WO2023070258A1 - 一种车辆的轨迹规划方法、装置及车辆 - Google Patents

一种车辆的轨迹规划方法、装置及车辆 Download PDF

Info

Publication number
WO2023070258A1
WO2023070258A1 PCT/CN2021/126059 CN2021126059W WO2023070258A1 WO 2023070258 A1 WO2023070258 A1 WO 2023070258A1 CN 2021126059 W CN2021126059 W CN 2021126059W WO 2023070258 A1 WO2023070258 A1 WO 2023070258A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
obstacle
trajectory
predicted trajectory
cost value
Prior art date
Application number
PCT/CN2021/126059
Other languages
English (en)
French (fr)
Inventor
王新宇
赵世杰
戴正晨
佘晓丽
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to CN202180057306.9A priority Critical patent/CN116457853A/zh
Priority to PCT/CN2021/126059 priority patent/WO2023070258A1/zh
Publication of WO2023070258A1 publication Critical patent/WO2023070258A1/zh

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units, or advanced driver assistance systems for ensuring comfort, stability and safety or drive control systems for propelling or retarding the vehicle
    • B60W30/08—Active safety systems predicting or avoiding probable or impending collision or attempting to minimise its consequences
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
    • G05D1/02—Control of position or course in two dimensions
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems

Definitions

  • the present application relates to the technical field of intelligent driving, and in particular to a vehicle trajectory planning method, device and vehicle.
  • Safety is the primary goal and pursuit of autonomous driving vehicles.
  • the automatic driving system is usually divided into three modules according to functions: perception, planning, and control. Based on the information of the environment, make corresponding decisions, and plan a collision-free drivable trajectory, and the control module is used to control the actuator to execute the planned results.
  • the planning module the planned drivable trajectory needs to consider the following factors from the perspective of safety: one is to satisfy the dynamics and kinematics constraints of the vehicle, and the other is that the planned trajectory does not collide with other obstacles, including Stationary and moving obstacles.
  • responsibility-sensitive safety model for automatic driving, which converts human drivers' ideas and concepts of safe driving into corresponding mathematical formulas, so as to help the division of responsibilities in traffic accidents.
  • the responsibility-sensitive safety model will check whether the result of the plan is safe. If the responsibility-sensitive safety model checks the plan If the result is not safe, return to the planning module for recalculation.
  • the responsibility-sensitive safety model plans the driving trajectory based on the assumption that other cars take aggressive actions, which will lead to the behavior of the own car tending to be conservative, and the user's comfort for the automatic driving experience is not good enough.
  • the present application provides a vehicle trajectory planning method, device and vehicle, so as to improve the comfort of automatic driving experience while ensuring the driving safety of the vehicle.
  • an embodiment of the present application provides a vehicle trajectory planning method, the method includes: the vehicle trajectory planning device obtains the first predicted trajectory of the obstacle, when the obstacle is the game target, according to the first predicted trajectory and Based on the information of the interaction scene between the obstacle and the vehicle, a feasible trajectory cluster of the obstacle is generated, and then a second predicted trajectory is selected from the feasible trajectory cluster, wherein the cost value of the second predicted trajectory is greater than the preset threshold, according to the second predicted trajectory, Generate the space-time constraints of the vehicle, and plan the trajectory of the vehicle based on the space-time constraints of the vehicle.
  • the trajectory planning device of the vehicle can generate a feasible trajectory that is safe to drive for the obstacle based on the first predicted trajectory of the obstacle and the information of the interaction scene between the obstacle and the vehicle cluster, and then according to the second predicted trajectory selected from the feasible trajectory cluster whose substitution value is greater than the preset threshold, the space-time constraints of the vehicle are generated, and the trajectory of the vehicle is planned based on the space-time constraints, which can make the self-vehicle as far as possible while maintaining safe driving.
  • the comfort of the first vehicle is achieved, so as to ensure the driving safety of the first vehicle and improve the comfort of the automatic driving experience of the first vehicle.
  • generating a feasible trajectory cluster of the obstacle includes: according to the first predicted trajectory, the information of the interaction scene and the sampling limit constraint, The lateral offset sampling and longitudinal acceleration sampling are respectively carried out to obtain the sampling data, and the feasible trajectory clusters of obstacles are generated according to the sampling data.
  • a series of feasible trajectories for obstacles that may be traveled in the future can be obtained through lateral offset sampling and longitudinal acceleration sampling.
  • a second predicted trajectory from the feasible trajectory cluster before selecting a second predicted trajectory from the feasible trajectory cluster, it may also include: determining each candidate in the feasible trajectory cluster according to the information of the interaction scene and the mapping relationship set corresponding to the information of the interaction scenario
  • the cost value of each evaluation dimension corresponding to the prediction trajectory, the mapping relationship set includes at least one mapping relationship, each mapping relationship includes the mapping relationship between the preset evaluation dimension and the preset cost value, according to each candidate prediction trajectory in the feasible trajectory cluster
  • Corresponding to the cost value of each evaluation dimension determine the cost value of each candidate prediction trajectory.
  • the candidate predicted trajectories in the feasible trajectory cluster can be evaluated from multiple evaluation dimensions, and then the trajectory planning device can determine the cost value of the candidate predicted trajectories.
  • mapping relationship set corresponding to the information of the interaction scene satisfies at least one of the following:
  • the right of way of the obstacle is higher than that of the vehicle, and the higher the right of way of the obstacle, the greater the cost value of the right of way corresponding to the candidate predicted trajectory;
  • the trafficability of obstacles is higher than that of vehicles, and the higher the trafficability of obstacles, the greater the cost value of trafficity corresponding to candidate predicted trajectories;
  • the safety of obstacles is higher than that of vehicles, and the higher the safety of obstacles, the greater the cost value of the safety corresponding to candidate predicted trajectories;
  • the comfort of the obstacle is higher than that of the vehicle, and the higher the comfort of the obstacle, the greater the cost value of the comfort corresponding to the candidate predicted trajectory;
  • the trajectory planning device of the vehicle may further include: when the obstacle is a non-game target, the trajectory planning device of the vehicle generates a vehicle trajectory according to the first predicted trajectory.
  • the space-time constraints of the vehicle based on the space-time constraints of the vehicle, plan the vehicle's driving trajectory.
  • the embodiment of the present application also provides a trajectory planning device, and the electronic device includes modules/units that implement the above-mentioned first aspect and any possible design method of the first aspect; these modules/units It can be realized by hardware, and it can also be realized by executing corresponding software by hardware.
  • the trajectory planning device includes an acquisition unit and a processing unit; wherein:
  • an acquisition unit configured to acquire the first predicted trajectory of the target obstacle
  • the processing unit is configured to generate a cluster of feasible trajectories of the obstacle according to the information of the first predicted trajectory and the interaction scene between the obstacle and the vehicle if the obstacle is the game target, and select a second predicted trajectory from the cluster of feasible trajectories.
  • the cost value of the second predicted trajectory is greater than a preset threshold, and the space-time constraints of the vehicle are generated according to the second predicted trajectory; and the driving trajectory of the vehicle is planned based on the space-time constraints of the vehicle.
  • the mapping relationship set corresponding to the information of the interaction scene satisfies at least one of the following: in the interaction scene, the right of way of the obstacle is higher than the right of way of the vehicle, and the higher the right of way of the obstacle, The greater the cost value of the right of way corresponding to the candidate predicted trajectory; in the interactive scene, the trafficability of obstacles is higher than that of vehicles, and the higher the trafficability of obstacles, the higher the cost value of traffic corresponding to candidate predicted trajectories. Large; the safety of obstacles is higher than that of vehicles in interactive scenarios, and the higher the safety of obstacles, the greater the cost of safety corresponding to candidate predicted trajectories; in interactive scenarios, the comfort of obstacles is higher than that of vehicles.
  • the processing unit is specifically configured to: generate a first encroachment area of the obstacle on the expected driving path of the vehicle on the SL coordinate system according to the second predicted trajectory, and generate a lateral offset area according to the first encroachment area.
  • the displacement constraint and the radius of curvature constraint ; generate the second encroachment area of the obstacle on the expected driving path of the vehicle on the ST coordinate system according to the second predicted trajectory, and generate the speed constraint, acceleration constraint and jerk constraint according to the second encroachment area.
  • the processing unit is further configured to: when the obstacle is a non-game target, generate a space-time constraint of the vehicle according to the first predicted trajectory; plan a trajectory of the vehicle based on the space-time constraint of the vehicle.
  • the embodiment of the present application also provides a trajectory planning device.
  • the trajectory planning device includes a processor and a memory, and the memory is used to store computer-executable instructions.
  • the processor executes the The computer in the memory executes instructions to use the hardware resources in the trajectory planning device to execute the operation steps of the above-mentioned first aspect and the method of any possible design of the first aspect.
  • a program product in the embodiment of the present application includes instructions, and when the program product is run on the vehicle-mounted device, the vehicle-mounted device executes the first aspect of the embodiment of the present application and any of the first aspects.
  • a chip system may include a processor.
  • the processor is coupled with the memory and can be used to execute the first aspect and the method in any possible implementation manner of the first aspect.
  • the chip system further includes a memory.
  • Memory used to store computer programs (also called code, or instructions).
  • the processor is configured to call and run a computer program from the memory, so that the device installed with the system-on-a-chip executes the first aspect and the method in any possible implementation manner of the first aspect.
  • the above trajectory planning device can be a chip
  • the input circuit can be an input pin
  • the output circuit can be an output pin
  • the processing circuit can be a transistor, a gate circuit, a flip-flop, and various logic circuits.
  • the input signal received by the input circuit may be received and input by, for example but not limited to, the receiver
  • the output signal of the output circuit may be, for example but not limited to, output to the transmitter and transmitted by the transmitter
  • the circuit may be the same circuit, which is used as an input circuit and an output circuit respectively at different times.
  • the embodiment of the present application does not limit the specific implementation manners of the processor and various circuits.
  • FIG. 1 is a schematic diagram of a vehicle driving scene provided by an embodiment of the present application
  • FIG. 2 is a schematic flowchart of a vehicle trajectory planning method provided in an embodiment of the present application
  • 3 to 7 are schematic diagrams of the types of obstacles in multiple scenarios provided by the embodiments of the present application.
  • FIG. 8 is a schematic diagram of boundary constraints for lateral path planning provided by an embodiment of the present application.
  • FIG. 9 is a schematic diagram of speed planning based on an ST diagram provided in the embodiment of the present application.
  • Fig. 10 is a schematic diagram of candidate prediction trajectories of the game object provided by the embodiment of the present application.
  • FIG. 13 is a schematic diagram of a trajectory planning device provided by an embodiment of the present application.
  • the vehicle can communicate with other objects based on the wireless communication technology between the vehicle and the outside world (for example, vehicle to everything (V2X)).
  • V2X vehicle to everything
  • the communication between the vehicle and other objects can be realized based on inter-vehicle wireless communication technology (eg, vehicle to vehicle (V2V)).
  • the communication between vehicles and other objects can be based on Wi-Fi (wireless fidelity), fifth generation (5th generation, 5G) mobile communication technology, long term evolution (long term evolution, LTE) and so on.
  • the driving automation grading standards proposed by the Society of Automotive Engineers International include six levels including L0-L5, among which, L0-L2,
  • the driver support system can provide some support functions for the driver, but regardless of whether the driver support function of the vehicle is turned on or not, the driver must drive the vehicle himself and supervise these support functions provided by the driver support system at all times, which must be carried out as needed Steering, braking, or accelerating to ensure safety, the difference between L0-L2 level support functions is: L0 level is no driving automation, support functions are limited to providing warnings and instant assistance, L1 level support functions provide the driver with steering or Braking/accelerating support, L2 support functions provide steering and braking/accelerating support for the driver.
  • the automatic driving system can complete certain driving tasks and monitor the driving environment under certain circumstances, but the driver needs to be ready to regain driving control at any time, for example, the driver must drive when the function is requested.
  • L4 level highly automatic driving the automatic driving system can complete driving tasks and monitor the driving environment under certain environments and specific conditions.
  • L5 level fully automatic driving the automatic driving system can complete all driving tasks under all conditions.
  • any vehicle can constrain the trajectory planning of the vehicle by sensing the driving trajectory of the surrounding vehicles within a preset period of time, thereby improving safety.
  • the safety factors that need to be considered by the planning module are transformed into constraints in the optimization problem, that is, for the planned trajectory, it needs to satisfy:
  • the planned trajectory needs to meet the executable criteria, that is, the planned trajectory of the first vehicle needs to meet the dynamic constraints and kinematic constraints of the ego vehicle, where the dynamic constraints include the maximum rotation angle of the steering wheel, the maximum speed, and the kinematic constraints Including longitudinal maximum acceleration and maximum jerk.
  • the planned trajectory also needs to meet the no-collision criterion, that is, the planned trajectory does not collide with other obstacles, and then, according to whether the obstacle allocates enough attention to the own vehicle to interact with the own vehicle, and the obstacle's Predict trajectories and generate spatio-temporal constraints for the ego vehicle.
  • the scene includes a first vehicle and a second vehicle located in the right front of the first vehicle.
  • the first vehicle has a partial or complete Vehicles capable of automatic driving
  • the first vehicle can be a vehicle of L2 level or above in the above-mentioned driving automation grading standard
  • the second vehicle can be any one of L0-L5 in the above-mentioned driving automation grading standard Vehicles, semi-motor vehicles, motorcycles, etc.
  • the first vehicle can obtain the vehicle information of the second vehicle based on the sensor of the own vehicle, such as camera, laser radar, millimeter wave radar, Global Navigation Satellite System (Global Navigation Satellite System, GNSS), etc.
  • GPS Global Navigation Satellite System
  • Step 201 the trajectory planning device of the first vehicle obtains the first predicted trajectory of the obstacle.
  • the first vehicle can acquire the vehicle information of the second vehicle based on the sensor of its own vehicle, such as camera, laser radar, millimeter wave radar, GNSS, etc.
  • the vehicle information of the second vehicle may include but not limited to the geographical location, driving speed, driving direction, turn signal information and other information of the second vehicle, and the first vehicle may also obtain surrounding traffic indication information, such as traffic light information, traffic indication sign information .
  • the first vehicle may predict the first predicted trajectory of the second vehicle in combination with information such as vehicle information of the second vehicle, current road information, and traffic instruction information.
  • Step 202 the trajectory planning device generates the space-time constraints of the first vehicle based on the first predicted trajectory and the type of obstacles.
  • step 202 if the type of obstacle is a non-game object, that is, an obstacle that does not need to allocate enough attention to the first vehicle, the trajectory planning device generates the first vehicle according to the first predicted trajectory space-time constraints.
  • the centerline of the road is used as the reference line, and the direction along which the first vehicle is along the reference line is called the longitudinal direction, that is, the S coordinate.
  • the normal direction of the line is the L coordinate
  • the distance between the projection point of the position of the first vehicle on the reference line and the position of the first vehicle is the lateral offset displacement
  • the distance between the starting point of the first vehicle driving and the projection point is The curve distance of is the longitudinal displacement. Since the first vehicle is constantly moving forward along the reference line, the lateral displacement L of the vehicle is constantly changing as the longitudinal displacement S changes.
  • boundary constraints can be used for lateral path planning:
  • l(s) is the lateral deviation constraint of the path.
  • the maximum lateral deviation that can be driven along the S axis can be calculated. shift.
  • Table 1 shows an example of the corresponding relationship between the vehicle speed, the maximum front wheel rotation angle and the maximum front wheel rotation speed.
  • k(s) is the curvature radius constraint of the path, ensuring that the curvature of each point on the path meets the minimum turning radius limit of the vehicle.
  • k(s) is related to the physical parameters of the vehicle (such as the turning limit of the vehicle, vehicle type, etc.), the state of motion, the friction of the road surface, etc.
  • the horizontal axis T is the time axis
  • the vertical axis S is the direction along the expected driving path of the first vehicle (that is, the reference line).
  • the parallelogram area on the ST diagram is based on The encroachment area on the expected travel path of the first vehicle (self-vehicle) generated by the predicted trajectory of the obstacle (other vehicle), and the curve is the speed planning curve of the first vehicle.
  • v(t), a(t), and jerk(t) are the upper and lower bound constraints of the speed, acceleration, and jerk allowed by the first vehicle, respectively, where v(t) is related to the current road speed limit (for example, the maximum speed limit of the expressway 120km/h), environmental risk speed limit (for example, the speed limit at the gate of the school is 30km/h, and for example, the maximum speed limit is 60km/h when the visibility is within 200 meters in foggy weather) and other factors.
  • v(t) is a positive value
  • v(t) is a negative value when the first vehicle reverses.
  • the first vehicle is driving on a city road
  • v min is
  • v max is the maximum speed limit of the city road.
  • v min of the expressway may be 60 km/h
  • v max may be 120 km/h.
  • a(t) is the acceleration.
  • a(t) takes a positive value
  • a(t) takes a negative value.
  • a min can be -4m/s 2
  • a max can be 3m/s 2 .
  • jerk(t) is the derivative of acceleration with respect to time, representing the rate of change of acceleration.
  • the value of jerk(t) is related to the current speed and acceleration.
  • jerk(t) can take a positive number, which means that the larger the value, the greater the acceleration, and jerk(t) can also take a negative number, and the larger the value, the smaller the acceleration.
  • the jerk min can be -10m/s 3
  • the jerk max can be 1.5m/s 3 .
  • the speed of the first vehicle is limited between v min and v max , so that the speed of the vehicle can neither be too low nor overspeed, and the acceleration of the first vehicle is limited between a min and a max , so that the acceleration It will not exceed the maximum acceleration capability of the first vehicle, and the jerk of the first vehicle is limited between jerk min and jerk max , so that the acceleration of the first vehicle will not change violently, so that according to the space-time constraints of the first vehicle Planning the driving trajectory of the first vehicle can ensure the safety of the vehicle while improving the comfort of the automatic driving experience.
  • the planned driving trajectory of the first vehicle can be made to satisfy the above-mentioned executable criterion.
  • the trajectory planning device can be based on the obstacle
  • the first predicted trajectory (the first predicted trajectory can be understood as the trajectory with the highest probability of the obstacle’s future driving) and the information of the interaction scene between the obstacle and the first vehicle, generate a series of candidate predictions for the possible future driving of the obstacle Trajectories can also be called feasible trajectory clusters, and then select a trajectory with a relatively high cost value for obstacles from the feasible trajectory clusters as the second predicted trajectory, which can be understood as the A trajectory that is less comfortable in terms of material.
  • the encroachment area of the obstacle is generated, the space-time constraint of the first vehicle is determined based on the encroachment area of the obstacle, and the driving trajectory of the first vehicle is planned based on the space-time constraint of the first vehicle.
  • the driving track of the vehicle makes the first vehicle as comfortable as possible under the premise of maintaining safe driving.
  • the trajectory planning device may perform lateral offset sampling and longitudinal acceleration sampling according to the first predicted trajectory to generate a series of candidate predicted trajectories that the obstacle may travel in the future, that is, a cluster of feasible trajectories.
  • sampling limit constraints can be set.
  • the sampling limit constraints can include sampling limit constraints corresponding to lateral offset sampling and sampling limit constraints corresponding to longitudinal acceleration sampling, and lateral offset sampling corresponding to Sampling constraints include but are not limited to information such as road boundaries, vehicle speeds, and whether it is an intersection scene.
  • Sampling constraints corresponding to longitudinal acceleration sampling include but are not limited to jerk, acceleration, and speed constraints, as well as information such as different target types.
  • the trajectory planning device can perform lateral offset sampling and longitudinal acceleration sampling respectively according to the first predicted trajectory, the information of the interaction scene between the obstacle and the vehicle, and the sampling limit constraints to obtain the sampling data, and generate the obstacle's possible future travel according to the sampling data.
  • At least one candidate predicted trajectory of the obstacle and then determine a candidate predicted trajectory with a relatively high cost value for the obstacle from at least one candidate predicted trajectory that the obstacle may travel in the future, as the second predicted trajectory of the obstacle.
  • the lateral offset sampling and longitudinal acceleration sampling are respectively performed to obtain the sampling data as shown in Figure 11, and the first row in the table described in Figure 11 represents the longitudinal acceleration sampling
  • a series of accelerations used are accelerations, and the first column indicates that the lateral offset sampling adopts a lateral offset distance of -3 ⁇ 3m respectively.
  • the ellipsis in each grid indicates the coordinates of multiple track points obtained by sampling with a certain acceleration and a certain lateral offset distance and the speed information of other vehicles.
  • the longitudinal acceleration is sampled with an acceleration of -4m/ s2 .
  • the lateral offset distance of 3m is used for lateral offset sampling to obtain the coordinates of multiple track points of other vehicles and the speed information of other vehicles at each track point, and store them in the grid in the second row and second column.
  • a series of feasible trajectory clusters are generated, that is, at least one candidate prediction of the possible future travel of other vehicles Trajectories, for example, three candidate prediction trajectories are planned in Figure 11.
  • trajectory The planning device can evaluate each candidate predicted trajectory from multiple evaluation dimensions to obtain the cost value of each candidate predicted trajectory, and then determine that the cost value of the obstacle is greater than the cost threshold from at least one candidate predicted trajectory that may be driven in the future
  • the candidate predicted trajectory of is used as the second predicted trajectory, so that the first vehicle generates the space-time constraints of the first vehicle according to the second predicted trajectory.
  • the trajectory planning device may also store a set of mapping relationships, or the trajectory planning device obtains a set of mapping relationships from other storage devices of the first vehicle, the set of mapping relationships includes at least one mapping relationship, and each mapping relationship includes The mapping relationship between preset evaluation dimensions and preset cost values. For example, suppose there are seven preset evaluation dimensions, namely, right of way, traffic, safety, comfort, attention, target type, target portrait, then the mapping relationship set includes seven mapping relationships. The various evaluation dimensions will be introduced later and will not be repeated here.
  • the information of each interaction scene may correspond to a mapping relationship set
  • the trajectory planning device may determine at least one candidate The cost value of each evaluation dimension corresponding to each candidate prediction trajectory in the prediction trajectory. Then, for each candidate predicted trajectory, the cost value of the candidate predicted trajectory may be determined according to the cost values of the evaluation dimensions corresponding to the candidate predicted trajectory.
  • x i is the cost value of different evaluation dimensions, the value is between 0 and 1, and w i is the weight value of the corresponding dimension.
  • the weight value of a certain evaluation dimension refers to the The relative importance of w i can take any value, for example, the value of w i is between 0 and 1, and the specific value of w i is not limited here, for example, the weight of comfort is 0.3 to 0.4.
  • Trajectory cost is the final cost value of each candidate prediction trajectory.
  • the trajectory planning device determines the second predicted trajectory from the candidate predicted trajectory whose cost value is greater than the cost threshold in at least one candidate predicted trajectory. For example, from the candidate predicted trajectories whose cost value is greater than the cost threshold, a candidate predicted trajectory with the largest cost value is determined as the second predicted trajectory, and a candidate predicted trajectory with the largest cost value can be understood as an obstacle for at least one candidate predicted trajectory A candidate prediction trajectory with the lowest comfort. For another example, from the candidate prediction trajectories whose cost value is greater than the cost threshold, determine N candidate prediction trajectories with larger cost values, N is an integer greater than 1, and then select a candidate prediction trajectory from the N candidate prediction trajectories as Second predicted trajectory.
  • Right of way refers to the right of an object to travel within a certain space and time. Examples of the right of way in different scenarios are as follows:
  • Non-intersection scene rules zebra crossing > go straight in the lane > change lanes > go straight in the opposite direction > cross > gap.
  • the trafficability of the first vehicle is high.
  • the higher the passability of the obstacle the more inclined it is to maintain the current state.
  • the passability of the obstacle is higher than that of the first vehicle, and the higher the passability of the obstacle, the greater the cost value of the passability corresponding to the candidate predicted trajectory.
  • Safety the safety of the interaction process between the first vehicle and the obstacle.
  • the distance between the first vehicle and the obstacle is always kept greater than a certain threshold.
  • the safety is high.
  • the first vehicle A collision is sent during the interaction between a vehicle and an obstacle, which is the least safe case.
  • the higher the security of the obstacle, the more inclined it is to maintain the current state. Examples of security sizes for different scenarios are as follows:
  • the comfort of the obstacle is higher than that of the first vehicle, and the higher the comfort of the obstacle, the greater the cost value of the comfort corresponding to the candidate predicted trajectory.
  • the trajectory is evaluated according to the degree of attention of the obstacle to the first vehicle.
  • the lower the degree of attention of the obstacle to the first vehicle the more inclined it is to maintain the current state.
  • the lower the obstacle's attention to the first vehicle the greater the cost value of the attention corresponding to the candidate predicted trajectory.
  • changeability can be understood as the ability to change the current state of motion.
  • a child has no ability to change the current state of motion.
  • the changeability of a child is very low, and the corresponding cost is large, while the small car’s
  • the changeability is high, and the corresponding cost is small.
  • Obstacle image Evaluate the trajectory based on the image of the obstacle. The more dangerous the image of the obstacle is, the more the obstacle tends to maintain the current state.
  • the example of the rule is as follows:
  • the obstacle portrait can be determined according to the state management of the obstacle in a life cycle. For example, if the obstacle is a smart car, its portrait can be determined according to the behavior of the smart car. For example, the smart car needs to turn left. If the smart car It is a forced left turn on the straight lane and blocked to the left turn lane, and its portrait can be considered dangerous. In some other examples, the portrait can also be determined according to the behavior of the smart car, the type of the smart car, the type of the driver, and what the driver is currently doing. The type of the smart car can be, for example, a sports car, a family car wait. The more dangerous the target image of the obstacle, the greater the cost value of the target image corresponding to the candidate predicted trajectory.
  • the trajectory planning device generates the second encroachment area of the obstacle on the expected travel path of the first vehicle on the ST coordinate system according to the second predicted trajectory, and generates the speed constraint, the acceleration constraint and the jerk constraint according to the second encroachment area, immediately Time constraints in null constraints.
  • the relevant content about generating the space-time constraints of the first vehicle according to the first predicted trajectory in the above-mentioned possible implementation a1, which will not be repeated here repeat.
  • Step 203 Determine the trajectory plan of the first vehicle according to the space-time constraints of the first vehicle.
  • the trajectory planning device may plan the driving trajectory of the first vehicle based on the space-time constraints of the first vehicle generated by the first predicted trajectory after generating the space-time constraints of the first vehicle according to the first predicted trajectory.
  • the trajectory planning device may plan the driving trajectory of the first vehicle based on the space-time constraints of the first vehicle generated by the second predicted trajectory after generating the space-time constraints of the first vehicle according to the second predicted trajectory.
  • the obstacles are divided into non-game goals and game goals.
  • non-game goals obstacles There is no interaction with the self-vehicle, and the space-time constraints of the self-vehicle are generated according to the first predicted trajectory of the obstacle; while for the game target, the obstacle interacts with the self-vehicle, and the feasible trajectory cluster of the obstacle is generated according to the interaction scene (that is, at least one of the above Candidate predicted trajectories), and then based on the feasible trajectories selected from the feasible trajectories clusters that are more costly for obstacles (that is, the above-mentioned second predicted trajectories), the space-time constraints of the self-vehicle are generated. Under the space-time constraints, it is possible to make The planned trajectory of the ego vehicle is as comfortable as possible while maintaining safe driving.
  • FIG. 12 is a schematic diagram of a trajectory planning device provided in an embodiment of the present application.
  • the trajectory planning device 1200 can implement the steps performed by the trajectory planning device in the above method embodiment.
  • the trajectory planning device may include an acquisition unit 1201 and a processing unit 1202 .
  • An acquisition unit 1201 configured to acquire a first predicted trajectory of the obstacle
  • the processing unit 1202 is configured to generate a feasible trajectory cluster of the obstacle according to the information of the first predicted trajectory and the interaction scene between the obstacle and the vehicle if the obstacle is the game target; select a second predicted trajectory from the feasible trajectory cluster, The cost value of the second predicted trajectory is greater than a preset threshold; according to the second predicted trajectory, a space-time constraint of the vehicle is generated, and the space-time constraint of the vehicle is used for the vehicle to plan the driving trajectory of the own vehicle.
  • the processing unit 1202 is specifically configured to: respectively perform lateral offset sampling and longitudinal acceleration sampling according to the first predicted trajectory, information of the interaction scene, and sampling restriction constraints to obtain sampling data; generate obstacle data according to the sampling data cluster of feasible trajectories.
  • the processing unit 1202 is further configured to: determine the cost value of each evaluation dimension corresponding to each candidate predicted trajectory in the feasible trajectory cluster according to the information of the interaction scene and the mapping relationship set corresponding to the information of the interaction scenario , the mapping relationship set includes at least one mapping relationship, and each mapping relationship includes a mapping relationship between a preset evaluation dimension and a preset cost value; according to the cost value of each evaluation dimension corresponding to each candidate prediction trajectory in the feasible trajectory cluster, determine The cost value of each candidate predicted trajectory.
  • mapping relationship set corresponding to the information of the interaction scene satisfies at least one of the following:
  • the right of way of the obstacle is higher than that of the vehicle, and the higher the right of way of the obstacle, the greater the cost value of the right of way corresponding to the candidate predicted trajectory;
  • the trafficability of obstacles is higher than that of vehicles, and the higher the trafficability of obstacles, the greater the cost value of trafficity corresponding to candidate predicted trajectories;
  • the safety of obstacles is higher than that of vehicles, and the higher the safety of obstacles, the greater the cost value of the safety corresponding to candidate predicted trajectories;
  • the comfort of the obstacle is higher than that of the vehicle, and the higher the comfort of the obstacle, the greater the cost value of the comfort corresponding to the candidate predicted trajectory;
  • the processing unit 1202 is specifically configured to: generate a first encroachment area of the obstacle on the expected driving path of the vehicle on the SL coordinate system according to the second predicted trajectory, and generate a lateral offset area according to the first encroachment area.
  • the displacement constraint and the radius of curvature constraint ; generate the second encroachment area of the obstacle on the expected driving path of the vehicle on the ST coordinate system according to the second predicted trajectory, and generate the speed constraint, acceleration constraint and jerk constraint according to the second encroachment area.
  • the processing unit 1202 is also configured to: when the obstacle is a non-game target, generate the space-time constraints of the vehicle according to the first predicted trajectory, and the space-time constraints of the vehicle are used for the vehicle to plan the driving trajectory of its own vehicle .
  • FIG. 13 is a schematic structural diagram of a trajectory planning device provided in an embodiment of the present application. As shown in FIG. The memory 1301 can be connected through a bus system.
  • the above processor 1302 may be a chip.
  • the processor 1302 may be a field programmable gate array (field programmable gate array, FPGA), may be an application specific integrated circuit (ASIC), may also be a system chip (system on chip, SoC), or It can be a central processing unit (central processor unit, CPU), or a network processor (network processor, NP), or a digital signal processing circuit (digital signal processor, DSP), or a microcontroller (micro controller) unit, MCU), it can also be a programmable controller (programmable logic device, PLD) or other integrated chips.
  • FPGA field programmable gate array
  • ASIC application specific integrated circuit
  • SoC system on chip
  • CPU central processing unit
  • NP network processor
  • DSP digital signal processing circuit
  • microcontroller micro controller
  • MCU microcontroller
  • PLD programmable logic device
  • each step of the above method may be completed by an integrated logic circuit of hardware in the processor 1302 or instructions in the form of software.
  • the steps of the methods disclosed in connection with the embodiments of the present application may be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules in the processor 1302 .
  • the software module can be located in a mature storage medium in the field such as random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, register.
  • the storage medium is located in the memory 1301, and the processor 1302 reads the information in the memory 1301, and completes the steps of the above method in combination with its hardware.
  • the processor 1302 in the embodiment of the present application may be an integrated circuit chip, which has a signal processing capability.
  • each step of the above-mentioned method embodiments may be completed by an integrated logic circuit of hardware in a processor or instructions in the form of software.
  • the above-mentioned processor may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components .
  • DSP digital signal processor
  • ASIC application-specific integrated circuit
  • FPGA field-programmable gate array
  • Various methods, steps, and logic block diagrams disclosed in the embodiments of the present application may be implemented or executed.
  • a general-purpose processor may be a microprocessor, or the processor may be any conventional processor, or the like.
  • the steps of the method disclosed in connection with the embodiments of the present application may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor.
  • the software module can be located in a mature storage medium in the field such as random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, register.
  • the storage medium is located in the memory, and the processor reads the information in the memory, and completes the steps of the above method in combination with its hardware.
  • the memory 1301 in the embodiment of the present application may be a volatile memory or a nonvolatile memory, or may include both volatile and nonvolatile memories.
  • the non-volatile memory can be read-only memory (read-only memory, ROM), programmable read-only memory (programmable ROM, PROM), erasable programmable read-only memory (erasable PROM, EPROM), electrically programmable Erases programmable read-only memory (electrically EPROM, EEPROM) or flash memory.
  • Volatile memory can be random access memory (RAM), which acts as external cache memory.
  • RAM random access memory
  • SRAM static random access memory
  • DRAM dynamic random access memory
  • DRAM synchronous dynamic random access memory
  • SDRAM double data rate synchronous dynamic random access memory
  • ESDRAM enhanced synchronous dynamic random access memory
  • SLDRAM direct memory bus random access memory
  • direct rambus RAM direct rambus RAM
  • the present application also provides a vehicle, which may include the trajectory planning device mentioned above.
  • the vehicle may be the first vehicle involved in this application.
  • the present application also provides a computer program product, the computer program product including: computer program code or instruction, when the computer program code or instruction is run on the computer, the computer is made to execute the above method The method of any one of the embodiments in the embodiments.
  • the present application also provides a computer-readable storage medium, the computer-readable medium stores program codes, and when the program codes are run on a computer, the computer executes the method described in the above-mentioned embodiments. The method of any one of the embodiments.
  • the present application further provides a chip system, where the chip system may include a processor.
  • the processor is coupled with the memory, and can be used to execute the method in any one of the above method embodiments.
  • the chip system further includes a memory.
  • Memory used to store computer programs (also called code, or instructions).
  • the processor is configured to invoke and run a computer program from the memory, so that the device installed with the system-on-a-chip executes the method in any one of the above method embodiments.
  • the above-mentioned embodiments may be implemented in whole or in part by software, hardware, firmware or other arbitrary combinations.
  • the above-described embodiments may be implemented in whole or in part in the form of computer program products.
  • the computer program product includes one or more computer instructions. When the computer program instructions are loaded or executed on the computer, the processes or functions according to the embodiments of the present application will be generated in whole or in part.
  • the computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable devices.
  • the computer instructions may be stored in or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from a website, computer, server or data center Transmission to another website site, computer, server, or data center by wired (eg, coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (eg, infrared, wireless, microwave, etc.).
  • the computer-readable storage medium may be any available medium that can be accessed by a computer, or a data storage device such as a server or a data center that includes one or more sets of available media.
  • the available media may be magnetic media (eg, floppy disk, hard disk, magnetic tape), optical media (eg, DVD), or semiconductor media.
  • the semiconductor medium may be a solid state drive (SSD).
  • the disclosed systems, devices and methods may be implemented in other ways.
  • the device embodiments described above are only illustrative.
  • the division of the units is only a logical function division. In actual implementation, there may be other division methods.
  • multiple units or components can be combined or May be integrated into another system, or some features may be ignored, or not implemented.
  • the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be in electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to realize the purpose of the technical solution of the present application.

Abstract

一种车辆的轨迹规划方法、装置（1200，1300）及车辆，轨迹规划方法包括：获取障碍物的第一预测轨迹（201），当障碍物为博弈目标时，根据第一预测轨迹以及障碍物与车辆的交互场景的信息，生成障碍物的可行轨迹簇；从可行轨迹簇中选取一条第二预测轨迹，第二预测轨迹的代价值大于预设阈值；据第二预测轨迹，生成车辆的时空约束（202）；基于车辆的时空约束，规划车辆的行驶轨迹（203）。车辆在时空约束规划自车的行驶轨迹，使得自车在保持安全行驶的前提下尽可能提高舒适性。

Description

一种车辆的轨迹规划方法、装置及车辆 技术领域
本申请涉及智能驾驶技术领域，尤其涉及一种车辆的轨迹规划方法、装置及车辆。
背景技术
安全性是自动驾驶车辆的首要目标和追求，通常将自动驾驶系统按照功能划分为感知、规划、控制三个模块，其中，感知模块用于获取车辆周围的环境信息，规划模块需要根据获取的周围环境的信息，做出相应的决策，并规划出一个无碰撞的可行驶轨迹，控制模块用于控制执行器执行规划的结果。在规划模块中，规划出的可行驶轨迹从安全性的角度出发需要考虑以下因素：一个是满足自车的动力学和运动学约束，另一个是规划的轨迹不与其他障碍物发生碰撞，包括静止障碍物和运动障碍物。
为保证规划结果的安全性，业界提出了一种责任敏感安全模型用于自动驾驶，将人类驾驶员对安全驾驶的理念和概念转化为对应的数学公式，以此来帮助交通事故的责任划分。使用责任敏感安全模型，即使他车采取激进的动作，也确保自车不会主动导致事故，并尽可能避免事故发生，责任敏感安全模型会检查规划的结果是否安全，若责任敏感安全模型检查规划的结果不安全，则返回规划模块重新计算。责任敏感安全模型基于他车采取激进动作的假设进行行驶轨迹规划，会导致自车的行为倾向于保守，用户对于自动驾驶体验的舒适性不够好。
发明内容
本申请提供一种车辆的轨迹规划方法、装置及车辆，以保证车辆行驶安全性的同时，提高自动驾驶体验的舒适性。
第一方面，本申请实施例提供了一种车辆的轨迹规划方法，该方法包括：车辆的轨迹规划装置获取障碍物的第一预测轨迹，当障碍物为博弈目标时，根据第一预测轨迹以及障碍物与车辆的交互场景的信息，生成障碍物的可行轨迹簇，然后从可行轨迹簇中选取一条第二预测轨迹，其中第二预测轨迹的代价值大于预设阈值，根据第二预测轨迹，生成车辆的时空约束，基于车辆的时空约束，规划车辆的行驶轨迹。
通过该方法，当障碍物为博弈目标时，车辆的轨迹规划装置可以根据障碍物的第一预测轨迹、以及障碍物与车辆的交互场景的信息，生成对于障碍物而言可安全行驶的可行轨迹簇，然后根据从可行轨迹簇中选取代价值大于预设阈值的第二预测轨迹，生成车辆的时空约束，基于此时空约束规划车辆的行驶轨迹，可以使得自车在保持安全行驶的前提下尽可能的具有舒适性，从而实现在保证第一车辆行驶安全性的同时，提高第一车辆自动驾驶体验的舒适性。
在一种可能的实现方式中，根据第一预测轨迹以及障碍物与车辆的交互场景的信息，生成障碍物的可行轨迹簇，包括：根据第一预测轨迹、交互场景的信息以及采样限制约束，分别进行横向偏移采样和纵向加速度采样得到采样数据，根据采样数据生成障碍物的可行轨迹簇。在该实现方式中，可以通过横向偏移采样和纵向加速度采样可以获得对于障碍物而言未来可能行驶的一系列可行轨迹。
在一种可能的实现方式中，从可行轨迹簇中选取一条第二预测轨迹之前，还可以包括：根据交互场景的信息以及交互场景的信息对应的映射关系集合，确定可行轨迹簇中每个候选预测轨迹对应的各个评价维度的代价值，映射关系集合包括至少一个映射关系，每个映射关系包括预设的评价维度与预设的代价值的映射关系，根据可行轨迹簇中每个候选预测轨迹对应的各个评价维度的代价值，确定每个候选预测轨迹的代价值。通过该实现方式中，可以从多个评价维度对可行轨迹簇中的候选预测轨迹进行评价，然后可以便于轨迹规划装置确定出候选预测轨迹的代价值。
在一种可能的实现方式中，与交互场景的信息对应的映射关系集合满足以下至少一项：
在交互场景下障碍物的路权高于车辆的路权、且障碍物的路权越高，侯选预测轨迹对应的路权的代价值越大；
在交互场景下障碍物的通行性高于车辆的通行性、且障碍物的通行性越高，候选预测轨迹对应的通行性的代价值越大；
在交互场景下障碍物的安全性高于车辆的安全性、且障碍物安全性越高，候选预测轨迹对应的安全性的代价值越大；
在交互场景下障碍物的舒适性高于车辆的舒适性、且障碍物的舒适性越高，候选预测轨迹对应的舒适性的代价值越大；
在交互场景下障碍物对车辆的注意力程度越低，候选预测轨迹对应的注意力的代价值越大；
障碍物所属的目标类型的可改变性越低，候选预测轨迹对应的目标类型的代价值越大；
障碍物的目标画像越危险，候选预测轨迹对应的目标画像的代价值越大。
在一种可能的实现方式中，根据第二预测轨迹，生成车辆的时空约束，包括：根据第二预测轨迹在Frenet坐标系上生成障碍物在车辆的期望行驶路径的第一侵占区域，并根据第一侵占区域生成横向偏移约束和曲率半径约束；根据第二预测轨迹在ST坐标系上生成障碍物在车辆的期望行驶路径的第二侵占区域，并根据第二侵占区域生成速度约束、加速度约束和加加速度约束。通过该实现方式，可以实现将物理问题转化为数学问题，从而将根据第二预测轨迹生成车辆的时空约束转变为数学可解的问题。
在一种可能的实现方式中，车辆的轨迹规划装置获取障碍物的第一预测轨迹之后，还可以包括：当障碍物为非博弈目标时，车辆的轨迹规划装置根据第一预测轨迹，生成车辆的时空约束，基于车辆的时空约束，规划车辆的行驶轨迹。通过该实现方式，当障碍物为非博弈目标，障碍物与车辆无交互，根据障碍物的第一预测轨迹生成车辆的时空约束，来轨迹车辆的行驶轨迹，可以保证车辆行驶的安全性。
第二方面，本申请实施例还提供了一种轨迹规划装置，所述电子设备包括执行上述第一方面、以及第一方面的任意一种可能的设计的方法的模块/单元；这些模块/单元可以通过硬件实现，也可以通过硬件执行相应的软件实现。
在一种可能的实现方式中，轨迹规划装置包括获取单元和处理单元；其中：
获取单元，用于获取目标障碍物的第一预测轨迹；
处理单元，用于若障碍物为博弈目标，则根据第一预测轨迹以及障碍物与车辆的交互场景的信息，生成障碍物的可行轨迹簇，从可行轨迹簇中选取一条第二预测轨迹，第二预测轨迹的代价值大于预设阈值，根据第二预测轨迹，生成车辆的时空约束；基于车辆的时空约束，规划车辆的行驶轨迹。
在一种可能的实现方式中，处理单元，具体用于：根据第一预测轨迹、交互场景的信息以及采样限制约束，分别进行横向偏移采样和纵向加速度采样得到采样数据；根据采样数据生成障碍物的可行轨迹簇。
在一种可能的实现方式中，处理单元，还用于：根据交互场景的信息以及交互场景的信息对应的映射关系集合，确定可行轨迹簇中每个候选预测轨迹对应的各个评价维度的代价值，映射关系集合包括至少一个映射关系，每个映射关系包括预设的评价维度与预设的代价值的映射关系；根据可行轨迹簇中每个候选预测轨迹对应的各个评价维度的代价值，确定每个候选预测轨迹的代价值。
在一种可能的实现方式中，与交互场景的信息对应的映射关系集合满足以下至少一项：在交互场景下障碍物的路权高于车辆的路权、且障碍物的路权越高，侯选预测轨迹对应的路权的代价值越大；在交互场景下障碍物的通行性高于车辆的通行性、且障碍物的通行性越高，候选预测轨迹对应的通行性的代价值越大；在交互场景下障碍物的安全性高于车辆的安全性、且障碍物安全性越高，候选预测轨迹对应的安全性的代价值越大；在交互场景下障碍物的舒适性高于车辆的舒适性、且障碍物的舒适性越高，候选预测轨迹对应的舒适性的代价值越大；在交互场景下障碍物对车辆的注意力程度越低，候选预测轨迹对应的注意力的代价值越大；障碍物所属的目标类型的可改变性越低，候选预测轨迹对应的目标类型的代价值越大；障碍物的目标画像越危险，候选预测轨迹对应的目标画像的代价值越大。
在一种可能的实现方式中，处理单元，具体用于：根据第二预测轨迹在SL坐标系上生成障碍物在车辆的期望行驶路径的第一侵占区域，并根据第一侵占区域生成横向偏移约束和曲率半径约束；根据第二预测轨迹在ST坐标系上生成障碍物在车辆的期望行驶路径的第二侵占区域，并根据第二侵占区域生成速度约束、加速度约束和加加速度约束。
在一种可能的实现方式中，处理单元，还用于：当障碍物为非博弈目标时，根据第一预测轨迹，生成车辆的时空约束；基于车辆的时空约束，规划车辆的行驶轨迹。
第三方面，本申请实施例还提供一种轨迹规划装置，轨迹规划装置包括处理器和存储器，所述存储器用于存储计算机执行指令，所述轨迹规划装置运行时，所述处理器执行所述存储器中的计算机执行指令以利用所述轨迹规划装置中的硬件资源执行上述第一方面以及第一方面的任一可能设计的方法的操作步骤。
第四方面，本申请实施例的一种计算机可读存储介质，所述计算机可读存储介质包括计算机指令，当计算机指令在车载设备上运行时，使得所述车载设备执行本申请实施例第一方面、以及第一方面的任一可能设计的技术方案。
第五方面，本申请实施例的中一种程序产品，包括指令，当所述程序产品在车载设备上运行时，使得所述车载设备执行本申请实施例第一方面、以及第一方面的任一可能设计的技术方案。
第六方面，提供了一种芯片系统，该芯片系统可以包括处理器。该处理器与存储器耦合，可用于执行第一方面、以及第一方面的任一种可能实现方式中的方法。可选地，该芯片系统还包括存储器。存储器，用于存储计算机程序(也可以称为代码，或指令)。处理器，用于从存储器调用并运行计算机程序，使得安装有芯片系统的设备执行第一方面、以及第一方面的任一种可能实现方式中的方法。
在具体实现过程中，上述轨迹规划装置可以为芯片，输入电路可以为输入管脚，输出电路可以为输出管脚，处理电路可以为晶体管、门电路、触发器和各种逻辑电路等。输入 电路所接收的输入的信号可以是由例如但不限于接收器接收并输入的，输出电路所输出的信号可以是例如但不限于输出给发射器并由发射器发射的，且输入电路和输出电路可以是同一电路，该电路在不同的时刻分别用作输入电路和输出电路。本申请实施例对处理器及各种电路的具体实现方式不做限定。
附图说明
图1为本申请实施例提供的一种车辆行驶的场景示意图；
图2为本申请实施例提供的一种车辆的轨迹规划方法的流程示意图；
图3至图7为本申请实施例提供的多个场景下障碍物的类型示意图；
图8为本申请实施例提供的横向路径规划的边界约束示意图；
图9为本申请实施例提供的基于ST图的速度规划示意图；
图10为本申请实施例提供的博弈目标的候选预测轨迹示意图；
图11为本申请实施例提供的博弈目标的第一预测轨迹的横纵采样数据示意图；
图12为本申请实施例提供的轨迹规划装置的示意图；
图13为本申请实施例提供的轨迹规划装置的示意图。
具体实施方式
下面将结合附图对本申请作进一步地详细描述。显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。方法实施例中的具体操作方法也可以应用于装置实施例中。其中，在本申请实施例的描述中，本领域普通技术人员可以理解：本申请中涉及的第一、第二等各种数字编号仅为描述方便进行的区分，并不用来限制本申请实施例的范围，也不用来表示先后顺序。“多个”的含义是两个或两个以上。“和/或”，描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。字符“/”一般表示前后关联对象是一种“或”的关系。“至少一个”是指一个或者多个。至少两个是指两个或者多个。“至少一个”、“任意一个”或其类似表达，是指的这些项中的任意组合，包括单项(个)或复数项(个)的任意组合。
本申请实施例中车辆可以基于车辆与外界无线通信技术(例如，vehicle to everything(V2X))与其它物体进行通信。例如，可以基于车辆间无线通信技术(例如，vehicle to vehicle(V2V))实现车辆与其它物体之间的通信。车辆与其它物体之间进行通信可以基于Wi-Fi(wireless fidelity)、第五代(5th generation，5G)移动通信技术、长期演进(long term evolution，LTE)等进行通信。
业界关于自动驾驶汽车提出了分级标准，其中，国际自动机工程师学会(Society of Automotive Engineers International，简称SAE International)提出的驾驶自动化分级标准包括L0-L5等六个级别，其中，L0-L2级，驾驶员支持系统能够为驾驶员提供一些支持功能，但是无论车辆的驾驶员支持功能是否已经开启，驾驶员都必须自己驾驶车辆，并时刻监督驾驶员支持系统提供的这些支持功能，必须根据需要进行转向、制动、或加速以保证安全，L0-L2级的支持功能的区别在于：L0级为无驾驶自动化，支持功能仅限于提供警告和瞬时协助，L1级的支持功能为驾驶员提供转向或制动/加速支持，L2级的支持功能为驾驶员提供转向和制动/加速支持。L3级半自动驾驶，自动驾驶系统既能完成某些驾驶任务，也能 在某些情况下监控驾驶环境，但驾驶员需要随时准备重新取得驾驶控制权，例如当功能请求时，驾驶员必须驾驶。L4级高度自动驾驶，自动驾驶系统在某些环境和特定条件下，能够完成驾驶任务并监控驾驶环境。L5级完全自动驾驶，自动驾驶系统在所有条件下都能完成的所有驾驶任务。本申请实施例的方案可以用于自动驾驶能力达到L2级及以上级别的车辆，后文不再赘述。
在车辆行驶过程中，任何一辆车辆均可以通过感知周围车辆预设时段内的行驶轨迹，来约束本车辆的轨迹规划，从而提高安全性。
在本申请实施例中，将规划模块需要考虑的安全因素转化为优化问题中的约束，即对于规划出来的轨迹，需要满足：
(1)规划的轨迹需要满足可执行准则，即规划的第一车辆的轨迹需要满足自车的动力学约束和运动学约束，其中，动力学约束包括方向盘的最大转角、最大转速，运动学约束包括纵向的最大加速度、最大加加速度。
(2)规划的轨迹还需要满足无碰撞准则，即规划的轨迹不与其他障碍物发生碰撞，然后，根据障碍物是否分配足够的注意力到自车而与自车产生交互、以及障碍物的预测轨迹，生成自车的时空约束。
例如，如图1所示的场景示意图中，该场景中包括第一车辆、以及位于第一车辆右前方的第二车辆，以第一车辆进行轨迹规划为例，第一车辆为具有部分或完全自动驾驶能力的车辆，例如第一车辆可以为上述驾驶自动化分级标准中的L2级及以上级别的车辆，第二车辆可以为上述驾驶自动化分级标准中的L0-L5级中的任一种级别的车辆、半机动车辆、摩托车等。第一车辆可以基于自车的传感器，例如摄像头、激光雷达、毫米波雷达、全球导航卫星系统(Global Navigation Satellite System，GNSS)等获取第二车辆的车辆信息，第二车辆的车辆信息可以包括但不限于第二车辆的地理位置、行驶速度、行驶方向、转向灯信息等信息。第一车辆可以结合第二车辆的车辆信息和当前的道路信息，预测第二车辆的第一预测轨迹。然后，第一车辆基于第二车辆预测轨迹、以及第二车辆是否分配足够的注意力到第一车辆，生成第一车辆的时空约束，例如图1所示的第二车辆在行驶方向上向左侧车道变道时，应当分配足够的注意力到第一车辆以避免与第一车辆发生碰撞，这样第一车辆与第二车辆这两车之间会存在交互过程，第一车辆可以根据交互信息，生成可以保证第二车辆的可行轨迹簇，再从可行轨迹簇中选择一条对于第二车辆而言代价值较大的第二预测轨迹，对于第二车辆而言代价值较大的第二预测轨迹也可以理解为第二预测轨迹对于第二车辆而言是可行轨迹簇中舒适性较低的一条轨迹，然后第一车辆再根据第二预测轨迹生成第一车辆的时空约束，以便第一车辆根据所述时空约束来进行第一车辆的轨迹规划。从而可以在保证第一车辆行驶安全性的同时，提高第一车辆自动驾驶体验的舒适性。
基于上述描述，本申请实施例提供了一种车辆的轨迹规划方法。该方法可以由第一车辆中的轨迹规划装置执行，为便于说明，下面以第一车辆的轨迹规划装置执行该方法为例进行说明。如图2所示，该方法包括：
步骤201，第一车辆的轨迹规划装置获取障碍物的第一预测轨迹。
其中，障碍物可以包括车辆、行人或其它物体，其它物体例如为路障。
在一种可能的实现方式中，以障碍物为第二车辆为例，第一车辆可以基于自车的传感器，例如摄像头、激光雷达、毫米波雷达、GNSS等获取第二车辆的车辆信息，第二车辆的车辆信息可以包括但不限于第二车辆的地理位置、行驶速度、行驶方向、转向灯信息等 信息，第一车辆还可以获取周围的交通指示信息，例如交通灯信息、交通指示路牌信息。第一车辆可以结合第二车辆的车辆信息、当前的道路信息以及交通指示信息等信息，预测第二车辆的第一预测轨迹。
步骤202，轨迹规划装置基于第一预测轨迹以及障碍物的类型，生成第一车辆的时空约束。
其中，障碍物的类型可以包括分配足够的注意力到第一车辆的障碍物或未分配足够的注意力到第一车辆的障碍物两种。
下面以障碍物为第二车辆为例，针对障碍物的类型进行详细介绍。从注意力的角度分析，当第二车辆(它车)与第一车辆(自车)存在交互过程，例如，存在轨迹交叉或行驶意图一致等，第二车辆(无论是否有自动驾驶能力)在接下来的行驶过程中，必须注意(由驾驶员或者自动驾驶系统)到第一车辆的存在以免与第一车辆发生碰撞，并且在这个过程中，第二车辆的行驶轨迹往往还要受第一车辆行驶轨迹的影响，则第二车辆要分配足够的注意力到第一车辆，即该第二车辆的类型为分配足够的注意力到第一车辆的障碍物，该第二车辆属于博弈目标，反之，如果第二车辆与第一车辆之间不存在交互过程，第二车辆不需要注意到第一车辆的存在，而且，第二车辆的行驶轨迹不会受到第一车辆行驶轨迹的影响，即第二车辆无需分配足够的注意力到第一车辆，则第二车辆的类型为未分配足够的注意力到第一车辆的障碍物，该第二车辆属于非博弈目标。
结合图3至图7说明博弈目标和非博弈目标，在图3至图7中车辆1为第一车辆(也就是自车)，如图3所示的路口左转的车辆2、如图4所示的横穿的车辆2、如图5所示的与车辆1汇入相同车道的车辆2、如图6所示的窄道通行时的对向车辆2、如图7所示的车辆1汇入车流的后车车辆2等障碍物与第一车辆存在轨迹冲突或意图冲突，这些障碍物的行为会因第一车辆的驾驶行为发生变化，因此这些障碍物必须分配足够的注意力到第一车辆的存在以免与第一车辆发生碰撞，均属于博弈目标。而如图4所示的在车辆1所在车道前方行驶的车辆3、如图6所示的路边的静止车(即车辆3)、如图7所示的车辆1汇入车流的前车车辆3等障碍物的行为不会因第一车辆行为发生变化，这些障碍物不需要分配足够的注意力到第一车辆，均属于非博弈目标。
基于上述内容，上述步骤202有多种可能的实施方式，下面通过下述可能的实施方式a1和可能的实施方式a2，来进行具体的描述。
可能的实施方式a1，步骤202中，若障碍物的类型为非博弈目标，即不需要分配足够的注意力到第一车辆的障碍物，则轨迹规划装置根据第一预测轨迹，生成第一车辆的时空约束。
在一种可能的实现方式中，轨迹规划装置根据第一预测轨迹在Frenet坐标系上生成标障碍物在第一车辆的期望行驶路径的第一侵占区域，并根据第一侵占区域生成横向偏移约束和曲率半径约束，即时空约束中的空间约束。然后，轨迹规划装置根据第一预测轨迹在ST坐标系上生成障碍物在第一车辆的期望行驶路径的第二侵占区域，并根据第二侵占区域生成速度约束、加速度约束和加加速度约束，即时空约束中的时间约束。
下面对用于横向路径规划的空间约束进行介绍。
以基于Frenet坐标系的横向路径规划为例，如图8所示，Frenet坐标系中，以道路中心线作为参考线，第一车辆沿着该参考线的方向称为纵向，即S坐标，参考线的法线方向为L坐标，第一车辆的位置在参考线上的投影点与第一车辆的位置之间的距离为横向偏移 位移，第一车辆行驶的起始点与该投影点之间的曲线距离为纵向位移。由于第一车辆是不断沿着参考线向前行驶的，所以车辆的横向位移L是随着纵向位移S的变化而不断变化的。
本申请实施例中，可以通过下列边界约束，来进行横向路径规划：
l low≤l(s)≤l up
k low≤k(s)≤k up
其中，l(s)为路径的横向偏移约束，可以根据不同车速下，自车对应的最大方向盘转角和最大方向盘的转速，以及道路的边界，计算得到沿S轴能够行驶到的最大横向偏移。其中，将方向盘的最大转角和最大转速对应到前轮的最大转速后，表1示出了车速与最大前轮转角、最大前轮转速的对应关系示例。
表1
车速km/h 20 30 40 50 100
最大前轮转角rad 0.56 0.15 0.096 0.074 0.022
最大前轮转速rad/s 0.285 0.233 0.116 0.07 0.06
示例的，假设第一车辆在一个车道A上行驶，该车道A的中心线为参考线，以参考线上某一点为原点，从该原点到该参考线左侧第一个车道实线或道路边界或不接跨越的障碍物的距离，可以设置为l low，从该原点到该参考线右侧第一个车道实线或不接跨越的障碍物的距离，可以设置为l up。例如，以车道A的两侧车道线均为实线，车道A的道路宽度为d，那么l low为
Figure PCTCN2021126059-appb-000001
l up为
Figure PCTCN2021126059-appb-000002
k(s)为路径的曲率半径约束，确保路径上每一点的曲率都满足车辆的最小转弯半径的限制。k(s)与车辆的物理参数(例如车辆的转弯极限、车型等)，运动状态，道路地面的摩擦力等相关。
下面对用于纵向路径规划的时间约束进行介绍。
以基于ST图的速度规划示例，横轴T为时间轴，纵轴S为沿第一车辆期望行驶路径(即参考线)的方向，如图9所示，ST图上的平行四边形区域为根据障碍物(它车)预测轨迹生成的在第一车辆(自车)期望行驶路径上的侵占区域，曲线为第一车辆的速度规划曲线。
基于ST图进行速度规划时，通过下列边界约束，保证速度规划的结果是可执行的：
v min≤v(t)≤v max
a min≤a(t)≤a max
jerk min≤jerk(t)≤jerk max
v(t)、a(t)、jerk(t)分别为第一车辆允许的速度、加速度、加加速度的上下界约束，其中，v(t)与当前道路限速(例如高速公路限制最高速度为120km/h)、环境风险限速(例如学校门口限速30km/h，又例如大雾天气能见度在200米以内高速限制最高速度为60km/h)等各种因素相关，第一车辆前进时v(t)为正值，第一车辆倒车时v(t)为负值。在一个示例中，第一车辆在城市道路行驶，v min为0，v max为城市道路的最高限速。在另一个示例中，第一车辆在最高限速120km/h、最低限速60km/h的高速公路上正常行驶时，高速公路v min可以取60km/h，v max可以取120km/h。
a(t)为加速度，车辆加速时a(t)取正值，车辆减速时a(t)取负值，示例性的，a min可以为-4m/s 2，a max可以为3m/s 2。
jerk(t)是加速度对时间的导数，表示加速度的变化率。jerk(t)的取值与当前速度、加速度相关。jerk(t)可以取正数，表示数值越大表示加速度越大，jerk(t)也可以取负数，数值越大，表示加速度越小。示例的，jerk min可以为-10m/s 3，jerk max可以为1.5m/s 3。
通过上述实施例，对于第一车辆的速度限制在v min和v max之间，使得车速既不能速度太低也不能超速，将第一车辆的加速度限制在a min和a max之间，使得加速度也不会超出第一车辆的最大加速度能力，将第一车辆的加加速度限制在jerk min和jerk max之间，使得第一车辆的加速度变化不会很猛烈，从而根据第一车辆的时空约束来规划第一车辆的行驶轨迹，可以实现保证车辆行驶的安全性的同时，提高自动驾驶体验的舒适性。
通过上述时空约束，可以使得第一车辆的规划的行驶轨迹满足上述可执行准则。
可能的实施方式a2，上述步骤202中，若障碍物的类型为博弈目标，则其必须分配足够的注意力到第一车辆，以避免与第一车辆发生碰撞，则轨迹规划装置可以根据障碍物的第一预测轨迹(第一预测轨迹可以理解为障碍物未来行驶概率最大的轨迹)以及障碍物与第一车辆的交互场景的信息，生成对于障碍物而言的未来可能行驶的一系列候选预测轨迹，也可以称为可行轨迹簇，然后从可行轨迹簇中选取一条对于障碍物而言代价值较大的轨迹，作为第二预测轨迹，该第二预测轨迹可以理解为可行轨迹簇中对于障碍物而言舒适性较低的一条轨迹。然后根据第二预测轨迹，生成障碍物的侵占区，基于障碍物的侵占区确定第一车辆的时空约束，基于第一车辆的时空约束，规划第一车辆的行驶轨迹，这样规划出的第一车辆的行驶轨迹使得第一车辆在保持安全行驶的前提下尽可能的具有舒适性。
下面针对如何确定第二预测轨迹进行详细说明。
在一种可能的实现方式中，轨迹规划装置可以根据第一预测轨迹进行横向偏移采样和纵向加速度采样，生成障碍物未来可能行驶的一系列候选预测轨迹，即可行轨迹簇。为了获取到障碍物未来可能行驶的一系列候选预测轨迹，可以设置采样限制约束，采样限制约束可以包括横向偏移采样对应的采样限制约束和纵向加速度采样对应的采样限制约束，横向偏移采样对应的采样限制约束包括但不限于道路边界、车速大小、是否为路口场景等信息，纵向加速度采样对应的采样限制约束包括但不限于加加速度、加速度、速度的限制，以及不同目标类型等信息。
然后，轨迹规划装置可以根据第一预测轨迹、障碍物与车辆的交互场景的信息以及采样限制约束，分别进行横向偏移采样和纵向加速度采样，得到采样数据，根据采样数据生成障碍物未来可能行驶的至少一个候选预测轨迹，再从障碍物未来可能行驶的至少一个候选预测轨迹中确定出一条对于障碍物而言代价值较大的候选预测轨迹，作为障碍物的第二预测轨迹。
如图10所示，根据他车的第一预测轨迹，分别进行横向偏移采样和纵向加速度采样，得到如图11所示的采样数据，图11所述的表格中第一行表示纵向加速度采样所采用的一系列加速度分别为加速度，第一列表示横向偏移采样分别采用横向偏移距离分别为-3～3m。每个格子中的省略号表示以某一加速度和某一横向偏移距离进行采样得到的多个轨迹点的坐标和它车的速度信息，例如，以-4m/s 2的加速度进行纵向加速度采样，3m的横向偏移距离进行横向偏移采样，得到它车的多个轨迹点的坐标和它车在每个轨迹点处的速度信息，存放在第二行第二列的格子中。然后根据采样数据中的横向偏移满足-1至1m的范围，纵 向加速度满足-3至0m/s 2的范围的数据，生成一系列可行轨迹簇，即他车未来可能行驶的至少一个候选预测轨迹，例如图11中规划出3个候选预测轨迹。
在得到障碍物未来可能行驶的至少一个候选预测轨迹之后，从未来可能行驶的至少一个候选预测轨迹中确定出一条对于障碍物而言代价值较大的候选预测轨迹，可以通过以下方式实现：轨迹规划装置可以从多个评价维度对每个候选预测轨迹进行评价，得到每个候选预测轨迹的代价值，再从未来可能行驶的至少一个候选预测轨迹中，确定出障碍物的代价值大于代价阈值的候选预测轨迹，作为第二预测轨迹，以便第一车辆根据第二预测轨迹生成第一车辆的时空约束。
下面介绍如何确定每个候选预测轨迹的代价值。
本申请实施例中，轨迹规划装置还可以存储有映射关系集合，或者轨迹规划装置从第一车辆的其它存储设备中获取到映射关系集合，映射关系集合包括至少一个映射关系，每个映射关系包括预设的评价维度与预设的代价值的映射关系，举个例子，假设有七个预设的评价维度，分别为路权、通行性、安全性、舒适性、注意力、目标类型、目标画像，那么映射关系集合包括七个映射关系。关于各个评价维度在后文进行介绍，此处不再赘述。
本申请实施例涉及到的映射关系集合可以包括以下至少一项映射关系：
路权与预设的代价值的映射关系；
通行性与预设的代价值的映射关系；
安全性与预设的代价值的映射关系；
舒适性与预设的代价值的映射关系；
注意力与预设的代价值的映射关系；
目标类型与预设的代价值的映射关系；
目标画像与预设的代价值的映射关系。
本申请实施例中，每个交互场景的信息可以对应一个映射关系集合，轨迹规划装置可以根据障碍物与第一车辆的交互场景的信息以及交互场景的信息对应的映射关系集合，确定至少一个候选预测轨迹中每个候选预测轨迹对应的各个评价维度的代价值。然后，针对每个候选预测轨迹来说，可以根据这个候选预测轨迹对应的各个评价维度的代价值，确定这个候选预测轨迹的代价值。
以第一车辆与障碍物的交互场景的信息对应的映射关系集合包括上述七个映射关系为例，轨迹规划装置基于交互场景的信息，从路权、通行性、安全性、舒适性、注意力、目标类型、目标画像等维度对每条候选预测轨迹进行评价，不同维度设置有不同的权重，对于每个候选预测轨迹来说，将各个评价维度对应代价值进行加权求和，得到每个候选预测轨迹的代价值。
其中，每条候选预测轨迹的代价值可以通过如下公式计算得到：
trajectory cost＝∑w i*x i……公式(1)
上述公式(1)中，x i为不同评价维度的代价值，取值0～1之间，w i为对应维度的权重值，某一个评价维度的权重值是指该评价维度在整体评价中的相对重要程度，可以取任意数值，例如w i的取值为0～1之间的值，此处不限定w i的具体取值，例如舒适性的权重为0.3～0.4。trajectory cost为每条候选预测轨迹的最终代价值。
然后，轨迹规划装置从至少一个候选预测轨迹中代价值大于代价阈值的候选预测轨迹 中，确定出第二预测轨迹。例如，从代价值大于代价阈值的候选预测轨迹中，确定出代价值最大的一个候选预测轨迹作为第二预测轨迹，代价值最大的一个候选预测轨迹可以理解为至少一个候选预测轨迹中对于障碍物而言舒适性最低的一条候选预测轨迹。又例如，从代价值大于代价阈值的候选预测轨迹中，确定出代价值较大的N个候选预测轨迹，N为大于1的整数，然后，从N个候选预测轨迹中选择一个候选预测轨迹作为第二预测轨迹。
下面针对不同评价维度进行说明。
(1)路权：是指物体在一定空间和时间内行驶的权利，不同场景的路权大小示例如下：
路口场景规则：斑马线>直行>左转>右转>掉头(U turn)；
非路口场景规则：斑马线>车道直行>换道>逆向直行>横穿>豁口。
例如在路口场景中，若障碍物在斑马线上，第一车辆若为直行，则障碍物的路权较高；若第一车辆为直行，障碍物为左转，则第一车辆的路权较高。
在障碍物与第一车辆的交互场景下，当障碍物的路权高于第一车辆的路权、且障碍物的路权越高时，障碍物越倾向保持当前状态，即采样之后的候选预测轨迹与其初始的第一预测轨迹的偏差越大，侯选预测轨迹对应的路权的代价值越大。
(2)通行性：是指第一车辆需要保持一定的速度通过一定的空间，以通过路口为例，一般越快通过路口，通行性越高，但是如果在中途刹车，就会很慢通过该路口，通行性就会较低。不同场景的通行性大小示例如下：
路口场景规则：直行>左转>右转>掉头(U turn)>斑马线；
非路口场景规则：车道直行>换道>逆向直行>横穿>豁口。
例如在路口场景，若障碍物在斑马线上，第一车辆若为直行，则第一车辆的通行性较高。障碍物的通行性越高，越倾向保持当前状态。在交互场景下障碍物的通行性高于第一车辆的通行性、且障碍物的通行性越高，候选预测轨迹对应的通行性的代价值越大。
(3)安全性：第一车辆与障碍物交互过程的安全，例如，第一车辆与障碍物之间交互过程中始终保持大于一定阈值的距离，这种情况下安全性高，又例如，第一车辆与障碍物之间交互过程中发送碰撞，这种情况下安全性最低。障碍物的安全性越高，越倾向保持当前状态。不同场景的安全性大小示例如下：
路口场景规则：斑马线>直行>左转>右转>掉头(U turn)；
非路口场景规则：横穿>车道直行>换道>逆向直行>豁口。
在交互场景下障碍物的安全性高于第一车辆的安全性、且障碍物安全性越高，候选预测轨迹对应的安全性的代价值越大。
(4)舒适性：可以用jerk值表示，jerk越大，代表舒适性越差。障碍物的舒适性越高，越倾向保持当前状态。舒适性在不同场景下，可进一步分为横向和纵向两方面，不同场景的舒适性大小示例如下：
路口场景规则：
横向：掉头(U turn)>左转≈右转>斑马线>直行；
纵向：斑马线>直行>左转≈右转>掉头(U turn)。
非路口场景规则：
横向：换道>豁口>逆向直行>车道直行≈横穿；
纵向：车道直行>换道>逆向直行>横穿>豁口。
在交互场景下障碍物的舒适性高于第一车辆的舒适性、且障碍物的舒适性越高，候选预测轨迹对应的舒适性的代价值越大。
(5)注意力：根据障碍物对第一车辆的注意力程度对轨迹进行评价，障碍物对第一车辆的注意力程度越低，越倾向保持当前状态。在交互场景下障碍物对第一车辆的注意力程度越低，候选预测轨迹对应的注意力的代价值越大。
(6)障碍物类型：根据障碍物类型的可改变性对轨迹进行评价，障碍物类型的可改变性越低，障碍物越倾向保持当前状态，不同目标类型的代价值的大小比较规则示例如下：
小孩>行人>非机动车≈机动单车>大型车>小型车。
其中，可改变性可以理解为有能力对当前的运动状态进行改变，例如，小孩没有能力去对当前运动状态进行改变，小孩的可改变性很低，其对应的代价值大，而小型车的可改变性较高，其对应的代价值小。障碍物所属的目标类型的可改变性越低，候选预测轨迹对应的目标类型的代价值越大。
(7)障碍物画像：根据障碍物的画像对轨迹进行评价，障碍物画像越危险，障碍物越倾向保持当前状态，规则示例如下：
危险>激进>正常>保守。
举例来说，障碍物画像可以根据在一个生命周期内障碍物的状态管理确定，例如，障碍物为智能车，可以根据智能车的行为确定其画像，例如，智能车需要左转，如果智能车是在直行道强行左转加塞到左转车道，其画像可以认为是危险。在其它一些示例中，画像也可以根据该智能车的行为、智能车的类型、驾驶员的类型、以及驾驶员当前正在做什么等来确定，其中，智能车的类型例如可以为跑车、家用车等。障碍物的目标画像越危险，候选预测轨迹对应的目标画像的代价值越大。
上述实施例中，在确定出障碍物的第二预测轨迹之后，轨迹规划装置根据第二预测轨迹，生成第一车辆的时空约束，以实现将物理问题转化为数学问题。一种可能的实现方式中，根据第二预测轨迹在Frenet坐标系上生成障碍物在第一车辆的期望行驶路径的第一侵占区域，并根据第一侵占区域生成横向偏移约束和曲率半径约束，即时空约束中的空间约束。然后，轨迹规划装置根据第二预测轨迹在ST坐标系上生成障碍物在第一车辆的期望行驶路径的第二侵占区域，并根据第二侵占区域生成速度约束、加速度约束和加加速度约束，即时空约束中的时间约束。此处，根据第二预测轨迹生成第一车辆的时空约束的具体实现方式，可以参见上述可能的实施方式a1中关于根据第一预测轨迹生成第一车辆的时空约束的相关内容，此处不再赘述。
步骤203，根据第一车辆的时空约束，确定第一车辆的轨迹规划。
当障碍物为非博弈目标，轨迹规划装置可以在根据第一预测轨迹生成第一车辆的时空约束之后，基于第一预测轨迹生成的第一车辆的时空约束来规划第一车辆的行驶轨迹。
当障碍物为博弈目标，轨迹规划装置可以在根据第二预测轨迹，生成第一车辆的时空约束之后，基于第二预测轨迹生成的第一车辆的时空约束来规划第一车辆的行驶轨迹。
通过上述内容，根据障碍物是否分配足够的注意力到自车(即第一车辆)，是否与自车产生交互，而将障碍物分为非博弈目标和博弈目标，针对非博弈目标，障碍物与自车无 交互，根据障碍物的第一预测轨迹，生成自车的时空约束；而针对博弈目标，障碍物与自车有交互，根据交互场景生成障碍物的可行轨迹簇(即上述至少一个候选预测轨迹)，然后基于从可行轨迹簇中选取的对于障碍物而言代价值较大的可行轨迹(即上述第二预测轨迹)，生成自车的时空约束，在此时空约束下，可以使得自车的规划轨迹在保持安全行驶的前提下尽可能的具有舒适性。
基于以上实施例以及相同构思，图12为本申请实施例提供的轨迹规划装置的示意图，如图12所示，该轨迹规划装置1200可以实现上述方法实施例中的轨迹规划装置所执行的步骤。该轨迹规划装置可以包括获取单元1201和处理单元1202。
获取单元1201，用于获取障碍物的第一预测轨迹；
处理单元1202，用于若障碍物为博弈目标，则根据第一预测轨迹以及障碍物与车辆的交互场景的信息，生成障碍物的可行轨迹簇；从可行轨迹簇中选取一条第二预测轨迹，第二预测轨迹的代价值大于预设阈值；根据第二预测轨迹，生成车辆的时空约束，车辆的时空约束用于车辆规划自车的行驶轨迹。
一种可能的实现方式中，处理单元1202，具体用于：根据第一预测轨迹、交互场景的信息以及采样限制约束，分别进行横向偏移采样和纵向加速度采样得到采样数据；根据采样数据生成障碍物的可行轨迹簇。
一种可能的实现方式中，处理单元1202，还用于：根据交互场景的信息以及交互场景的信息对应的映射关系集合，确定可行轨迹簇中每个候选预测轨迹对应的各个评价维度的代价值，映射关系集合包括至少一个映射关系，每个映射关系包括预设的评价维度与预设的代价值的映射关系；根据可行轨迹簇中每个候选预测轨迹对应的各个评价维度的代价值，确定每个候选预测轨迹的代价值。
一种可能的实现方式中，与交互场景的信息对应的映射关系集合满足以下至少一项：
在交互场景下障碍物的路权高于车辆的路权、且障碍物的路权越高，侯选预测轨迹对应的路权的代价值越大；
在交互场景下障碍物的通行性高于车辆的通行性、且障碍物的通行性越高，候选预测轨迹对应的通行性的代价值越大；
在交互场景下障碍物的安全性高于车辆的安全性、且障碍物安全性越高，候选预测轨迹对应的安全性的代价值越大；
在交互场景下障碍物的舒适性高于车辆的舒适性、且障碍物的舒适性越高，候选预测轨迹对应的舒适性的代价值越大；
在交互场景下障碍物对车辆的注意力程度越低，候选预测轨迹对应的注意力的代价值越大；
障碍物所属的目标类型的可改变性越低，候选预测轨迹对应的目标类型的代价值越大；
障碍物的目标画像越危险，候选预测轨迹对应的目标画像的代价值越大。
一种可能的实现方式中，处理单元1202，具体用于：根据第二预测轨迹在SL坐标系上生成障碍物在车辆的期望行驶路径的第一侵占区域，并根据第一侵占区域生成横向偏移约束和曲率半径约束；根据第二预测轨迹在ST坐标系上生成障碍物在车辆的期望行驶路径的第二侵占区域，并根据第二侵占区域生成速度约束、加速度约束和加加速度约束。
一种可能的实现方式中，处理单元1202，还用于：当障碍物为非博弈目标时，根据第 一预测轨迹，生成车辆的时空约束，车辆的时空约束用于车辆规划自车的行驶轨迹。
该轨迹规划装置所涉及的与本申请实施例提供的技术方案相关的概念，解释和详细说明及其他步骤请参见前述方法或其他实施例中关于这些内容的描述，此处不做赘述。
根据前述方法，图13为本申请实施例提供的轨迹规划装置的结构示意图，如图13所示，该轨迹规划装置1300可以包括存储器1301、处理器1302，还可以包括总线系统，处理器1302和存储器1301可以通过总线系统相连。
应理解，上述处理器1302可以是一个芯片。例如，该处理器1302可以是现场可编程门阵列(field programmable gate array，FPGA)，可以是专用集成芯片(application specific integrated circuit，ASIC)，还可以是系统芯片(system on chip，SoC)，还可以是中央处理器(central processor unit，CPU)，还可以是网络处理器(network processor，NP)，还可以是数字信号处理电路(digital signal processor，DSP)，还可以是微控制器(micro controller unit，MCU)，还可以是可编程控制器(programmable logic device，PLD)或其他集成芯片。
在实现过程中，上述方法的各步骤可以通过处理器1302中的硬件的集成逻辑电路或者软件形式的指令完成。结合本申请实施例所公开的方法的步骤可以直接体现为硬件处理器执行完成，或者用处理器1302中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器1301，处理器1302读取存储器1301中的信息，结合其硬件完成上述方法的步骤。
应注意，本申请实施例中的处理器1302可以是一种集成电路芯片，具有信号的处理能力。在实现过程中，上述方法实施例的各步骤可以通过处理器中的硬件的集成逻辑电路或者软件形式的指令完成。上述的处理器可以是通用处理器、数字信号处理器(DSP)、专用集成电路(ASIC)、现场可编程门阵列(FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成，或者用译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器，处理器读取存储器中的信息，结合其硬件完成上述方法的步骤。
可以理解，本申请实施例中的存储器1301可以是易失性存储器或非易失性存储器，或可包括易失性和非易失性存储器两者。其中，非易失性存储器可以是只读存储器(read-only memory，ROM)、可编程只读存储器(programmable ROM，PROM)、可擦除可编程只读存储器(erasable PROM，EPROM)、电可擦除可编程只读存储器(electrically EPROM，EEPROM)或闪存。易失性存储器可以是随机存取存储器(random access memory，RAM)，其用作外部高速缓存。通过示例性但不是限制性说明，许多形式的RAM可用，例如静态随机存取存储器(static RAM，SRAM)、动态随机存取存储器(dynamic RAM，DRAM)、同步动态随机存取存储器(synchronous DRAM，SDRAM)、双倍数据速率同步动态随机存取存储器(double data rate SDRAM，DDR SDRAM)、增强型同步动态随机 存取存储器(enhanced SDRAM，ESDRAM)、同步连接动态随机存取存储器(synchlink DRAM，SLDRAM)和直接内存总线随机存取存储器(direct rambus RAM，DR RAM)。应注意，本文描述的系统和方法的存储器旨在包括但不限于这些和任意其它适合类型的存储器。
本申请还提供一种车辆，所述车辆可以包含上述涉及的轨迹规划装置。在一种示例中，所述车辆可以为本申请涉及的第一车辆。
根据本申请实施例提供的方法，本申请还提供一种计算机程序产品，该计算机程序产品包括：计算机程序代码或指令，当该计算机程序代码或指令在计算机上运行时，使得该计算机执行上述方法实施例中任意一个实施例的方法。
根据本申请实施例提供的方法，本申请还提供一种计算机可读存储介质，该计算机可读介质存储有程序代码，当该程序代码在计算机上运行时，使得该计算机执行上述方法实施例中任意一个实施例的方法。
根据本申请实施例提供的方法，本申请还提供一种芯片系统，该芯片系统可以包括处理器。该处理器与存储器耦合，可用于执行上述方法实施例中任意一个实施例的方法。可选地，该芯片系统还包括存储器。存储器，用于存储计算机程序(也可以称为代码，或指令)。处理器，用于从存储器调用并运行计算机程序，使得安装有芯片系统的设备执行上述方法实施例中任意一个实施例的方法。
上述实施例，可以全部或部分地通过软件、硬件、固件或其他任意组合来实现。当使用软件实现时，上述实施例可以全部或部分地以计算机程序产品的形式实现。所述计算机程序产品包括一个或多个计算机指令。在计算机上加载或执行所述计算机程序指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以为通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集合的服务器、数据中心等数据存储设备。所述可用介质可以是磁性介质(例如，软盘、硬盘、磁带)、光介质(例如，DVD)、或者半导体介质。半导体介质可以是固态硬盘(solid state drive，SSD)。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统、装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本申请的技术方案的目 的。
以上所述，仅为本申请的具体实施方式。熟悉本技术领域的技术人员根据本申请提供的具体实施方式，可想到变化或替换，都应涵盖在本申请的保护范围之内。

Claims (16)

  1. 一种车辆的轨迹规划方法，其特征在于，所述方法包括：
    获取障碍物的第一预测轨迹；
    当所述障碍物为博弈目标时，根据所述第一预测轨迹以及所述障碍物与车辆的交互场景的信息，生成所述障碍物的可行轨迹簇；
    从所述可行轨迹簇中选取一条第二预测轨迹，所述第二预测轨迹的代价值大于预设阈值；
    根据所述第二预测轨迹，生成车辆的时空约束；
    基于所述车辆的时空约束，规划所述车辆的行驶轨迹。
  2. 如权利要求1所述的方法，其特征在于，所述根据所述第一预测轨迹以及所述障碍物与所述车辆的交互场景的信息，生成所述障碍物的可行轨迹簇，包括：
    根据所述第一预测轨迹、所述交互场景的信息以及采样限制约束，分别进行横向偏移采样和纵向加速度采样得到采样数据；
    根据所述采样数据生成所述障碍物的可行轨迹簇。
  3. 如权利要求2所述的方法，其特征在于，所述从所述可行轨迹簇中选取一条第二预测轨迹之前，还包括：
    根据所述交互场景的信息以及所述交互场景的信息对应的映射关系集合，确定所述可行轨迹簇中每个候选预测轨迹对应的各个评价维度的代价值，所述映射关系集合包括至少一个映射关系，每个映射关系包括预设的评价维度与预设的代价值的映射关系；
    根据所述可行轨迹簇中每个候选预测轨迹对应的各个评价维度的代价值，确定每个候选预测轨迹的代价值。
  4. 如权利要求3所述的方法，其特征在于，所述与所述交互场景的信息对应的映射关系集合满足以下至少一项：
    在所述交互场景下所述障碍物的路权高于所述车辆的路权、且所述障碍物的路权越高，所述侯选预测轨迹对应的路权的代价值越大；
    在所述交互场景下所述障碍物的通行性高于所述车辆的通行性、且所述障碍物的通行性越高，所述候选预测轨迹对应的通行性的代价值越大；
    在所述交互场景下所述障碍物的安全性高于所述车辆的安全性、且所述障碍物安全性越高，所述候选预测轨迹对应的安全性的代价值越大；
    在所述交互场景下所述障碍物的舒适性高于所述车辆的舒适性、且所述障碍物的舒适性越高，所述候选预测轨迹对应的舒适性的代价值越大；
    在所述交互场景下所述障碍物对所述车辆的注意力程度越低，所述候选预测轨迹对应的注意力的代价值越大；
    所述障碍物所属的目标类型的可改变性越低，所述候选预测轨迹对应的目标类型的代价值越大；
    所述障碍物的目标画像越危险，所述候选预测轨迹对应的目标画像的代价值越大。
  5. 如权利要求2-4任一项所述的方法，其特征在于，所述根据所述第二预测轨迹，生成所述车辆的时空约束，包括：
    根据所述第二预测轨迹在Frenet坐标系上生成所述障碍物在所述车辆的期望行驶路径 的第一侵占区域，并根据所述第一侵占区域生成横向偏移约束和曲率半径约束；
    根据所述第二预测轨迹在ST坐标系上生成所述障碍物在所述车辆的期望行驶路径的第二侵占区域，并根据所述第二侵占区域生成速度约束、加速度约束和加加速度约束。
  6. 如权利要求1所述的方法，其特征在于，所述获取障碍物的第一预测轨迹之后，还包括：
    当所述障碍物为非博弈目标时，根据所述第一预测轨迹，生成所述车辆的时空约束，基于所述车辆的时空约束，规划所述车辆的行驶轨迹。
  7. 一种轨迹规划装置，其特征在于，包括：
    获取单元，用于获取目标障碍物的第一预测轨迹；
    处理单元，用于若所述障碍物为博弈目标，则根据所述第一预测轨迹以及所述障碍物与车辆的交互场景的信息，生成所述障碍物的可行轨迹簇；从所述可行轨迹簇中选取一条第二预测轨迹，所述第二预测轨迹的代价值大于预设阈值；根据所述第二预测轨迹，生成车辆的时空约束；基于所述车辆的时空约束，规划所述车辆的行驶轨迹。
  8. 如权利要求7所述的装置，其特征在于，处理单元，具体用于：
    根据所述第一预测轨迹、所述交互场景的信息以及采样限制约束，分别进行横向偏移采样和纵向加速度采样得到采样数据；
    根据所述采样数据生成所述障碍物的可行轨迹簇。
  9. 如权利要求8所述的装置，其特征在于，所述处理单元，还用于：
    根据所述交互场景的信息以及所述交互场景的信息对应的映射关系集合，确定所述可行轨迹簇中每个候选预测轨迹对应的各个评价维度的代价值，所述映射关系集合包括至少一个映射关系，每个映射关系包括预设的评价维度与预设的代价值的映射关系；
    根据所述可行轨迹簇中每个候选预测轨迹对应的各个评价维度的代价值，确定每个候选预测轨迹的代价值。
  10. 如权利要求9所述的装置，其特征在于，所述与所述交互场景的信息对应的映射关系集合满足以下至少一项：
    在所述交互场景下所述障碍物的路权高于所述车辆的路权、且所述障碍物的路权越高，所述侯选预测轨迹对应的路权的代价值越大；
    在所述交互场景下所述障碍物的通行性高于所述车辆的通行性、且所述障碍物的通行性越高，所述候选预测轨迹对应的通行性的代价值越大；
    在所述交互场景下所述障碍物的安全性高于所述车辆的安全性、且所述障碍物安全性越高，所述候选预测轨迹对应的安全性的代价值越大；
    在所述交互场景下所述障碍物的舒适性高于所述车辆的舒适性、且所述障碍物的舒适性越高，所述候选预测轨迹对应的舒适性的代价值越大；
    在所述交互场景下所述障碍物对所述车辆的注意力程度越低，所述候选预测轨迹对应的注意力的代价值越大；
    所述障碍物所属的目标类型的可改变性越低，所述候选预测轨迹对应的目标类型的代价值越大；
    所述障碍物的目标画像越危险，所述候选预测轨迹对应的目标画像的代价值越大。
  11. 如权利要求7-10任一项所述的装置，其特征在于，所述处理单元，具体用于：
    根据所述第二预测轨迹在SL坐标系上生成所述障碍物在所述车辆的期望行驶路径的 第一侵占区域，并根据所述第一侵占区域生成横向偏移约束和曲率半径约束；
    根据所述第二预测轨迹在ST坐标系上生成所述障碍物在所述车辆的期望行驶路径的第二侵占区域，并根据所述第二侵占区域生成速度约束、加速度约束和加加速度约束。
  12. 如权利要求7-11任一项所述的装置，其特征在于，所述处理单元，还用于：
    当所述障碍物为非博弈目标时，根据所述第一预测轨迹，生成所述车辆的时空约束；
    基于所述车辆的时空约束，规划所述车辆的行驶轨迹。
  13. 一种轨迹规划装置，其特征在于，包括：存储器与处理器，所述存储器用于存储指令，所述处理器用于执行所述存储器存储的指令，并且执行所述存储器中存储的指令，以使得所述处理器执行如权利要求1至6中任一项所述的方法。
  14. 一种车辆，其特征在于，所述车辆包括如权利要求7至12中任一项所述的轨迹规划装置。
  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中存储有计算机可读指令，所述计算机读取并执行所述计算机可读指令时，使得所述计算机执行如权利要求1至6中任一项所述的方法。
  16. 一种计算机程序产品，其特征在于，包括计算机可读指令，当计算机读取并执行所述计算机可读指令，使得所述计算机执行如权利要求1至6中任一项所述的方法。
PCT/CN2021/126059 2021-10-25 2021-10-25 一种车辆的轨迹规划方法、装置及车辆 WO2023070258A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202180057306.9A CN116457853A (zh) 2021-10-25 2021-10-25 一种车辆的轨迹规划方法、装置及车辆
PCT/CN2021/126059 WO2023070258A1 (zh) 2021-10-25 2021-10-25 一种车辆的轨迹规划方法、装置及车辆

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/126059 WO2023070258A1 (zh) 2021-10-25 2021-10-25 一种车辆的轨迹规划方法、装置及车辆

Publications (1)

Publication Number Publication Date
WO2023070258A1 true WO2023070258A1 (zh) 2023-05-04

Family

ID=86159905

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/126059 WO2023070258A1 (zh) 2021-10-25 2021-10-25 一种车辆的轨迹规划方法、装置及车辆

Country Status (2)

Country Link
CN (1) CN116457853A (zh)
WO (1) WO2023070258A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116540745A (zh) * 2023-07-05 2023-08-04 新石器慧通（北京）科技有限公司 轨迹规划方法和装置、无人车和存储介质
CN116572994A (zh) * 2023-07-10 2023-08-11 之江实验室 一种车辆速度规划方法、装置及计算机可读介质
CN116734882A (zh) * 2023-08-14 2023-09-12 禾昆科技(北京)有限公司 车辆路径规划方法、装置、电子设备和计算机可读介质
CN117141474A (zh) * 2023-10-30 2023-12-01 深圳海星智驾科技有限公司 障碍物轨迹预测方法、装置、车辆控制器、系统及车辆
CN117636270A (zh) * 2024-01-23 2024-03-01 南京理工大学 基于单目摄像头的车辆抢道事件识别方法及设备

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106527452A (zh) * 2016-12-30 2017-03-22 广州汽车集团股份有限公司 一种无人驾驶汽车遇障时运动路径规划方法及系统
CN107168305A (zh) * 2017-04-01 2017-09-15 西安交通大学 路口场景下基于Bezier和VFH的无人车轨迹规划方法
CN109204311A (zh) * 2017-07-04 2019-01-15 华为技术有限公司 一种汽车速度控制方法和装置
CN109213134A (zh) * 2017-07-03 2019-01-15 百度在线网络技术（北京）有限公司 生成自动驾驶策略的方法和装置
CN110362096A (zh) * 2019-08-13 2019-10-22 东北大学 一种基于局部最优性的无人驾驶车辆动态轨迹规划方法
CN110674723A (zh) * 2019-09-19 2020-01-10 北京三快在线科技有限公司 一种确定无人驾驶车辆行驶轨迹的方法及装置
CN111965968A (zh) * 2019-05-20 2020-11-20 华为技术有限公司 一种切换控制方法、系统及装置
US20210114625A1 (en) * 2019-10-18 2021-04-22 WeRide Corp. System and method for autonomous collision avoidance

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106527452A (zh) * 2016-12-30 2017-03-22 广州汽车集团股份有限公司 一种无人驾驶汽车遇障时运动路径规划方法及系统
CN107168305A (zh) * 2017-04-01 2017-09-15 西安交通大学 路口场景下基于Bezier和VFH的无人车轨迹规划方法
CN109213134A (zh) * 2017-07-03 2019-01-15 百度在线网络技术（北京）有限公司 生成自动驾驶策略的方法和装置
CN109204311A (zh) * 2017-07-04 2019-01-15 华为技术有限公司 一种汽车速度控制方法和装置
CN111965968A (zh) * 2019-05-20 2020-11-20 华为技术有限公司 一种切换控制方法、系统及装置
CN110362096A (zh) * 2019-08-13 2019-10-22 东北大学 一种基于局部最优性的无人驾驶车辆动态轨迹规划方法
CN110674723A (zh) * 2019-09-19 2020-01-10 北京三快在线科技有限公司 一种确定无人驾驶车辆行驶轨迹的方法及装置
US20210114625A1 (en) * 2019-10-18 2021-04-22 WeRide Corp. System and method for autonomous collision avoidance

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116540745A (zh) * 2023-07-05 2023-08-04 新石器慧通（北京）科技有限公司 轨迹规划方法和装置、无人车和存储介质
CN116540745B (zh) * 2023-07-05 2023-09-12 新石器慧通（北京）科技有限公司 轨迹规划方法和装置、无人车和存储介质
CN116572994A (zh) * 2023-07-10 2023-08-11 之江实验室 一种车辆速度规划方法、装置及计算机可读介质
CN116572994B (zh) * 2023-07-10 2023-09-22 之江实验室 一种车辆速度规划方法、装置及计算机可读介质
CN116734882A (zh) * 2023-08-14 2023-09-12 禾昆科技(北京)有限公司 车辆路径规划方法、装置、电子设备和计算机可读介质
CN116734882B (zh) * 2023-08-14 2023-11-24 禾昆科技(北京)有限公司 车辆路径规划方法、装置、电子设备和计算机可读介质
CN117141474A (zh) * 2023-10-30 2023-12-01 深圳海星智驾科技有限公司 障碍物轨迹预测方法、装置、车辆控制器、系统及车辆
CN117141474B (zh) * 2023-10-30 2024-01-30 深圳海星智驾科技有限公司 障碍物轨迹预测方法、装置、车辆控制器、系统及车辆
CN117636270A (zh) * 2024-01-23 2024-03-01 南京理工大学 基于单目摄像头的车辆抢道事件识别方法及设备
CN117636270B (zh) * 2024-01-23 2024-04-09 南京理工大学 基于单目摄像头的车辆抢道事件识别方法及设备

Also Published As

Publication number Publication date
CN116457853A (zh) 2023-07-18

Similar Documents

Publication Publication Date Title
WO2023070258A1 (zh) 一种车辆的轨迹规划方法、装置及车辆
US11858508B2 (en) Trajectory prediction from precomputed or dynamically generated bank of trajectories
CN110488802B (zh) 一种网联环境下的自动驾驶车辆动态行为决策方法
US10235881B2 (en) Autonomous operation capability configuration for a vehicle
CN114222691A (zh) 基于成本的路径确定
CN114600053A (zh) 基于碰撞区域的行驶路线修改
JP2023510136A (ja) 知覚、予測又は計画のための地理的位置特定モデル
US11945440B2 (en) Data driven rule books
US11814075B2 (en) Conditional motion predictions
US20220234618A1 (en) Homotopic-based planner for autonomous vehicles
US20220314968A1 (en) Electronic control device
JP2023531962A (ja) 移動制限及び観察された走行挙動から生成されたデルタコストボリュームを使用した経路計画
US11887324B2 (en) Cross-modality active learning for object detection
US20220219727A1 (en) Vehicle path planning
CN114834451A (zh) 自主车辆的主动式车道变换
WO2023213200A1 (zh) 一种会车方法及相关装置
US20230256999A1 (en) Simulation of imminent crash to minimize damage involving an autonomous vehicle
US20220357453A1 (en) Lidar point cloud segmentation using box prediction
EP4270352A1 (en) Controlling a future traffic state on a road segment
US20230003533A1 (en) Fast collision free path generation by connecting c-slices through cell decomposition
EP4202476A1 (en) Anomaly prioritization using dual-mode adaptive radar
US20230192144A1 (en) Uncertainty prediction for a predicted path of an object that avoids infeasible paths
US20230192133A1 (en) Conditional mode anchoring
US20230243952A1 (en) Unified radar perception architecture
US20230053243A1 (en) Hybrid Performance Critic for Planning Module's Parameter Tuning in Autonomous Driving Vehicles

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 202180057306.9

Country of ref document: CN

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

Ref document number: 21961649

Country of ref document: EP

Kind code of ref document: A1