and the second driving state, determine when the second vehicle is in the yield prediction state, Predict the first lane change benefit of the offset, and determine the second lane change benefit of each predicted offset when the second vehicle is in a non-yield prediction state; obtain the predicted yield probability of the second vehicle, according to Predict the yield probability, the first lane change benefit and the second lane change benefit of each predicted offset, generate the target lane change benefit of each predicted offset, and determine the predicted offset with the largest target lane change benefit as Target predicted offset; the target predicted offset is used to represent the predicted lateral lane change distance for the first vehicle.
  • a decision is made on the first vehicle (that is, the self-vehicle), and the decision is used to represent the possible offset distance of the first vehicle (that is, at least two predicted offsets), and by obtaining the benefit value of each decision,
  • the driving of the first vehicle can be controlled based on the decision (target prediction offset), so that the first vehicle can have the right of way to a certain extent, and can be actively based on Decide to change lanes to improve the efficiency of lane changing of the first vehicle.
  • FIG. 10 is a schematic structural diagram of a computer device provided by an embodiment of the present application.
  • the computer device in this embodiment of the present application may include: one or more processors 1001 , a memory 1002 , and an input-output interface 1003 .
  • the processor 1001 , the memory 1002 and the input/output interface 1003 are connected through a bus 1004 .
  • the memory 1002 is used to store a computer program, the computer program includes program instructions, and the input and output interface 1003 is used to receive data and output data; the processor 1001 is used to execute the program instructions stored in the memory 1002, and perform the following operations:
  • the second vehicle is a reference for the first vehicle when changing lanes means of transportation;
  • the first driving state and the second driving state determine the first lane change benefit of each predicted offset when the second vehicle is in the yield prediction state, and determine when the second vehicle is in the non-yield prediction state , the second lane change benefit of each predicted offset;
  • the predicted offset of the target lane change benefit of is determined as the target predicted offset; the target predicted offset is used to represent the predicted lateral lane change travel distance for the first vehicle.
  • the processor 1001 may be a central processing unit (central processing unit, CPU), the processor may also be other general-purpose processors, digital signal processors (digital signal processors, DSP), dedicated integrated Circuit (application specific integrated circuit, ASIC), off-the-shelf programmable gate array (field-programmable gate array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
  • a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
  • the memory 1002 may include read-only memory and random access memory, and provides instructions and data to the processor 1001 and the input-output interface 1003 .
  • a portion of memory 1002 may also include non-volatile random access memory.
  • the memory 1002 may also store information of device type.
  • the computer device can execute the implementation manner provided by each step in FIG. 4 through its built-in functional modules.
  • the implementation manner provided by each step in FIG. 4 can be executed through its built-in functional modules.
  • the embodiment of the present application provides a computer device, including: a processor, an input and output interface, and a memory, and obtains computer instructions in the memory through the processor, executes each step of the method shown in FIG. Handling operations.
  • the embodiment of the present application realizes the determination of at least two predicted offsets of the first vehicle, the first driving state of the first vehicle, and the second driving state of the second vehicle; the second vehicle is the first vehicle in The vehicle that is referenced when changing lanes; according to the first driving state and the second driving state, determine the first lane-changing benefit of each predicted offset when the second vehicle is in the yield prediction state, and determine the first lane-changing benefit of each predicted offset.
  • the second lane change benefit of each predicted offset determines the predicted yield probability of the second vehicle, according to the predicted yield probability and the first lane change of each predicted offset Benefit and second lane change benefit, generate the target lane change benefit of each predicted offset, and determine the predicted offset with the largest target lane change benefit as the target predicted offset; the target predicted offset is used to represent the target The predicted lateral lane change travel distance of the first vehicle.
  • a decision is made on the first vehicle (that is, the self-vehicle), and the decision is used to represent the possible offset distance of the first vehicle (that is, at least two predicted offsets), and by obtaining the benefit value of each decision,
  • the driving of the first vehicle can be controlled based on the decision (target prediction offset), so that the first vehicle can have the right of way to a certain extent, and can be actively based on Decide to change lanes to improve the efficiency of lane changing of the first vehicle.
  • Embodiments of the present application further provide a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and the computer program includes program instructions.
  • the program instructions When the program instructions are executed by the processor, each step in FIG. 4 can be implemented.
  • the implementation manner provided by each step in FIG. 4 which will not be repeated here.
  • the description of the beneficial effects of using the same method will not be repeated.
  • program instructions may be deployed to execute on one computer device, or on multiple computer devices located at one site, or alternatively, on multiple computer devices distributed across multiple sites and interconnected by a communications network implement.
  • the computer-readable storage medium may be the vehicle-based data processing apparatus provided in any of the foregoing embodiments or an internal storage unit of the computer device, such as a hard disk or a memory of the computer device.
  • the computer-readable storage medium can also be an external storage device of the computer device, such as a plug-in hard disk, a smart media card (smart media card, SMC), a secure digital (secure digital, SD) card equipped on the computer device, Flash card (flash card), etc.
  • the computer-readable storage medium may also include both an internal storage unit of the computer device and an external storage device.
  • the computer-readable storage medium is used to store the computer program and other programs and data required by the computer device.
  • the computer-readable storage medium can also be used to temporarily store data that has been or will be output.
  • Embodiments of the present application also provide a computer program product or computer program, where the computer program product or computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium.
  • the processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method provided in the various exemplary manners in FIG.
  • the determination of the target predicted offset is to control the first vehicle to travel based on the predicted offset trajectory determined by the target predicted offset, so as to improve the lane changing efficiency of the first vehicle.
  • each process and/or the schematic structural diagrams of the method flowcharts and/or structural schematic diagrams can be implemented by computer program instructions. or blocks, and combinations of processes and/or blocks in flowcharts and/or block diagrams.
  • These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor or other programmable data processing device to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing device produce a function
  • These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory result in an article of manufacture comprising instruction means, the instructions
  • the apparatus implements the functions specified in one or more of the flowcharts and/or one or more blocks of the structural diagram.
  • These computer program instructions can also be loaded on a computer or other programmable data processing device to cause a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process such that The instructions provide steps for implementing the functions specified in the block or blocks of the flowchart and/or structural representation.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Transportation (AREA)
  • Automation & Control Theory (AREA)
  • Mechanical Engineering (AREA)
  • Physics & Mathematics (AREA)
  • Human Computer Interaction (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Mathematical Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Computing Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computational Linguistics (AREA)
  • Medical Informatics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Multimedia (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Traffic Control Systems (AREA)
  • Control Of Driving Devices And Active Controlling Of Vehicle (AREA)

Abstract

一种基于交通工具的数据处理方法、装置、计算机及可读存储介质，涉及自动驾驶技术，该方法包括：确定第一交通工具的至少两个预测偏移量、第一交通工具的第一行驶状态以及第二交通工具的第二行驶状态（S401）；根据第一行驶状态及第二行驶状态，确定各个预测偏移量在第二交通工具处于让行预测状态时的第一变道效益，以及在第二交通工具处于非让行预测状态时的第二变道效益（S402）；确定第二交通工具的预测让行概率，根据预测让行概率、各个预测偏移量的第一变道效益和第二变道效益，生成各个预测偏移量的目标变道效益，将具有最大的目标变道效益的预测偏移量确定为目标预测偏移量（S403）。

Description

基于交通工具的数据处理方法、装置、计算机及存储介质
本申请要求2020年09月10日提交的申请号为202010947834.9、发明名称为“基于交通工具的数据处理方法、装置、计算机及存储介质”的中国专利申请的优先权。
技术领域
本申请涉及计算机技术领域，尤其涉及一种基于交通工具的数据处理方法、装置、计算机及可读存储介质。
背景技术
自动驾驶车辆又称无人驾驶车辆或电脑驾驶车辆等，是一种通过电脑***实现无人驾驶的智能车辆。随着科技的发展，针对自动驾驶车辆的研发也越来越广泛，其中，无人驾驶一般划分为第0级(Level 0，L0)至第5级(Level 5，L5)，即从无自动化到完全自动化。其中，现有的自动驾驶车辆技术一般是基于凯迪拉克的CT6自动驾驶***或特斯拉(Tesla)的自动驾驶(Autopilot)***等。当自车(无人驾驶车辆)要进行变道时，在一定程度上，需要环境车辆的配合、让行等，自车才能获得足够的变道空间进行变道，因此，自车变道的关键在于环境车辆是否对自车让行，即自车不具备路权。而在实际驾驶中，并不是所有车辆都会基于自车的转向灯信号，为自车让行，故而，自车一般是由驾驶员手动触发变道，或者，等待需要驶入的车行道中的环境车辆为自车让行，以实现自车的变道。
发明内容
本申请实施例提供了一种基于交通工具的数据处理方法、装置、计算机及可读存储介质，可以提高确定当前交通工具的变道效率。
本申请实施例一方面提供了一种基于交通工具的数据处理方法，该方法包括：
确定第一交通工具的至少两个预测偏移量、第一交通工具的第一行驶状态以及第二交通工具的第二行驶状态；第二交通工具为第一交通工具在变换车行道时参考的交通工具；
根据第一行驶状态及第二行驶状态，确定在第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，各个预测偏移量的第二变道效益；
确定第二交通工具的预测让行概率，根据预测让行概率、各个预测偏移量的第一变道效益和第二变道效益，生成各个预测偏移量的目标变道效益，将具有最大的目标变道效益的预测偏移量确定为目标预测偏移量；目标预测偏移量用于表示针对第一交通工具所预测的侧向变道行驶距离。
本申请实施例一方面提供了一种基于交通工具的数据处理装置，该装置包括：
状态获取模块，用于确定第一交通工具的至少两个预测偏移量、第一交通工具的第一行驶状态以及第二交通工具的第二行驶状态；第二交通工具为第一交通工具在变换车行道时参考的交通工具；
效益获取模块，用于根据第一行驶状态及第二行驶状态，确定在第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，各个预测偏移量的第二变道效益；
偏移选取模块，用于确定第二交通工具的预测让行概率，根据预测让行概率、各个预测偏移量的第一变道效益和第二变道效益，生成各个预测偏移量的目标变道效益，将具有最大的目标变道效益的预测偏移量确定为目标预测偏移量；目标预测偏移量用于表示针对第一交通工具所预测的侧向变道行驶距离。
本申请实施例一方面提供了一种计算机可读存储介质，计算机可读存储介质存储有计算机程序，计算机程序包括程序指令，程序指令当被处理器执行时，执行本申请实施例一方面中的基于交通工具的数据处理方法。
本申请实施例一方面提供了一种计算机程序产品或计算机程序，该计算机程序产品或计算机程序包括计算机指令，该计算机指令存储在计算机可读存储介质中。计算机设备的处理器从计算机可读存储介质读取该计算机指令，处理器执行该计算机指令，使得该计算机设备执行本申请实施例一方面中的各种实施方式中提供的方法。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，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的附图。
图1是本申请实施例提供的一种基于交通工具的数据处理的网络架构图；
图2A至图2C是本申请实施例提供的一种应用场景示意图；
图3是本申请实施例提供的一种引导交通工具的确定场景示意图；
图4是本申请实施例提供的一种基于交通工具的数据处理的方法流程图；
图5是本申请实施例提供了一种行驶决策场景示意图；
图6是本申请实施例提供的一种让行距离确定场景示意图；
图7是本申请实施例提供的一种基于交通工具的数据处理的具体方法流程图；
图8是本申请实施例提供的一种决策树的树形结构示意图；
图9是本申请实施例提供的一种基于交通工具的数据处理装置示意图；
图10是本申请实施例提供的一种计算机设备的结构示意图。
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清楚、完整地描述，显然，所描述的实施例仅仅是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
本申请实施例可以由交通工具中的自动驾驶***实现。该自动驾驶***可以包括但不限于算法端、客户端及云端。其中，算法端包括面向传感、感知及决策等的相关算法，客户端包括机器人操作***及硬件平台，云端则可以进行数据存储、模拟、高精度地图绘制及深度学习模型训练或预测等。
自动驾驶***可以是由自动驾驶技术所实现的。自动驾驶技术通常包括高精地图、环境感知、行为决策、路径规划、运动控制等技术。自动驾驶技术有着广泛的应用前景。
算法端用于从传感器采集到的原始数据中提取有效信息，以获取自车(Ego car)的周围环境信息，并基于周围环境信息作出决策(如，沿什么路线行驶，以什么速度行驶或如何躲避障碍物等)。现有的自动驾驶***中所使用的传感器一般包括全球定位***(Global Positioning System，GPS)/惯性测量单元(Inertial measurement unit，IMU)、激光雷达(Light Detection and Ranging，LIDAR)、摄像头、雷达及声呐等。算法端一般通过传感器等获取自车的引导车(Leading car)、逻辑引导车(Putative Leader，PL)或逻辑追随者(Putative Follower，PF)等的相关信息。引导车通俗来讲是指自车行驶过程中，出现或即将出现在自车前方、且与自车较近的车辆，可以作为自车行驶的一个参考物。在本申请中，可以对自车与逻辑追随者PF 之间的行驶状态进行协调。自车可以通过传感器获取自车所要进入的车行道中存在的交通工具。例如，自车获取到所要进入的车行道中存在的交通工具A及交通工具B，其中，自车要***交通工具A与交通工具B之间进行变道，交通工具A位于交通工具B的前方，则该交通工具A为自车的逻辑引导车PL，交通工具B为自车的逻辑追随者PF。“前方”是基于各个交通工具的行驶方向而言的。
感知部分可以从传感部分获取有效数据，根据该有效数据对第二交通工具进行定位、物体识别及物体追踪等。决策部分可以包括行为预测(如对周围环境的预测，对第一交通工具后续操作的预测等)、对第一交通工具的路径规划及避障机制等。本申请实施例包括对决策部分的改进。
具体地，请参见图1，图1是本申请实施例提供的一种基于交通工具的数据处理的网络架构图。本申请实施例所实现的功能可以应用于任意一个具有自动驾驶***的交通工具，将该交通工具记作第一交通工具。当第一交通工具确定要进行变道，而第一交通工具获取到逻辑追随者PF时，第一交通工具可以通过本申请实施例所实现的功能，进行变道。
如图1所示，第一交通工具(即自车)101的自动驾驶***可以包括感知模块、预测模块及决策模块等；此处的感知模块用于实现上述传感部分及感知部分的功能。第一交通工具101的自动驾驶***也可以包括传感模块、感知模块、预测模块及决策模块等。举例来说，第一交通工具101通过感知模块检测其他交通工具，如交通工具102a、交通工具102b或交通工具102c等。当第一交通工具101要进行变道时，第一交通工具101通过感知模块获取要进入的车行道中的交通工具，从中确定第二交通工具(即逻辑追随者PF)，并对该第一交通工具101进行决策，得到第一交通工具101的至少两个预测偏移量。每个预测偏移量相当于该第一交通工具101的一个决策。通过获取各个预测偏移量的效益值，可以选取最大的效益值对应的预测偏移量，作为目标预测偏移量，即确定该第一交通工具101的最优决策，以确定该第一交通工具101在下一时刻的行驶路线，实现主动挤占第二交通工具的行车空间，迫使第二交通工具对自车进行让行，使得自车进行变道时，在一定程度上可以拥有路权，从而提高第一交通工具的变道效率。
本申请实施例可以使用人工智能技术。人工智能(Artificial Intelligence，AI)是利用数字计算机或者数字计算机控制的机器模拟、延伸和扩展人的智能，感知环境、获取知识并使用知识获得最佳结果的理论、方法、技术及应用***。换句话说，人工智能是计算机科学的一个综合技术，它企图了解智能的实质，并生产出一种新的能以人类智能相似的方式做出反应的智能机器。人工智能也就是研究各种智能机器的设计原理与实现方法，使机器具有感知、推理与决策的功能。
人工智能技术是一门综合学科，涉及领域广泛，既有硬件层面的技术也有软件层面的技术。人工智能基础技术一般包括如传感器、专用人工智能芯片、云计算、分布式存储、大数据处理技术、操作/交互***、机电一体化等技术。人工智能软件技术主要包括计算机视觉技术、语音处理技术、自然语言处理技术以及机器学习/深度学习等几大方向。
随着人工智能技术研究和进步，人工智能技术在多个领域展开研究和应用，例如常见的智能家居、智能穿戴设备、虚拟助理、智能音箱、智能营销、无人驾驶、自动驾驶、无人机、机器人、智能医疗、智能客服等，相信随着技术的发展，人工智能技术将在更多的领域得到应用，并发挥越来越重要的价值，其中，本申请实施例则是人工智能在自动驾驶领域的应用。
本申请实施例可以用于自动驾驶车辆的任意一个变道场景中。举例来说，可以参见图2A至图2C，图2A至图2C是本申请实施例提供的一种应用场景示意图。如图2A所示，当第一交通工具(即自车)2011在匝道200中行驶并且要进入车行道210时(on-ramp merge)，确定该第一交通工具2011所要进入的车行道210中存在的交通工具，具体是确定该第一交通工具2011在变道时的逻辑引导车2013及逻辑追随者2012等，其中，第一交通工具2011要进入逻辑引导车2013及逻辑追随者2012所在的车行道210。第一 交通工具2011中的自动驾驶***可以基于本申请实施例，确定目标预测偏移量，基于目标预测偏移量行驶，以挤占逻辑追随者2012的行车空间，占据一定的主动路权，提高第一交通工具2011的变道效率。其中，匝道是互通式立体交叉不可缺少的组成部分，是供上、下相交的道路。在t型(y型)互通立交中，通常将相交的主要道路定义为主线，相交次要道路定义为引线，连接引线与主线互通的线路称为匝道。
如图2B所示，当第一交通工具(即自车)2021要驶入匝道202时(off-ramp merge)，确定第一交通工具2021的逻辑引导车2023及逻辑追随者2022，并且采用本申请实施例所实现的方法，确定目标预测偏移量，基于目标预测偏移量行驶，以挤占逻辑追随者2022的行车空间，占据一定的主动路权，提高第一交通工具2021的变道效率。其中，与第一交通工具2021位于同一车行道203、且在第一交通工具2021的前方行驶的交通工具，为第一交通工具2021的引导车2024。引导车可以是指与自车位于同一车行道、且在自车前方行驶的交通工具，逻辑引导车可以是指在自车待进入的车行道中在自车前方行驶的交通工具。
同理，如图2C所示，当第一交通工具2031进入该第一交通工具2031的逻辑引导车2033及逻辑追随者2032所在的车行道，即***逻辑引导车2033与逻辑追随者2032之间时，可以通过本申请实施例实现对第一交通工具2031的行驶路线的规划，提高第一交通工具2031的变道效率。其中，图2A至图2C仅为列举本申请适用的几种可能的应用场景，其他变道场景也可以应用本申请实施例所实现的方案，在此不做限制。
举例来说，参见图3，图3是本申请实施例提供的一种引导交通工具的确定场景示意图。如图3所示，假定第一交通工具3011所在的交通道路包括左一道、左二道、左三道及左四道，第一交通工具3011所在的车行道为左二道，第一交通工具3011要从左二道变道至左三道，以***逻辑引导车3012及逻辑追随者3013之间。将该逻辑追随者3013记作第二交通工具。第一交通工具3011中的自动驾驶***可以对第一交通工具3011进行决策，确定该第一交通工具3011的至少两个预测偏移量，该预测偏移量用于指示预测的第一交通工具3011在下一时刻向左三道(待变换到的车道)偏移的距离。该至少两个预测偏移量包括n个预测偏移量，如预测偏移量1、…及预测偏移量n，其中，n为正整数。自动驾驶***基于第一交通工具3011及第二交通工具3013的行驶状态，确定每个预测偏移量的目标变道效益，得到预测偏移量1的目标变道效益1、…及预测偏移量n的目标变道效益n。假定目标变道效益1至目标变道效益n中，目标变道效益3最大，则自动驾驶***将目标变道效益3对应的预测偏移量3确定为目标预测偏移量，基于目标预测偏移量确定第一交通工具3011的预测偏移轨迹302，控制第一交通工具3011沿预测偏移轨迹302行驶，以增加第二交通工具3013在预测偏移轨迹302的基础上，对第一交通工具3011进行让道的概率，使得第一交通工具3011成功变道至左三道的概率增加，提高了第一交通工具3011的变道效率。
进一步地，请参见图4，图4是本申请实施例提供的一种基于交通工具的数据处理的方法流程图。如图4所示，该基于交通工具的数据处理过程包括如下步骤：
步骤S401，确定第一交通工具的至少两个预测偏移量、第一交通工具的第一行驶状态以及第二交通工具的第二行驶状态。
在本申请实施例中，第一交通工具中的自动驾驶***在控制自车(即第一交通工具)进行变道时，确定自车所要变道进入的车行道中的交通信息，并基于所确定的交通信息，检测该车行道中是否存在第一交通工具的逻辑追随者PF。该逻辑追随者PF可以认为是对第一交通工具的变道造成影响的交通工具，将该逻辑追随者PF记作第二交通工具。换句话说，若第二交通工具不对第一交通工具让行，则第一交通工具无法进入该第二交通工具所在的车行道。第一交通工具中的自动驾驶***在检测到第二交通工具后，对第一交通工具向第二交通工具所在的车行道行驶时的偏移量进行决策，得到第一交通工具的至少两个预测偏移量，并确定第一交通工具的第一行驶状态以及第二交通工具的第二行驶状态。若该自动驾驶***检测到 自车所要进入的车行道中，不存在第一交通工具的逻辑追随者PF及逻辑引导车PL，则控制第一交通工具直接进行变道。若该自动驾驶***检测到自车所要进入的车行道中不存在第一交通工具的逻辑追随者PF，而存在第一交通工具的逻辑引导车PL，则控制第一交通工具调整与逻辑引导车PL之间在纵向上的距离，调整好后，控制第一交通工具进行变道。
第二交通工具为第一交通工具在变换车行道时参考的交通工具。自动驾驶***可以确定车行道的车道宽度及决策数量。所述决策数量可以预先确定，在此情况下，自动驾驶***可以例如从存储器获取预设的决策数量。自动驾驶***还可以确定第一交通工具与第一交通工具所在第一车行道的车道线之间的横向距离，并基于车道宽度、横向距离及决策数量，确定第一交通工具的至少两个预测偏移量。所述至少两个预测偏移量的数量可等于决策数量。举例来说，参见图5，图5是本申请实施例提供的一种行驶决策场景示意图。如图5所示，第一交通工具501的自动驾驶***确定逻辑引导车PL及逻辑追随者PF。将该逻辑追随者PF记作第二交通工具502，将该逻辑引导车PL记作第三交通工具503，将车道宽度记作lane_width，假定决策数量为n。自动驾驶***确定第一交通工具501与第一交通工具501所在的第一车行道504的车道线511之间的横向距离5041。车道线511为第一车行道504及第二交通工具502所在的第二车行道的共同边线。自动驾驶***可以基于车道宽度lane_width及横向距离5041，确定n个预测偏移量offset。例如，该n个预测偏移量最小可以为0，最大可以为车道宽度lane_width。自动驾驶***可以在0与车道宽度lane_width之间确定n个预测偏移量。例如，车道宽度lane_width为3.5米，n为5，第一交通工具501与第一车行道504的车道线之间的横向距离5041为0.9米，则自动驾驶***确定的至少两个预测偏移量可以包括“预测偏移量offset＝0、预测偏移量offset＝0.3米、预测偏移量offset＝0.6米、预测偏移量offset＝0.9米及预测偏移量offset＝3.5米”等。其中，该至少两个预测偏移量可以分为原车行道偏移量及变车道偏移量等。原车行道偏移量是指小于或等于横向距离5041的预测偏移量，变车道偏移量是指大于横向距离5041、且小于或等于车道宽度lane_width的预测偏移量。自动驾驶***可以基于决策数量，确定原车行道偏移量包括的预测偏移量的数量n1以及变车道偏移量包括的预测偏移量的数量n2，基于横向距离5041及原车行道偏移量包括的预测偏移量的数量n1，确定n1个预测偏移量，基于横向距离5041、车道宽度lane_width及变车道偏移量包括的预测偏移量的数量n2，确定n2个预测偏移量，其中，n1与n2均为正整数，n1与n2之和为n。
进一步地，自动驾驶***可以确定第一交通工具所在的第一车行道，以第一车行道的中心线作为坐标纵轴，以第一交通工具映射到坐标纵轴上的点作为坐标原点，将坐标纵轴对应的法线作为坐标横轴，根据坐标原点、坐标横轴以及坐标纵轴，建立道路坐标系。其中，将第一交通工具映射到坐标纵轴上时，可以是确定第一交通工具与坐标纵轴的最短距离，将该最短距离在坐标纵轴上对应的点作为坐标原点，或者，可以将第一交通工具映射到坐标纵轴上，该第一交通工具到坐标纵轴的映射路线与坐标纵轴垂直，将第一交通工具映射到坐标纵轴上的点，确定为坐标原点。
第一交通工具的第一行驶状态可以包括但不限于第一交通工具的第一位置信息、第一行驶速度及第一行驶方向等。第二交通工具的第二行驶状态可以包括但不限于第二交通工具的第二位置信息、第二行驶速度及第二行驶方向等。第一交通工具的第一行驶状态及第二交通工具的第二行驶状态，可以是由自动驾驶***中的感知模块确定的。具体地，自动驾驶***可以确定第一交通工具在道路坐标系中的第一位置信息，并根据第一位置信息确定第一交通工具的第一行驶状态；确定第二交通工具在道路坐标系中的第二位置信息，并根据第二位置信息确定第二交通工具的第二行驶状态。该第一行驶速度、第一行驶方向、第二行驶速度及第二行驶方向等也可以是基于道路坐标系确定的。
举例来说，参见图5，自动驾驶***确定坐标原点O、坐标纵轴S及坐标横轴D。该坐标原点O、坐 标纵轴S及坐标横轴D组成道路坐标系505。该坐标纵轴S的方向可以是第一交通工具501的行驶方向。例如，在图5中，该第一交通工具501的第一位置信息为(x1，y1)。如第一交通工具501在坐标原点O处，则x1为0，y1为0。第一行驶速度可以直接通过获取第一交通工具501的速度(如仪表盘上显示的速度)等确定。第一行驶方向可以通过第一单位向量进行表示。如第一单位向量为(0，1)时，表示第一行驶方向为沿坐标纵轴S的行驶方向。第一位置信息、第一行驶速度及第一行驶方向等组成第一交通工具501的第一行驶状态。第二交通工具502的第二位置信息为(x2，y2)，第二行驶方向可以通过第二单位向量进行表示。该第二位置信息、第二行驶方向及第二行驶速度等是基于道路坐标系505所确定的，其中，该第二位置信息、第二行驶方向及第二行驶速度等组成第二交通工具502的第二行驶状态。
步骤S402，根据第一行驶状态及第二行驶状态，确定在第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益(payoff)，以及确定在第二交通工具处于非让行预测状态时，各个预测偏移量的第二变道效益。
在本申请实施例中，第一交通工具中的自动驾驶***确定的各个预测偏移量的第一变道效益，以及各个预测偏移量的第二变道效益，可以通过表1进行表示：
表1
Figure PCTCN2021116193-appb-000001
其中，如表1所示，该表1包括自车EV(即第一交通工具)的至少两个预测偏移量以及第二交通工具(即逻辑追随者PF)的两种预测状态。举例来说，该至少两个预测偏移量包括n个预测偏移量，假定n为5。可以将该至少两个预测偏移量记作
Figure PCTCN2021116193-appb-000002
表示第i个预测偏移量。可以将第二交通工具的两种预测状态记作
Figure PCTCN2021116193-appb-000003
表示第二交通工具的第j种预测状态。该至少两个预测偏移量包括“预测偏移量1：offset＝0；预测偏移量2：offset＝0.3米；预测偏移量3：offset＝0.6米；预测偏移量4：offset＝0.9米及预测偏移量5：offset＝3.5米”。第二交通工具的两种预测状态包括让行预测状态(Yield)及非让行预测状态(Not Yield)。a i,j可以表示在第二交通工具处于第j种预测状态时，第i个预测偏移量的第j个变道效益，其中，i及j为正整数，i小于或等于n，n为至少两个预测偏移量包括的预测偏移量的数量，j小于或等于2。例如，a 1,1用于表示在第二交通工具处于让行预测状态时，第1个预测偏移量(即预测偏移量1)的第一变道效益。
Figure PCTCN2021116193-appb-000004
其中，S EV用于表示第一交通工具的第一行驶状态，S PF用于表示第二交通工具的第二行驶状态。第一行驶状态和第二行驶状态例如为当前时刻的第一行驶状态和第二行驶状态。
在一示例中，自动驾驶***可以获取让行概率模型。该让行概率模型可以是预先确定的。将第一行驶状态、第二行驶状态及第i个预测偏移量，输入让行概率模型，得到第i个预测偏移量对应的让行预测状态的概率及非让行预测状态的概率；i为正整数，i小于或等于至少两个预测偏移量的数量。将让行预测状 态的概率，确定为在第二交通工具处于让行预测状态时，第i个预测偏移量的第一变道效益。将非让行预测状态的概率，确定为在第二交通工具处于非让行预测状态时，第i个预测偏移量的第二变道效益。
步骤S403，确定第二交通工具的预测让行概率，根据预测让行概率、各个预测偏移量的第一变道效益和第二变道效益，生成各个预测偏移量的目标变道效益，将具有最大的目标变道效益的预测偏移量确定为目标预测偏移量。
在本申请实施例中，自动驾驶***可以获取历史纵向碰撞时间及历史纵向碰撞时间对应的历史碰撞距离，并根据历史纵向碰撞时间及历史碰撞距离，确定碰撞时间均值及碰撞时间标准差。历史纵向碰撞时间以及历史纵向碰撞时间对应的历史碰撞距离，可以是事先在测试过程中采集的一个或多个交通工具的纵向碰撞时间样本以及和纵向碰撞时间样本对应的碰撞距离样本。比如统计多次车辆A打转向灯准备变道、其侧后方车辆对车辆A进行让行时的纵向碰撞时间的分布规律，以及对应的纵向碰撞距离。历史纵向碰撞距离，可以是在不同的碰撞距离区间分别进行测试与数据采集得到的。将碰撞时间均值记作μ ttc，将碰撞时间标准差记作σ ttc。历史碰撞距离可以用于对历史纵向碰撞时间进行划分。例如，获取统计距离范围，获取处于该统计距离范围内的历史碰撞距离，根据处于该统计距离范围内的历史碰撞距离对应的历史纵向碰撞时间，确定碰撞时间均值及碰撞时间标准差。若交通工具1与交通工具2之间的碰撞距离属于该统计距离范围，则表示该交通工具1与交通工具2可能会对要变道***两者(即交通工具1与交通工具2)之间的交通工具的变道造成影响。碰撞时间(Time to Collision，TTC)也称为预计碰撞时间，表示如果两车保持当前速度和行驶轨迹不变，且不采取任何避碰行为，从当前时刻至发生碰撞的时间。例如，TTC＝两车车距/两车的相对车速。碰撞时间有时也指纵向碰撞时间。碰撞距离(Distance to Collision)或预计碰撞距离，与碰撞时间对应。
对于第二交通工具(PF)是否对第一交通工具(EV)进行让行，可以考虑两个因素：纵向的碰撞风险与侧向的距离安全。对于纵向的因素，由于第二交通工具占有路权，当第一交通工具给出换道意图时，如果第二交通工具没有充分的时间避让，往往会选择不进行避让，会保持原来的状态甚至是加速行驶。这个概率关系可以用高斯变量描述。自动驾驶***可以确定第二交通工具与第一交通工具的纵向碰撞时间，将纵向碰撞时间映射到由碰撞时间均值及碰撞时间标准差生成的第一概率密度函数中，以确定第二交通工具的初始让行概率。该初始让行概率可以通过公式①进行表示：
Figure PCTCN2021116193-appb-000005
其中，p(Yield；ttc(EV,PF))表示在给定第一交通工具与第二交通工具的纵向碰撞时间时，第二交通工具的初始让行概率，ttc(EV,PF)用于表示第一交通工具与第二交通工具的纵向碰撞时间，∝表示正比于，
Figure PCTCN2021116193-appb-000006
为第一概率密度函数。
进一步地，对于侧向的因素，后车一般会因为侧前方车辆侧向距离过近而进行避让，但是当侧向距离较远时，往往不会进行避让。自动驾驶***可以获取历史交通侧向间距，根据历史交通侧向间距确定间距均值及间距标准差，将间距均值记作μ dy，将间距标准差记作σ dy。与历史纵向碰撞时间及历史碰撞距离类似，历史交通侧向间距是指事先采集的、或获取的一个或多个交通工具的历史交通侧向间距样本。自动驾驶***还可以获取第二交通工具与第一交通工具的交通侧向间距，将交通侧向间距映射到由间距均值及间距标准差生成的第二概率密度函数中，以确定第二交通工具的行驶保持概率。该行驶保持概率可以用于表示第二交通工具不对第一交通工具让行的概率(即第二交通工具处于第二预测状态的概率)。该行驶保持概率可以通过公式②进行表示：
Figure PCTCN2021116193-appb-000007
其中，p(Not Yield；dy(EV,PF))表示在给定第一交通工具与第二交通工具之间的交通侧向间距的情况下，第二交通工具的行驶保持概率，dy(EV,PF)用于表示第一交通工具与第二交通工具之间的交通侧向间距，∝表示正比于，
Figure PCTCN2021116193-appb-000008
为第二概率密度函数。
在一示例中，在获取本申请实施例中所提及的各个标准差(如碰撞时间标准差或间距标准差等)时，也可以是获取各个标准差分别对应的方差(如碰撞时间标准差对应的碰撞时间方差，或者间距标准差对应的间距方差)，其中，标准差为对应方差的算术平方根。无论是获取方差，还是获取标准差，并不影响本申请实施例的实现，或者说，不影响公式①及公式②的实现，在此不再进行更多说明。
自动驾驶***可以根据初始让行概率及行驶保持概率，确定第二交通工具的预测让行概率。该预测让行概率可以通过公式③进行表示：
P(Yield)＝P(Yield；ttc(EV,PF))*(1-P(Not Yield；dy(EV,PF)))      ③
公式③是在初始让行概率与行驶保持概率为独立事件的情况下，生成预测让行概率的公式。若初始让行概率与行驶保持概率是非独立事件时，根据初始让行概率与行驶保持概率，生成预测让行概率的公式则可以基于初始让行概率与行驶保持概率之间的关系确定。
自动驾驶***在根据预测让行概率、各个预测偏移量的第一变道效益和第二变道效益，生成各个预测偏移量的目标变道效益后，可以将具有最大的目标变道效益的预测偏移量确定为目标预测偏移量，其中，该目标预测偏移量用于表示针对所述第一交通工具所预测的侧向变道行驶距离。
进一步地，在步骤S402中，确定第一交通工具的第一变道效益及第二变道效益时，具体过程如下：
其中，表1中的每个变道效益可以是由多个效益参数所确定的。假定存在m个效益参数，每个变道效益可以由m个效益参数中的任意一个或任意h个效益参数所确定的，其中，m为正整数，h为小于或等于m的正整数。例如，m为3，即存在3个效益参数，则每个变道效益可以是由1个效益参数所确定的，也可以是由任意两个效益参数所确定的，或者可以是由3个效益参数所确定的，在此不做限制。可以用
Figure PCTCN2021116193-appb-000009
表示根据第p个效益参数，确定第二交通工具处于第j个预测状态时，第i个预测偏移量的第j个变道效益，其中，p为正整数，p小于或等于m。
在一种变道效益确定方式中，自动驾驶***可以根据第一位置信息确定第一交通工具的偏移距离，该偏移距离为第一交通工具与第一交通工具所在的第一车行道的中心线之间的距离；获取第i个预测偏移量，根据第i个预测偏移量与偏移距离之间的差值，确定在第二交通工具处于让行预测状态时，第i个预测偏移量的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，第i个预测偏移量的第二变道效益；i为正整数，i小于或等于至少两个预测偏移量的数量。可以将根据第i个预测偏移量与偏移距离之间的差值所确定的效益参数，记作在第二交通工具处于不同预测状态下时，各个预测偏移量的第一个效益参数，将各个预测偏移量的第一个效益参数确定为各个预测偏移量的第一变道效益及第二变道效益。例如，该第一个效益参数用于表示自车舒适性，该自车舒适性可以通过公式④进行表示：
Figure PCTCN2021116193-appb-000010
其中，
Figure PCTCN2021116193-appb-000011
表示第一个效益参数，可以表示第一交通工具在第i个预测偏移量、且第二交通工具处于第j个预测状态时的舒适性，
Figure PCTCN2021116193-appb-000012
为第一交通工具与第一交通工具所在的第一车行道的中心线之间的距离。假定
Figure PCTCN2021116193-appb-000013
则以表1为例，
Figure PCTCN2021116193-appb-000014
若直接根据第一个效益参数确定第一变道效益及第二变道效益，则可以记作
Figure PCTCN2021116193-appb-000015
在一种变道效益确定方式中，自动驾驶***可以获取第一位置信息中的第一纵向坐标值及第二位置信息中的第二纵向坐标值，根据第一纵向坐标值、第一行驶速度、第二纵向坐标值及第二行驶速度，确定第一交通工具与第二交通工具的纵向碰撞时间。另外，自动驾驶***可以获取第一位置信息中的第一横向坐标值及第二位置信息中的第二横向坐标值，获取第i个预测偏移量，根据第一横向坐标值、第二横向坐标值及第i个预测偏移量，确定第i个预测偏移量对应的交通侧向间距；i为正整数，i小于或等于至少两个预测偏移量的数量。另外，自动驾驶***还可以根据纵向碰撞时间及第i个预测偏移量对应的交通侧向间距，确定在第二交通工具处于让行预测状态时，第i个预测偏移量的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，第i个预测偏移量的第二变道效益。自动驾驶***还可以根据纵向碰撞时间及第i个预测偏移量对应的交通侧向间距，确定多个效益参数，如第二个效益参数“侧向安全性”、第三个效益参数“碰撞安全性”及第四个效益参数“换道奖励”等，并可以根据直接根据其中一个效益参数确定第一变道效益及第二变道效益，也可以将各个效益参数进行随机组合，根据任意一组效益参数确定第一变道效益及第二变道效益。
例如，该第二个效益参数用于表示侧向安全性，该侧向安全性可以通过公式⑤进行表示：
Figure PCTCN2021116193-appb-000016
其中，
Figure PCTCN2021116193-appb-000017
用于表示第二个效益参数，可以表示第一交通工具在第i个预测偏移量、且第二交通工具处于第j个预测状态时的侧向安全性，其中，ttc(EV,PF)用于表示第一交通工具与第二交通工具的纵向碰撞时间，
Figure PCTCN2021116193-appb-000018
则表示在给定第i个预测偏移量
Figure PCTCN2021116193-appb-000019
的情况下，第i个预测偏移量对应的交通侧向间距。公式⑤可以表示当纵向碰撞时间小于2秒、且第i个预测偏移量的交通侧向间距小于1米时，确定第i个预测偏移量的第二个效益参数为-1；在其他情况下(即otherwise)，则可以确定第i个预测偏移量的第二个效益参数为0。若直接根据第二个效益参数确定第一变道效益及第二变道效益，则可以记作
Figure PCTCN2021116193-appb-000020
在上述公式⑤中，与纵向碰撞时间进行对比的第一时间阈值，及与交通侧向间距进行对比的第一间距阈值可以根据需要进行修改。如公式⑤中的第一时间阈值“2秒”及第一间距阈值“1米”为一种经验值，也可以是其他数值，在此不做限制。当纵向碰撞时间小于第一时间阈值、且第i个预测偏移量的交通侧向间距小于第一间距阈值时，确定的第i个预测偏移量的第二个效益参数也可以基于需要进行更改。
其中，ttc(EV,PF)可以通过公式⑥进行表示：
ttc(EV,PF)＝(l EV-l PF)/(v PF-v EV)，v PF>v EV      ⑥其中，l EV表示第一交通工具在道路坐标系下的第一纵向坐标值，l PF用于表示第二交通工具在道路坐标系下的第二纵向坐标值，v PF表示第二交通工具的第二行驶速度，v EV表示第一交通工具的第一行驶速度。
例如，该第三个效益参数用于表示碰撞安全性，该碰撞安全性可以通过公式⑦进行表示：
Figure PCTCN2021116193-appb-000021
其中，
Figure PCTCN2021116193-appb-000022
用于表示第三个效益参数，可以表示第一交通工具在第i个预测偏移量、且第二交通工具处于第j个预测状态时的侧向安全性，其中，ttc(EV,PF)与
Figure PCTCN2021116193-appb-000023
的含义可以参见公式⑤中的相关描述。其中，公式⑦可以表示当纵向碰撞时间小于2秒、且第i个预测偏移量的交通侧向间距小于0米时，确定第i个预测偏移量的第三个效益参数为-10；其他情况下，则可以确定第i个预测偏移量的第三个效益参数为0。若直接根据第三个效益参数，确定第一变道效益及第二变道效益，则可以记作
Figure PCTCN2021116193-appb-000024
在上述公式⑦中，与纵向碰撞时间进行对比的第二时间阈值，及与交通侧向间距进行对比的第二间距阈值可以根据需要进行修改，如公式⑦中的第二时间阈值“2秒”及第二间距阈值“0米”为一种经验值，也可以是其他数值，在此不做限制。当纵向碰撞时间小于第二时间阈值、且第i个预测偏移量的交通侧向间距小于第二间距阈值时，确定的第i个预测偏移量的第三个效益参数也可以基于需要进行更改。
例如，该第四个效益参数用于表示换道奖励，该换道奖励可以通过公式⑧进行表示：
Figure PCTCN2021116193-appb-000025
其中，
Figure PCTCN2021116193-appb-000026
用于表示第四个效益参数，可以表示第一交通工具在第i个预测偏移量、且第二交通工具处于第j个预测状态时的换道奖励，其中，ttc(EV,PF)与
Figure PCTCN2021116193-appb-000027
的含义可以参见公式⑤中的相关描述。公式⑧可以表示当纵向碰撞时间大于3秒、且第i个预测偏移量的交通侧向间距小于0米时，确定第i个预测偏移量的第四个效益参数为5；其他情况下，则可以确定第i个预测偏移量的第四个效益参数为0。若直接根据第四个效益参数，确定第一变道效益及第二变道效益，则可以记作
Figure PCTCN2021116193-appb-000028
上述公式⑧中，与纵向碰撞时间进行对比的第三时间阈值，及与交通侧向间距进行对比的第三间距阈值可以根据需要进行修改。如公式⑧中的第三时间阈值“3秒”及第三间距阈值“0米”为一种经验值，也可以是其他数值，在此不做限制。其中，当纵向碰撞时间大于第三时间阈值、且第i个预测偏移量的交通侧向间距小于第三间距阈值时，确定的第i个预测偏移量的第四个效益参数也可以基于需要进行更改。
在另一种变道效益获取方式中，自动驾驶***可以确定第三交通工具的第三位置信息，根据第二位置信息及第三位置信息，确定第二交通工具及第三交通工具之间的引导纵向距离。第二交通工具与第三交通工具处于同一车行道，第三交通工具与第二交通工具的行驶方向相同，第三交通工具在第二交通工具的前方，该前方是以第二交通工具的行驶方向为基准的。该第三交通工具可以认为是第一交通工具的逻辑引导车PL。自动驾驶***可以再获取第三交通工具的第三行驶速度。根据第i个预测偏移量及引导纵向距离，自动驾驶***可以确定第二交通工具的第i个让行距离，基于第二行驶速度、第三行驶速度及第i个让行距离，确定在第二交通工具处于让行预测状态时，第i个预测偏移量的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，第i个预测偏移量的第二变道效益；i为正整数，i小于或等于至少两个预测偏移量的数量。第i个让行距离是指在第i个预测偏移量、且第二交通工具处于让行预测状态时，第二交通工具与第三交通工具之间基于引导纵向距离所增加的距离。
可以参见图6，图6是本申请实施例提供的一种让行距离确定场景示意图。如图6所示，第一交通工具6011的自动驾驶***获取第二交通工具6012的第二纵向坐标值、第三交通工具6013的第三位置信息、该第三位置信息中的第三纵向坐标值，根据第二纵向坐标值及第三纵向坐标值，确定第二交通工具6012及第三交通工具6013之间的引导纵向距离602。自动驾驶***获取第i个预测偏移量603，基于第i个预测偏移量603及引导纵向距离602，确定所述第二交通工具的第i个让行距离。该第i个让行距离是指在第i个预测偏移量603、且第二交通工具6012处于让行预测状态时，第二交通工具6012与第三交通工具6013之间基于引导纵向距离602所增加的距离604。
另外，自动驾驶***可以基于第二行驶速度、第三行驶速度及第i个让行距离，确定第五个效益参数，该第五个效益参数用于表示交通堵塞成本(Traffic block cost)。该交通堵塞成本可以通过公式⑨进行表示：
Figure PCTCN2021116193-appb-000029
其中，
Figure PCTCN2021116193-appb-000030
用于表示在给定第一交通工具的第一行驶状态S EV、第i个预测偏移量
Figure PCTCN2021116193-appb-000031
及第j个预测状态
Figure PCTCN2021116193-appb-000032
的情况下，若第二交通工具为第一交通工具让行(即若第二交通工具处于让行预测状态)，则第二交通工具所需要的减速度。在一实施方式中，也可以将第一行驶状态、第二行驶状态、第i个预测偏移量及第j个预测状态等，输入智能驾驶员模型(Intelligent Driver Model，IDM)，基于该智能驾驶员模型，确定第二交通工具的减速度。公式⑨表示当第二交通工具的减速度小于-1时，确定第i个预测偏移量的第五个效益参数为-5；在其他情况下(即otherwise)，则可以确定第i个预测偏移量的第五个效益参数为0。若直接根据第五个效益参数，确定第一变道效益及第二变道效益，则可以记作
Figure PCTCN2021116193-appb-000033
以上第一个效益参数至第五个效益参数为示例的几种效益参数(此时的m为5)，也可以根据需要增加其他的效益参数，在此并不做限制。该第一变道效益及第二变道效益可以是基于m个效益参数中的任意一个或任意h个效益参数所确定的。若基于h个效益参数确定第一变道效益及第二变道效益，则可以对在第i个预测偏移量和第j个预测状态下的h个效益参数进行加权求和，以得到在第i个预测偏移量和第j个预测状态下的第j个变道效益。例如，该第一变道效益及第二变道效益是根据m个效益参数所确定的，则该第一变道效益及第二变道效益可以通过公式⑩进行表示：
Figure PCTCN2021116193-appb-000034
其中，λ p可以根据各个效益参数的重要程度确定。例如，各个效益参数的重要程度相同，则无论p的取值是几，λ p均为1。
本申请实施例通过确定第一交通工具的至少两个预测偏移量、第一交通工具的第一行驶状态以及第二交通工具的第二行驶状态；根据第一行驶状态及第二行驶状态，确定在第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，各个预测偏移量的第二变道效益；确定第二交通工具的预测让行概率，根据预测让行概率、各个预测偏移量的第一变道效益和第二变道效益，生成各个预测偏移量的目标变道效益，将具有最大的目标变道效益的预测偏移量确定为目标预测偏移量；目标预测偏移量用于表示针对第一交通工具所预测的侧向变道行驶距离。通过以上过程，对第一交通工具(即自车)进行决策，该决策用于表示第一交通工具可能的偏移距离(即至少两个预测偏移量)，通过获取各个决策的效益值，以得到可以使第一交通工具产生最大效益的决策，可以基于该决策(目标预测偏移量)控制第一交通工具行驶，使得第一交通工具可以在在一定程度上拥有路权，可以主动基于决策进行变道，以提高第一交通工具的变道效率。
进一步地，请参见图7，图7是本申请实施例提供的一种基于交通工具的数据处理的具体方法流程图。在该方法中，自动驾驶***可以构建决策树；决策树中的决策边包括至少两个预测偏移量、让行预测状态及非让行预测状态；决策树中的树节点包括第一交通工具及第二交通工具。在决策树的至少两个决策层中，逐层根据预测让行概率以及决策边，对各个预测偏移量的第一变道效益和第二变道效益进行加权求和，直至得到决策树的根节点中各个预测偏移量的树形效益值。将根节点中各个预测偏移量的树形效益值，确定为各个预测偏移量的目标变道效益。
如图7所示，该数据处理过程可以包括如下步骤：
步骤S701，确定决策树以及决策树每一层中，第一交通工具的第一行驶状态及第二交通工具的第二行驶状态。
在本申请实施例中，自动驾驶***可以确定决策树。假定该决策树中包括T个决策层，每个决策层包括由第一交通工具及第二交通工具交替作为树节点的一个结构。即第一交通工具的父节点可为第二交通工具，第二交通工具的父节点可为第一交通工具。每个决策层包括第一交通工具的至少两个预测偏移量，以及第二交通工具的让行预测状态及非让行状态等所组成的决策边。其中，T为正整数，T为决策树包括的决策层的数量。例如，第一交通工具对应n个决策边，每个决策边对应第一交通工具的一个预测偏移量；第二交通工具对应2个决策边，分别对应让行预测状态及非让行预测状态。自动驾驶***可以获取决策树中每一层中第一交通工具的第一行驶状态及第二交通工具的第二行驶状态。第一行驶状态包括第一实际行驶状态和(T-1)个第一预测行驶状态，其中，每个第一行驶状态对应一个决策层。第二行驶状态包括第二实际行驶状态和(T-1)个第二预测行驶状态，其中，每个第二行驶状态对应一个决策层。例如，自动驾驶***可以获取到第一实际行驶状态及第二实际行驶状态，对第一实际行驶状态进行预测，得到第2个决策层的第一预测行驶状态，对第二实际行驶状态进行预测，得到第2个决策层的第二预测行驶状态；对第2个决策层的第一预测行驶状态进行预测，得到第3个决策层的第一预测行驶状态，对第2个决策层的第二预测行驶状态进行预测，得到第3个决策层的第二预测行驶状态；…；对第(T-1)个决策层的第一预测行驶状态进行预测，得到第T个决策层的第一预测行驶状态，对第(T-1)个决策层的第二预测行驶状态进行预测，得到第T个决策层的第二预测行驶状态。
步骤S702，根据第k层的第一行驶状态及第二行驶状态，确定各个预测偏移量在第k层的第一变道效益及第二变道效益。
在本申请实施例中，自动驾驶***可以根据第k层的第一预测行驶状态及第二预测行驶状态，确定各个预测偏移量在第k层的第一变道效益及第二变道效益，其中，k为正整数，k小于或等于T。其中，自动驾驶***根据第k层的第一预测行驶状态及第二预测行驶状态，确定各个预测偏移量在第k层的第一变道效益及第二变道效益的过程，可以参见图4中步骤S402所示的具体描述，在此不再进行赘述。
步骤S703，确定第二交通工具在第k层的预测让行概率，基于第k层的预测让行概率、第(k+1)层的参数值各个预测偏移量在第k层的第一变道效益及第二变道效益，生成各个预测偏移量在第k层的树形变道效益。
在本申请实施例中，自动驾驶***可以确定第二交通工具在第k层的预测让行概率。第二交通工具在第k层的预测让行概率的确定过程，可以参见图4中步骤S403所示的具体描述。例如，公式①中ttc(EV,PF)表示，根据第k层的第一预测行驶状态及第二预测行驶状态，所确定的第一交通工具与第二交通工具之间的纵向碰撞时间。其中，可以根据第k层的第一预测行驶状态及第二预测行驶状态，确定第二交通工具在第k层的预测让行概率。其中，第k层中各个预测偏移量的树形效益值可以参见公式(1)所示：
Figure PCTCN2021116193-appb-000035
其中，
Figure PCTCN2021116193-appb-000036
表示第一交通工具在第k层的行驶状态，
Figure PCTCN2021116193-appb-000037
表示第一交通工具在第(k+1)层的行驶状态，
Figure PCTCN2021116193-appb-000038
是对
Figure PCTCN2021116193-appb-000039
进行预测得到的；
Figure PCTCN2021116193-appb-000040
表示第二交通工具在第k层的行驶状态，
Figure PCTCN2021116193-appb-000041
表示第二交通工具在第(k+1)层的行驶状态，
Figure PCTCN2021116193-appb-000042
是对
Figure PCTCN2021116193-appb-000043
进行预测得到的。其中，P(Yield)表示第二交通工具在第k层的预测让行概率，a i,1为当第二交通工具处于让行预测状态时，第i个预测偏移量的第一变道效益，a i,2为当第二交通工具处于非让行预测状态时，第i个预测偏移量的第二变道效益。其中，
Figure PCTCN2021116193-appb-000044
用于表示第i个预测偏移量在第k层中的树形效益值。
步骤S704，检测k＝1。
在本申请实施例中，自动驾驶***判断k是否为1，若k为1，则执行步骤S706，若k不为1，则执行步骤S705。
步骤S705，根据各个预测偏移量在第k层的树形变道效益，确定第k层的参数值。
在本申请实施例中，自动驾驶***可以根据各个预测偏移量在第k层的树形变道效益，确定第k层的参数值，该第k层的参数值的确定方式可以参见公式(2)所示：
Figure PCTCN2021116193-appb-000045
其中，当k＝T+1时，将该参数值置为预设参数值，如公式(2)中，预设参数值为0。当k≤T时，第k层的参数值为各个预测偏移量在第k层的树形变道效益中最大的树形变道效益。
步骤S706，确定各个预测偏移量的目标变道效益。
在本申请实施例中，自动驾驶***可以将根节点中各个预测偏移量的树形效益值，确定为各个预测偏移量的目标变量效益，将具有最大的目标变道效益的预测偏移量确定为目标预测偏移量，该目标预测偏移量的确定方式可以参见公式(3)所示：
Figure PCTCN2021116193-appb-000046
其中，argmax(f(x))是使f(x)取得最大值的变量点x(或x的集合)。在一实施方式中，若根据公式(3)，确定具有最大的目标变道效益的预测偏移量不止一个，则可以将目标变道效益最大且值最大的预测偏移量，确定为目标预测偏移量。
自动驾驶***可以根据目标预测偏移量及第二交通工具所在的第二车行道，确定预测偏移轨迹；预测偏移轨迹在车行道中对应的侧向行驶距离为目标预测偏移量。控制第一交通工具沿预测偏移轨迹行驶。
举例来说，若决策树包括2个决策层，即该决策树中的至少两个决策层包括决策层k1和决策层k2，决策层k1包括根节点。第一行驶状态包括第一实际行驶状态和第一预测行驶状态，第二行驶状态包括第二实际行驶状态和第二预测行驶状态。第一变道效益包括决策层k1的第一变道效益和决策层k2的第一变道效益，第二变道效益包括决策层k1的第二变道效益和决策层k2的第二变道效益。例如，参见图8，图8是本申请实施例提供的一种决策树的树形结构示意图。如图8所示，该决策树中圆形表示第二交通工具，七边形表示第一交通工具，该决策树包括决策层k1(即k＝1)及决策层k2(k＝2)。该决策树的根节点为第二交通工具a1。根节点a1包括两个决策边，分别表示让行预测状态及非让行预测状态(不让行)。这两个决策边分别连接第一交通工具b21及第一交通工具b22。第一交通工具b21对应n个决策边，分别表示预测偏移量1、…及预测偏移量n，第一交通工具b21对应的n个决策边分别连接第二交通工具a31、…及第二交通工具a32。第一交通工具b22对应n个决策边，分别表示预测偏移量1、…及预测偏移量n，第一交通工具b22对应的n个决策边分别连接第二交通工具a33、…及第二交通工具a34；…直至得到该决策树的叶子节点，如第二交通工具a51、第二交通工具a52、…及第二交通工具a66等，具体参见图8中各个树节点的标号，生成该决策树。其中，若T大于2时，则在图8的两个决策层的基础上继续生成决策层，以得到T个决策层，生成决策树。
具体地，自动驾驶***可以根据第一实际行驶状态和第二实际行驶状态，确定在第二交通工具处于让行预测状态时，各个预测偏移量在决策层k1中的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，各个预测偏移量在决策层k1中的第二变道效益。根据第一实际行驶状态预测第一交通工具的第一预测行驶状态，根据第二实际行驶状态预测第二交通工具的第二预测行驶状态。根据第一预测行驶状 态和第二预测行驶状态，确定在第二交通工具处于让行预测状态时，各个预测偏移量在决策层k2中的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，各个预测偏移量在决策层k2中的第二变道效益。
自动驾驶***可以获取第二交通工具在决策层k1中的预测让行概率，以及在决策层k2中的预测让行概率。根据决策层k2中的预测让行概率，对各个预测偏移量在决策层k2中的第一变道效益及第二变道效益进行加权求和，得到各个预测偏移量在决策层k2的树形效益值。将各个预测偏移量在决策层k2的树形效益值中最大的树形效益值，确定为决策层k2的参数值。根据在决策层k1中的预测让行概率，对决策层k2的参数值、各个预测偏移量在决策层k1中的第一变道效益及第二变道效益进行加权求和，得到各个预测偏移量在决策层k1的树形效益值；各个预测偏移量在决策层k1的树形效益值，为决策树的根节点中各个预测偏移量的树形效益值。将根节点中，具有最大的树形效益值的预测偏移量确定为目标预测偏移量，基于该目标预测偏移量确定第一交通工具的预测偏移轨迹，控制第一交通工具沿该预测偏移轨迹行驶，以主动挤占第二交通工具的行车空间，使得第一交通工具可以在尽可能大的向第二交通工具所在的第二车行道偏移，且与第二交通工具之间发生交通意外最小的情况下，实现向第二车行道的变道操作，提高第一交通工具的变道效率。
本申请实施例通过以上过程，可以使得第一交通工具不用等待第二交通工具进行主动让行，再进行变道，而是尽可能向第二交通工具所在的第二车行道进行偏移，挤占第二交通工具的行车空间，以更为明确地向第二交通工具表达自己变道的意愿，从而提高第一交通工具的变道效率。
进一步地，请参见图9，图9是本申请实施例提供的一种基于交通工具的数据处理装置示意图。该基于交通工具的数据处理装置可以是运行于计算机设备中的一个计算机程序(包括程序代码)，例如该基于交通工具的数据处理装置为一个应用软件；该装置可以用于执行本申请实施例提供的方法中的相应步骤。如图9所示，该基于交通工具的数据处理装置900可以用于图4所对应实施例中的计算机设备。具体地，该装置可以包括：状态获取模块11、效益获取模块12及偏移选取模块13。
状态获取模块11，用于确定第一交通工具的至少两个预测偏移量、第一交通工具的第一行驶状态以及第二交通工具的第二行驶状态；第二交通工具为第一交通工具在变换车行道时参考的交通工具；
效益获取模块12，用于根据第一行驶状态及第二行驶状态，确定在第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，各个预测偏移量的第二变道效益；
偏移选取模块13，用于确定第二交通工具的预测让行概率，根据预测让行概率、各个预测偏移量的第一变道效益和第二变道效益，生成各个预测偏移量的目标变道效益，将具有最大的目标变道效益的预测偏移量确定为目标预测偏移量；目标预测偏移量用于表示针对第一交通工具所预测的侧向变道行驶距离。
其中，在获取第一交通工具的至少两个预测偏移量方面，该状态获取模块11包括：
决策获取单元111，用于确定车行道的车道宽度及决策数量；
决策生成单元112，用于确定第一交通工具与第一交通工具所在第一车行道的车道线之间的横向距离，基于车道宽度、横向距离及决策数量，确定第一交通工具的至少两个预测偏移量；至少两个预测偏移量的数量为决策数量。
其中，该装置900还包括：
坐标建立模块14，用于确定第一交通工具所在的第一车行道，以第一车行道的中心线作为坐标纵轴，以第一交通工具映射到坐标纵轴上的点作为坐标原点，将坐标纵轴对应的法线作为坐标横轴，根据坐标原点、坐标横轴以及坐标纵轴，建立道路坐标系；
在确定第一交通工具的第一行驶状态以及第二交通工具的第二行驶状态方面，状态获取模块11包括：
第一获取单元113，用于确定第一交通工具在道路坐标系中的第一位置信息，根据第一位置信息确定第一交通工具的第一行驶状态；
第二获取单元114，用于确定第二交通工具在道路坐标系中的第二位置信息，根据第二位置信息确定第二交通工具的第二行驶状态。
其中，第一行驶状态包括第一位置信息，第二行驶状态包括第二位置信息；
该效益获取模块12，包括：
距离确定单元121，用于根据第一位置信息确定第一交通工具的偏移距离；偏移距离为第一交通工具与第一交通工具所在的第一车行道的中心线之间的距离；
效益获取单元122，用于获取第i个预测偏移量，根据第i个预测偏移量与偏移距离之间的差值，确定在第二交通工具处于让行预测状态时，第i个预测偏移量的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，第i个预测偏移量的第二变道效益；i为正整数，i小于或等于至少两个预测偏移量的数量。
其中，第一行驶状态包括第一位置信息及第一行驶速度，第二行驶状态包括第二位置信息及第二行驶速度；
该效益获取模块12，包括：
时间获取单元123，用于获取第一位置信息中的第一纵向坐标值及第二位置信息中的第二纵向坐标值，根据第一纵向坐标值、第一行驶速度、第二纵向坐标值及第二行驶速度，确定第一交通工具与第二交通工具的纵向碰撞时间；
间距获取单元124，用于获取第一位置信息中的第一横向坐标值及第二位置信息中的第二横向坐标值，并且获取第i个预测偏移量，根据第一横向坐标值、第二横向坐标值及第i个预测偏移量，确定第i个预测偏移量对应的交通侧向间距；i为正整数，i小于或等于至少两个预测偏移量的数量；
该效益获取单元122，还用于根据纵向碰撞时间及第i个预测偏移量对应的交通侧向间距，确定在第二交通工具处于让行预测状态时，第i个预测偏移量的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，第i个预测偏移量的第二变道效益。
其中，第一行驶状态包括第一位置信息及第一行驶速度，第二行驶状态包括第二位置信息及第二行驶速度；
该效益获取模块12，包括：
车辆获取单元125，用于确定第三交通工具的第三位置信息，根据第二位置信息及第三位置信息，确定第二交通工具及第三交通工具之间的引导纵向距离；第二交通工具与第三交通工具处于同一车行道，第三交通工具与第二交通工具的行驶方向相同；
速度获取单元126，用于确定第三交通工具的第三行驶速度；
该效益获取单元122，还用于根据第i个预测偏移量及引导纵向距离，确定第二交通工具的第i个让行距离，基于第二行驶速度、第三行驶速度及第i个让行距离，确定在第二交通工具处于让行预测状态时，第i个预测偏移量的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，第i个预测偏移量的第二变道效益；i为正整数，i小于或等于至少两个预测偏移量的数量；第i个让行距离是指在第i个预测偏移量、且第二交通工具处于让行预测状态时，第二交通工具与第三交通工具之间基于引导纵向距离所增加的距离。
其中，该效益获取模块12，包括：
模块获取单元127，用于获取让行概率模型；
概率预测单元128，用于将第一行驶状态、第二行驶状态及第i个预测偏移量，输入让行概率模型，得到第i个预测偏移量对应的让行预测状态的概率及非让行预测状态的概率；i为正整数，i小于或等于至少两个预测偏移量的数量；
该效益获取单元122，还用于将让行预测状态的概率，确定为在第二交通工具处于让行预测状态时，第i个预测偏移量的第一变道效益；
该效益获取单元122，还用于将非让行预测状态的概率，确定为在第二交通工具处于非让行预测状态时，第i个预测偏移量的第二变道效益。
其中，在获取第二交通工具的预测让行概率方面，该偏移选取模块13包括：
碰撞获取单元131，用于获取历史纵向碰撞时间及历史纵向碰撞时间对应的历史碰撞距离，根据历史纵向碰撞时间及历史碰撞距离，确定碰撞时间均值及碰撞时间标准差；
第一概率获取单元132，用于确定第二交通工具与第一交通工具的纵向碰撞时间，将纵向碰撞时间映射到由碰撞时间均值及碰撞时间标准差生成的第一概率密度函数中，以确定第二交通工具的初始让行概率；
历史间距获取单元133，用于获取历史交通侧向间距，根据历史交通侧向间距确定间距均值及间距标准差；
第二概率获取单元134，用于获取第二交通工具与第一交通工具的交通侧向间距，将交通侧向间距映射到由间距均值及间距标准差生成的第二概率密度函数中，以根据第二概率密度函数确定第二交通工具的行驶保持概率；
让行概率确定单元135，用于根据初始让行概率及行驶保持概率，确定第二交通工具的预测让行概率。
其中，在根据预测让行概率、各个预测偏移量的第一变道效益和第二变道效益，生成各个预测偏移量的目标变道效益方面，该偏移选取模块13包括：
树获取单元136，用于构建决策树；决策树中的决策边包括至少两个预测偏移量、让行预测状态及非让行预测状态；决策树中的树节点包括第一交通工具及第二交通工具；
效益迭代单元137，用于在决策树的至少两个决策层中，逐层根据预测让行概率以及决策边，对各个预测偏移量的第一变道效益和第二变道效益进行加权求和，直至得到决策树的根节点中各个预测偏移量的树形效益值；
效益确定单元138，用于将根节点中各个预测偏移量的树形效益值，确定为各个预测偏移量的目标变道效益。
其中，决策树中的至少两个决策层包括决策层k1和决策层k2，决策层k1包括根节点；第一行驶状态包括第一实际行驶状态和第一预测行驶状态，第二行驶状态包括第二实际行驶状态和第二预测行驶状态；第一变道效益包括决策层k1中的第一变道效益和决策层k2中的第一变道效益，第二变道效益包括决策层k1中的第二变道效益和决策层k2中的第二变道效益；
该效益获取模块12，包括：
实际处理单元129a，根据第一实际行驶状态和第二实际行驶状态，确定在第二交通工具处于让行预测状态时，各个预测偏移量在决策层k1中的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，各个预测偏移量在决策层k1中的第二变道效益；
状态预测单元129b，用于根据第一实际行驶状态预测第一交通工具的第一预测行驶状态，根据第二实际行驶状态预测第二交通工具的第二预测行驶状态；
预测处理单元129c，用于根据第一预测行驶状态和第二预测行驶状态，确定在第二交通工具处于让行 预测状态时，各个预测偏移量在决策层k2中的第一变道效益以及确定在第二交通工具处于非让行预测状态时，各个预测偏移量在决策层k2中的第二变道效益。
其中，预测让行概率包括决策层k1中的预测让行概率和决策层k2中的预测让行概率；
该效益迭代单元137，包括：
层效益获取子单元1371，用于根据决策层k2中的预测让行概率，对各个预测偏移量在决策层k2中的的第一变道效益及第二变道效益进行加权求和，得到各个预测偏移量在决策层k2的树形效益值；
参数确定子单元1372，用于将各个预测偏移量在决策层k2的树形效益值中最大的树形效益值，确定为决策层k2的参数值；
该层效益获取子单元1371，还用于根据决策层k1中的预测让行概率，对决策层k2的参数值、各个预测偏移量在决策层k1中的的第一变道效益及第二变道效益进行加权求和，得到各个预测偏移量在决策层k1的树形效益值；各个预测偏移量在决策层k1的树形效益值，为决策树的根节点中各个预测偏移量的树形效益值。
其中，该装置900还包括：
轨迹确定模块15，用于根据目标预测偏移量确定预测偏移轨迹；预测偏移轨迹在车行道中对应的侧向行驶距离为目标预测偏移量；
行驶控制模块16，用于控制第一交通工具沿预测偏移轨迹行驶。
本申请实施例提供了一种基于交通工具的数据处理装置，该装置通过获取第一交通工具的至少两个预测偏移量，获取第一交通工具的第一行驶状态，获取第二交通工具的第二行驶状态；第二交通工具为第一交通工具在变换车行道时参考的交通工具；根据第一行驶状态及第二行驶状态，确定在第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，各个预测偏移量的第二变道效益；获取第二交通工具的预测让行概率，根据预测让行概率、各个预测偏移量的第一变道效益和第二变道效益，生成各个预测偏移量的目标变道效益，将具有最大的目标变道效益的预测偏移量确定为目标预测偏移量；目标预测偏移量用于表示针对第一交通工具所预测的侧向变道行驶距离。通过以上过程，对第一交通工具(即自车)进行决策，该决策用于表示第一交通工具可能的偏移距离(即至少两个预测偏移量)，通过获取各个决策的效益值，以得到可以使第一交通工具产生最大效益的决策，可以基于该决策(目标预测偏移量)控制第一交通工具行驶，使得第一交通工具可以在在一定程度上拥有路权，可以主动基于决策进行变道，以提高第一交通工具的变道效率。
参见图10，图10是本申请实施例提供的一种计算机设备的结构示意图。如图10所示，本申请实施例中的计算机设备可以包括：一个或多个处理器1001、存储器1002和输入输出接口1003。该处理器1001、存储器1002和输入输出接口1003通过总线1004连接。存储器1002用于存储计算机程序，该计算机程序包括程序指令，输入输出接口1003用于接收数据及输出数据；处理器1001用于执行存储器1002存储的程序指令，执行如下操作：
确定第一交通工具的至少两个预测偏移量、第一交通工具的第一行驶状态以及第二交通工具的第二行驶状态；第二交通工具为第一交通工具在变换车行道时参考的交通工具；
根据第一行驶状态及第二行驶状态，确定在第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，各个预测偏移量的第二变道效益；
确定第二交通工具的预测让行概率，根据预测让行概率、各个预测偏移量的第一变道效益和第二变道效益，生成各个预测偏移量的目标变道效益，将具有最大的目标变道效益的预测偏移量确定为目标预测偏移量；目标预测偏移量用于表示针对第一交通工具所预测的侧向变道行驶距离。
在一些可行的实施方式中，该处理器1001可以是中央处理单元(central processing unit，CPU)，该处理器还可以是其他通用处理器、数字信号处理器(digital signal processor，DSP)、专用集成电路(application specific integrated circuit，ASIC)、现成可编程门阵列(field-programmable gate array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。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。
该存储器1002可以包括只读存储器和随机存取存储器，并向处理器1001和输入输出接口1003提供指令和数据。存储器1002的一部分还可以包括非易失性随机存取存储器。例如，存储器1002还可以存储设备类型的信息。
具体实现中，该计算机设备可通过其内置的各个功能模块执行如该图4中各个步骤所提供的实现方式，具体可参见该图4中各个步骤所提供的实现方式，在此不再赘述。
本申请实施例通过提供一种计算机设备，包括：处理器、输入输出接口、存储器，通过处理器获取存储器中的计算机指令，执行该图4中所示方法的各个步骤，进行基于交通工具的数据处理操作。本申请实施例实现了确定第一交通工具的至少两个预测偏移量、第一交通工具的第一行驶状态以及第二交通工具的第二行驶状态；第二交通工具为第一交通工具在变换车行道时参考的交通工具；根据第一行驶状态及第二行驶状态，确定在第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在第二交通工具处于非让行预测状态时，各个预测偏移量的第二变道效益；确定第二交通工具的预测让行概率，根据预测让行概率、各个预测偏移量的第一变道效益和第二变道效益，生成各个预测偏移量的目标变道效益，将具有最大的目标变道效益的预测偏移量确定为目标预测偏移量；目标预测偏移量用于表示针对第一交通工具所预测的侧向变道行驶距离。通过以上过程，对第一交通工具(即自车)进行决策，该决策用于表示第一交通工具可能的偏移距离(即至少两个预测偏移量)，通过获取各个决策的效益值，以得到可以使第一交通工具产生最大效益的决策，可以基于该决策(目标预测偏移量)控制第一交通工具行驶，使得第一交通工具可以在在一定程度上拥有路权，可以主动基于决策进行变道，以提高第一交通工具的变道效率。
本申请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存储有计算机程序，该计算机程序包括程序指令，当该程序指令被该处理器执行时，可以实现图4中各个步骤所提供的基于交通工具的数据处理方法，具体可参见该图4中各个步骤所提供的实现方式，在此不再赘述。另外，对采用相同方法的有益效果描述，也不再进行赘述。对于本申请所涉及的计算机可读存储介质实施例中未披露的技术细节，请参照本申请方法实施例的描述。作为示例，程序指令可被部署为在一个计算机设备上执行，或者在位于一个地点的多个计算机设备上执行，又或者，在分布在多个地点且通过通信网络互连的多个计算机设备上执行。
该计算机可读存储介质可以是前述任一实施例提供的基于交通工具的数据处理装置或者该计算机设备的内部存储单元，例如计算机设备的硬盘或内存。该计算机可读存储介质也可以是该计算机设备的外部存储设备，例如该计算机设备上配备的插接式硬盘，智能存储卡(smart media card，SMC)，安全数字(secure digital，SD)卡，闪存卡(flash card)等。进一步地，该计算机可读存储介质还可以既包括该计算机设备的内部存储单元也包括外部存储设备。该计算机可读存储介质用于存储该计算机程序以及该计算机设备所需的其他程序和数据。该计算机可读存储介质还可以用于暂时地存储已经输出或者将要输出的数据。
本申请实施例还提供了一种计算机程序产品或计算机程序，该计算机程序产品或计算机程序包括计算机指令，该计算机指令存储在计算机可读存储介质中。计算机设备的处理器从计算机可读存储介质读取该计算机指令，处理器执行该计算机指令，使得该计算机设备执行图4中的各种示例方式中所提供的方法，实现对第一交通工具的目标预测偏移量的确定，以控制第一交通工具基于该目标预测偏移量所确定的预测 偏移轨迹行驶，提高第一交通工具的变道效率。
本申请实施例的说明书和权利要求书及附图中的术语“第一”、“第二”等是用于区别不同对象，而非用于描述特定顺序。此外，术语“包括”以及它们任何变形，意图在于覆盖不排他的包含。例如包含了一系列步骤或单元的过程、方法、装置、产品或设备没有限定于已列出的步骤或模块，而是还可包括没有列出的步骤或模块，或还可包括对于这些过程、方法、装置、产品或设备固有的其他步骤单元。
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、计算机软件或者二者的结合来实现，为了清楚地说明硬件和软件的可互换性，在该说明中已经按照功能一般性地描述了各示例的组成及步骤。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。
本申请实施例提供的方法及相关装置是参照本申请实施例提供的方法流程图和/或结构示意图来描述的，具体可由计算机程序指令实现方法流程图和/或结构示意图的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。这些计算机程序指令可提供到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或结构示意图一个方框或多个方框中指定的功能的装置。这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或结构示意图一个方框或多个方框中指定的功能。这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或结构示意一个方框或多个方框中指定的功能的步骤。
以上所揭露的仅为本申请较佳实施例而已，当然不能以此来限定本申请之权利范围，因此依本申请权利要求所作的等同变化，仍属本申请所涵盖的范围。

Claims (15)

  1. 一种基于交通工具的数据处理方法，其特征在于，所述方法包括：
    确定第一交通工具的至少两个预测偏移量、所述第一交通工具的第一行驶状态以及第二交通工具的第二行驶状态；所述第二交通工具为所述第一交通工具在变换车行道时参考的交通工具；
    根据所述第一行驶状态及所述第二行驶状态，确定在所述第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在所述第二交通工具处于非让行预测状态时，所述各个预测偏移量的第二变道效益；
    确定所述第二交通工具的预测让行概率，根据所述预测让行概率、所述各个预测偏移量的所述第一变道效益和所述第二变道效益，生成所述各个预测偏移量的目标变道效益，将具有最大的目标变道效益的预测偏移量确定为目标预测偏移量；所述目标预测偏移量用于表示针对所述第一交通工具所预测的侧向变道行驶距离。
  2. 如权利要求1所述的方法，其特征在于，确定所述第一交通工具的至少两个预测偏移量，包括：
    确定所述车行道的车道宽度及决策数量；
    确定所述第一交通工具与所述第一交通工具所在的第一车行道的车道线之间的横向距离，基于所述车道宽度、所述横向距离及所述决策数量，确定所述第一交通工具的至少两个预测偏移量。
  3. 如权利要求1-2任一项所述的方法，其特征在于，所述方法还包括：
    确定所述第一交通工具所在的第一车行道，以所述第一车行道的中心线作为坐标纵轴，以所述第一交通工具映射到所述坐标纵轴上的点作为坐标原点，将所述坐标纵轴对应的法线作为坐标横轴，根据所述坐标原点、所述坐标横轴以及所述坐标纵轴，建立道路坐标系；
    所述确定所述第一交通工具的第一行驶状态以及第二交通工具的第二行驶状态，包括：
    确定所述第一交通工具在所述道路坐标系中的第一位置信息，根据所述第一位置信息确定所述第一交通工具的第一行驶状态；
    确定所述第二交通工具在所述道路坐标系中的第二位置信息，根据所述第二位置信息确定所述第二交通工具的第二行驶状态。
  4. 如权利要求1所述的方法，其特征在于，所述第一行驶状态包括第一位置信息，所述第二行驶状态包括第二位置信息；
    所述根据所述第一行驶状态及所述第二行驶状态，确定在所述第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在所述第二交通工具处于非让行预测状态时，所述各个预测偏移量的第二变道效益，包括：
    根据所述第一位置信息确定所述第一交通工具的偏移距离，所述偏移距离为所述第一交通工具与所述第一交通工具所在的第一车行道的中心线之间的距离；
    从所述至少两个预测偏移量中获取第i个预测偏移量，根据所述第i个预测偏移量与所述偏移距离之间的差值，确定在所述第二交通工具处于让行预测状态时，所述第i个预测偏移量的第一变道效益，以及确定在所述第二交通工具处于非让行预测状态时，所述第i个预测偏移量的第二变道效益，其中，i为正整数，i小于或等于所述至少两个预测偏移量的数量。
  5. 如权利要求1所述的方法，其特征在于，所述第一行驶状态包括第一位置信息及第一行驶速度，所述第二行驶状态包括第二位置信息及第二行驶速度；
    所述根据所述第一行驶状态及所述第二行驶状态，确定在所述第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在所述第二交通工具处于非让行预测状态时，所述各个预测偏移 量的第二变道效益，包括：
    获取所述第一位置信息中的第一纵向坐标值及所述第二位置信息中的第二纵向坐标值，根据所述第一纵向坐标值、所述第一行驶速度、所述第二纵向坐标值及所述第二行驶速度，确定所述第一交通工具与所述第二交通工具的纵向碰撞时间；
    获取所述第一位置信息中的第一横向坐标值及所述第二位置信息中的第二横向坐标值，并且获取第i个预测偏移量，根据所述第一横向坐标值、所述第二横向坐标值及所述第i个预测偏移量，确定所述第i个预测偏移量对应的交通侧向间距；i为正整数，i小于或等于所述至少两个预测偏移量的数量；
    根据所述纵向碰撞时间及所述第i个预测偏移量对应的交通侧向间距，确定在所述第二交通工具处于让行预测状态时，所述第i个预测偏移量的第一变道效益，以及确定在所述第二交通工具处于非让行预测状态时，所述第i个预测偏移量的第二变道效益。
  6. 如权利要求1所述的方法，其特征在于，所述第一行驶状态包括第一位置信息及第一行驶速度，所述第二行驶状态包括第二位置信息及第二行驶速度；
    根据所述第一行驶状态及所述第二行驶状态，确定在所述第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在所述第二交通工具处于非让行预测状态时，所述各个预测偏移量的第二变道效益，包括：
    确定第三交通工具的第三位置信息，根据所述第二位置信息及所述第三位置信息，确定所述第二交通工具及所述第三交通工具之间的引导纵向距离，其中，所述第二交通工具与所述第三交通工具处于同一车行道，所述第三交通工具与所述第二交通工具的行驶方向相同；
    确定所述第三交通工具的第三行驶速度；
    根据第i个预测偏移量及所述引导纵向距离，确定所述第二交通工具的第i个让行距离，基于所述第二行驶速度、所述第三行驶速度及所述第i个让行距离，确定在所述第二交通工具处于让行预测状态时，所述第i个预测偏移量的第一变道效益，以及确定在所述第二交通工具处于非让行预测状态时，所述第i个预测偏移量的第二变道效益，其中，i为正整数，i小于或等于所述至少两个预测偏移量的数量，所述第i个让行距离是指在第i个预测偏移量、且所述第二交通工具处于所述让行预测状态时，所述第二交通工具与所述第三交通工具之间基于所述引导纵向距离所增加的距离。
  7. 如权利要求1所述的方法，其特征在于，所述根据所述第一行驶状态及所述第二行驶状态，确定在所述第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在所述第二交通工具处于非让行预测状态时，所述各个预测偏移量的第二变道效益，包括：
    获取让行概率模型；
    将所述第一行驶状态、所述第二行驶状态及第i个预测偏移量，输入所述让行概率模型，得到所述第i个预测偏移量对应的所述让行预测状态的概率及所述非让行预测状态的概率，其中，i为正整数，i小于或等于所述至少两个预测偏移量的数量；
    将所述让行预测状态的概率，确定为在所述第二交通工具处于所述让行预测状态时，所述第i个预测偏移量的第一变道效益；
    将所述非让行预测状态的概率，确定为在所述第二交通工具处于所述非让行预测状态时，所述第i个预测偏移量的第二变道效益。
  8. 如权利要求1所述的方法，其特征在于，确定所述第二交通工具的预测让行概率，包括：
    获取历史纵向碰撞时间及所述历史纵向碰撞时间对应的历史碰撞距离，根据所述历史纵向碰撞时间及所述历史碰撞距离，确定碰撞时间均值及碰撞时间标准差；
    确定所述第二交通工具与所述第一交通工具的纵向碰撞时间，将所述纵向碰撞时间映射到由所述碰撞时间均值及所述碰撞时间标准差生成的第一概率密度函数中，以根据所述第一概率密度函数确定所述第二交通工具的初始让行概率；
    获取历史交通侧向间距，根据所述历史交通侧向间距确定间距均值及间距标准差；
    确定所述第二交通工具与所述第一交通工具的交通侧向间距，将所述交通侧向间距映射到由所述间距均值及所述间距标准差生成的第二概率密度函数中，以根据所述第二概率密度函数确定所述第二交通工具的行驶保持概率；
    根据所述初始让行概率及所述行驶保持概率，确定所述第二交通工具的预测让行概率。
  9. 如权利要求1所述的方法，其特征在于，所述根据所述预测让行概率、所述各个预测偏移量的所述第一变道效益和所述第二变道效益，生成所述各个预测偏移量的目标变道效益，包括：
    构建决策树；所述决策树中的决策边包括所述至少两个预测偏移量、所述让行预测状态及所述非让行预测状态；所述决策树中的树节点包括所述第一交通工具及所述第二交通工具；
    在所述决策树的至少两个决策层中，逐层根据所述预测让行概率以及所述决策边，对所述各个预测偏移量的所述第一变道效益和所述第二变道效益进行加权求和，直至得到所述决策树的根节点中所述各个预测偏移量的树形效益值；
    将所述根节点中所述各个预测偏移量的树形效益值，确定为所述各个预测偏移量的目标变道效益。
  10. 如权利要求9所述的方法，其特征在于，所述决策树中的至少两个决策层包括决策层k1和决策层k2，所述决策层k1包括所述根节点；所述第一行驶状态包括第一实际行驶状态和第一预测行驶状态，所述第二行驶状态包括第二实际行驶状态和第二预测行驶状态；所述第一变道效益包括决策层k1中的第一变道效益和决策层k2中的第一变道效益，所述第二变道效益包括决策层k1中的第二变道效益和决策层k2中的第二变道效益；
    所述根据所述第一行驶状态及所述第二行驶状态，确定在所述第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在所述第二交通工具处于非让行预测状态时，所述各个预测偏移量的第二变道效益，包括：
    根据所述第一实际行驶状态和所述第二实际行驶状态，确定在所述第二交通工具处于让行预测状态时，各个预测偏移量在决策层k1中的所述第一变道效益，以及确定在所述第二交通工具处于非让行预测状态时，所述各个预测偏移量在决策层k1中的第二变道效益；
    根据所述第一实际行驶状态预测所述第一交通工具的所述第一预测行驶状态，根据所述第二实际行驶状态预测所述第二交通工具的所述第二预测行驶状态；
    根据所述第一预测行驶状态和所述第二预测行驶状态，确定在所述第二交通工具处于所述让行预测状态时，所述各个预测偏移量在决策层k2中的所述第一变道效益，以及确定在所述第二交通工具处于所述非让行预测状态时，所述各个预测偏移量在决策层k2中的所述第二变道效益。
  11. 如权利要求10所述的方法，其特征在于，所述预测让行概率包括决策层k1中的预测让行概率和决策层k2中的预测让行概率；
    所述在所述决策树的至少两个决策层中，逐层根据所述预测让行概率以及所述决策边对所述各个预测偏移量的所述第一变道效益和所述第二变道效益进行加权求和，直至得到所述决策树的根节点中所述各个预测偏移量的树形效益值，包括：
    根据决策层k2中的所述预测让行概率，对所述各个预测偏移量在决策层k2中的所述第一变道效益及所述第二变道效益进行加权求和，得到所述各个预测偏移量在所述决策层k2的树形效益值；
    将所述各个预测偏移量在所述决策层k2的树形效益值中最大的树形效益值，确定为所述决策层k2的参数值；
    根据所述决策层k1中的预测让行概率，对所述决策层k2的参数值、所述各个预测偏移量在决策层k1中的所述第一变道效益及所述第二变道效益进行加权求和，得到所述各个预测偏移量在所述决策层k1的树形效益值；所述各个预测偏移量在所述决策层k1的树形效益值，为所述决策树的根节点中所述各个预测偏移量的树形效益值。
  12. 如权利要求1所述的方法，其特征在于，所述方法还包括：
    根据所述目标预测偏移量及所述第二交通工具所在的第二车行道，确定预测偏移轨迹；所述预测偏移轨迹在所述第一车行道和/或所述第二车行道中对应的侧向行驶距离为所述目标预测偏移量；
    控制所述第一交通工具沿所述预测偏移轨迹行驶。
  13. 一种基于交通工具的数据处理装置，其特征在于，所述装置包括：
    状态获取模块，用于确定第一交通工具的至少两个预测偏移量、所述第一交通工具的第一行驶状态以及第二交通工具的第二行驶状态；所述第二交通工具为所述第一交通工具在变换车行道时参考的交通工具；
    效益获取模块，用于根据所述第一行驶状态及所述第二行驶状态，确定在所述第二交通工具处于让行预测状态时，各个预测偏移量的第一变道效益，以及确定在所述第二交通工具处于非让行预测状态时，所述各个预测偏移量的第二变道效益；
    偏移选取模块，用于确定所述第二交通工具的预测让行概率，根据所述预测让行概率、所述各个预测偏移量的所述第一变道效益和所述第二变道效益，生成所述各个预测偏移量的目标变道效益，将具有最大的目标变道效益的预测偏移量确定为目标预测偏移量；所述目标预测偏移量用于表示针对所述第一交通工具所预测的侧向变道行驶距离。
  14. 一种计算机设备，其特征在于，包括处理器、存储器、输入输出接口；
    所述处理器分别与所述存储器和所述输入输出接口相连，其中，所述输入输出接口用于接收数据及输出数据，所述存储器用于存储计算机程序，所述处理器用于调用所述计算机程序，以执行如权利要求1-12任一项所述的方法。
  15. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质存储有计算机程序，所述计算机程序包括程序指令，所述程序指令当被处理器执行时，执行如权利要求1-12任一项所述的方法。
PCT/CN2021/116193 2020-09-10 2021-09-02 基于交通工具的数据处理方法、装置、计算机及存储介质 WO2022052856A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
JP2022565903A JP7520444B2 (ja) 2020-09-10 2021-09-02 乗り物に基づくデータ処理方法、データ処理装置、コンピュータ機器、及びコンピュータプログラム
EP21865910.0A EP4119412A4 (en) 2020-09-10 2021-09-02 VEHICLE-BASED DATA PROCESSING METHOD AND APPARATUS, COMPUTER AND RECORDING MEDIUM
US17/971,495 US20230053459A1 (en) 2020-09-10 2022-10-21 Vehicle-based data processing method and apparatus, computer, and storage medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010947834.9A CN112406905B (zh) 2020-09-10 2020-09-10 基于交通工具的数据处理方法、装置、计算机及存储介质
CN202010947834.9 2020-09-10

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/971,495 Continuation US20230053459A1 (en) 2020-09-10 2022-10-21 Vehicle-based data processing method and apparatus, computer, and storage medium

Publications (1)

Publication Number Publication Date
WO2022052856A1 true WO2022052856A1 (zh) 2022-03-17

Family

ID=74855176

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/116193 WO2022052856A1 (zh) 2020-09-10 2021-09-02 基于交通工具的数据处理方法、装置、计算机及存储介质

Country Status (5)

Country Link
US (1) US20230053459A1 (zh)
EP (1) EP4119412A4 (zh)
JP (1) JP7520444B2 (zh)
CN (1) CN112406905B (zh)
WO (1) WO2022052856A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2024039997A1 (en) * 2022-08-15 2024-02-22 Motional Ad Llc Determination of an action for an autonomous vehicle in the presence of intelligent agents

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112406905B (zh) * 2020-09-10 2022-01-28 腾讯科技（深圳）有限公司 基于交通工具的数据处理方法、装置、计算机及存储介质
US20240092398A1 (en) * 2022-08-31 2024-03-21 Zoox, Inc. Trajectory prediction based on a decision tree
CN116729374B (zh) * 2023-08-16 2023-11-10 北京斯年智驾科技有限公司 一种自动驾驶车道保持控制方法、***、装置和存储介质

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170291603A1 (en) * 2014-08-11 2017-10-12 Nissan Motor Co., Ltd. Travel Control Device and Method for Vehicle
CN108313054A (zh) * 2018-01-05 2018-07-24 北京智行者科技有限公司 自动驾驶自主换道决策方法和装置及自动驾驶车辆
CN108595823A (zh) * 2018-04-20 2018-09-28 大连理工大学 一种联合驾驶风格和博弈理论的自主车换道策略的计算方法
CN110298131A (zh) * 2019-07-05 2019-10-01 西南交通大学 一种混合驾驶环境下自动驾驶换道决策模型建立方法
CN110908375A (zh) * 2019-11-14 2020-03-24 北京三快在线科技有限公司 获取变道决策信息的方法、装置、存储介质和车辆
CN111383464A (zh) * 2018-12-28 2020-07-07 沈阳美行科技有限公司 车辆变道识别方法、装置、电子设备和介质
CN112406905A (zh) * 2020-09-10 2021-02-26 腾讯科技（深圳）有限公司 基于交通工具的数据处理方法、装置、计算机及存储介质

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3208786B1 (en) * 2016-02-22 2023-06-07 Volvo Car Corporation Method and system for evaluating inter-vehicle traffic gaps and time instances to perform a lane change manoeuvre
JP6629897B2 (ja) * 2018-02-26 2020-01-15 本田技研工業株式会社 車両制御装置
DE102018109883A1 (de) 2018-04-24 2018-12-20 Continental Teves Ag & Co. Ohg Verfahren und Vorrichtung zum kooperativen Abstimmen von zukünftigen Fahrmanövern eines Fahrzeugs mit Fremdmanövern zumindest eines Fremdfahrzeugs
JP6710722B2 (ja) * 2018-06-15 2020-06-17 本田技研工業株式会社 車両制御装置、車両制御方法、およびプログラム
US11091158B2 (en) 2018-06-24 2021-08-17 Mitsubishi Electric Research Laboratories, Inc. System and method for controlling motion of vehicle with variable speed
CN109739246B (zh) * 2019-02-19 2022-10-11 阿波罗智能技术(北京)有限公司 一种变换车道过程中的决策方法、装置、设备及存储介质
WO2020198189A1 (en) 2019-03-25 2020-10-01 Zoox, Inc. Pedestrian prediction based on attributes
CN110136254B (zh) * 2019-06-13 2019-12-13 吉林大学 基于动态概率驾驶地图的驾驶辅助信息显示方法
US11548512B2 (en) 2019-08-23 2023-01-10 Zoox, Inc. Yield behavior modeling and prediction
CN110910663A (zh) 2019-10-16 2020-03-24 清华大学 一种车路协同环境下多智能车交叉口通行协调控制方法
CN111169476B (zh) * 2020-01-18 2021-09-03 重庆长安汽车股份有限公司 一种运动趋势预测方法、装置、控制器及汽车
CN111428943B (zh) * 2020-04-23 2021-08-03 福瑞泰克智能***有限公司 障碍物车辆轨迹预测的方法、设备和计算机设备

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170291603A1 (en) * 2014-08-11 2017-10-12 Nissan Motor Co., Ltd. Travel Control Device and Method for Vehicle
CN108313054A (zh) * 2018-01-05 2018-07-24 北京智行者科技有限公司 自动驾驶自主换道决策方法和装置及自动驾驶车辆
CN108595823A (zh) * 2018-04-20 2018-09-28 大连理工大学 一种联合驾驶风格和博弈理论的自主车换道策略的计算方法
CN111383464A (zh) * 2018-12-28 2020-07-07 沈阳美行科技有限公司 车辆变道识别方法、装置、电子设备和介质
CN110298131A (zh) * 2019-07-05 2019-10-01 西南交通大学 一种混合驾驶环境下自动驾驶换道决策模型建立方法
CN110908375A (zh) * 2019-11-14 2020-03-24 北京三快在线科技有限公司 获取变道决策信息的方法、装置、存储介质和车辆
CN112406905A (zh) * 2020-09-10 2021-02-26 腾讯科技（深圳）有限公司 基于交通工具的数据处理方法、装置、计算机及存储介质

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4119412A4

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2024039997A1 (en) * 2022-08-15 2024-02-22 Motional Ad Llc Determination of an action for an autonomous vehicle in the presence of intelligent agents

Also Published As

Publication number Publication date
JP7520444B2 (ja) 2024-07-23
CN112406905B (zh) 2022-01-28
JP2023523350A (ja) 2023-06-02
US20230053459A1 (en) 2023-02-23
CN112406905A (zh) 2021-02-26
EP4119412A4 (en) 2023-10-11
EP4119412A1 (en) 2023-01-18

Similar Documents

Publication Publication Date Title
WO2022052856A1 (zh) 基于交通工具的数据处理方法、装置、计算机及存储介质
US11734832B1 (en) Prediction on top-down scenes based on object motion
US11631200B2 (en) Prediction on top-down scenes based on action data
US11875681B2 (en) Drive envelope determination
US10324469B2 (en) System and method for controlling motion of vehicle in shared environment
CN110562258B (zh) 一种车辆自动换道决策的方法、车载设备和存储介质
JP7200371B2 (ja) 車両速度を決定する方法及び装置
US20210173402A1 (en) Systems and methods for determining vehicle trajectories directly from data indicative of human-driving behavior
CN112888612A (zh) 自动驾驶车辆规划
US20180374359A1 (en) Evaluation framework for predicted trajectories in autonomous driving vehicle traffic prediction
CN110389584A (zh) 用于评估自动驾驶车辆的轨迹候选项的方法
EP4296133A1 (en) Intelligent driving method and apparatus, and storage medium and computer program
US10832439B1 (en) Locating entities in a mapped environment
US11597388B2 (en) System and method for connected vehicle lane merge
US11353877B2 (en) Blocked region guidance
Mandal et al. Lyft 3D object detection for autonomous vehicles
CN114906164A (zh) 用于自主驾驶的轨迹验证
Morales et al. Proactive driving modeling in blind intersections based on expert driver data
WO2024049925A1 (en) Trajectory prediction based on a decision tree
WO2023135271A1 (en) Motion prediction and trajectory generation for mobile agents
WO2021170140A1 (zh) 一种车道结构检测方法及装置
KR20230024392A (ko) 주행 의사 결정 방법 및 장치 및 칩
WO2022052556A1 (zh) 一种车辆行为预测方法、装置及车辆
US11919514B1 (en) Safety filter with preview data to improve the safety of steer commands
US20240174266A1 (en) Prediction model with variable time steps

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: 21865910

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2021865910

Country of ref document: EP

Effective date: 20221012

ENP Entry into the national phase

Ref document number: 2022565903

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE