Benchmark test system for automatic parking assist function

Abstract

The invention relates to a benchmark test system for an automatic parking assistance function, and the system comprises a scene creation module which is used for creating a parking lot scene which at least comprises the current position of a vehicle and a target parking position; the path planning module is used for planning a vehicle track from the current position of the vehicle to the target parking position in the parking lot scene by utilizing one or more path planning algorithms of the automatic parking assistance function; and the track evaluation module is used for evaluating the vehicle track based on a plurality of indexes. The invention further relates to a control method of the automatic parking auxiliary function, a computer readable storage medium and a computer program product.

Description

Technical Field

[0001] The present application relates to the field of automatic parking assistance function testing and control, and more particularly, to a benchmark testing system for an automatic parking assistance function, a control method for an automatic parking assistance function, a computer-readable storage medium, and a computer program product. Background Art

[0002] Currently, the APA (automatic parking assist) function or system mainly uses ultrasonic sensors and / or close-range cameras as environmental perception sensors, and adopts a path planning algorithm to plan the trajectory from the vehicle's stopping position to the target parking final position.

[0003] However, existing path planning algorithms all have different limitations. For example, the geometric algorithm has limited path selection (straight lines, circles, clothoid curves, etc.), which means that some paths that human drivers can drive cannot be planned by this algorithm. Therefore, in certain scenarios, the geometry-based algorithm may not be able to find an effective trajectory. Moreover, even if a valid trajectory is found based on the path planning algorithm, it is still impossible to determine whether the trajectory is the best trajectory. Summary of the invention

[0004] According to one aspect of the present application, a benchmarking system for an automatic parking assistance function is provided, the system comprising: a scenario creation module, for creating a parking lot scenario, the parking lot scenario at least comprising a current position of a vehicle and a target parking position; a path planning module, for planning a vehicle trajectory from the current position of the vehicle to the target parking position in the parking lot scenario using one or more path planning algorithms of the automatic parking assistance function; and a trajectory evaluation module, for evaluating the vehicle trajectory based on multiple indicators.

[0005] As a supplement or alternative to the above solution, in the above benchmark test system, the parking lot scene also includes configurable objects or parameters, including vehicle parameters, obstacles around the vehicle, and ground signs.

[0006] As a supplement or replacement for the above solution, in the above benchmark system, the path planning algorithm includes: a geometry-based algorithm, a graph search-based algorithm, and a reinforcement learning-based algorithm.

[0007] As a supplement or alternative to the above solution, in the above benchmark system, the path planning module is configured to: receive a user's selection of the path planning algorithm; and call the selected path planning algorithm to plan the vehicle trajectory in the parking lot scenario.

[0008] As a supplement or alternative to the above solution, in the above benchmark test system, the path planning module is further configured to: receive an ideal vehicle trajectory determined based on expert knowledge.

[0009] As a supplement or alternative to the above scheme, in the above benchmarking system, the trajectory evaluation module is configured to: receive multiple indicators and their priorities input or selected by a user, the indicators including trajectory length, trajectory smoothness, number of gear shifts and occupied space; and evaluate the vehicle trajectory based on the multiple indicators to determine the most suitable path planning algorithm or vehicle trajectory in the parking lot scenario, and the path planning algorithm that needs to be improved.

[0010] According to another aspect of the present application, a control method for an automatic parking assist function is provided, the method comprising: determining a vehicle trajectory based on the aforementioned benchmark test system; and controlling the vehicle according to the vehicle trajectory so as to perform the automatic parking assist function.

[0011] As a supplement or alternative to the above-mentioned scheme, the above-mentioned method may also include: using an environmental perception sensor to obtain the scene around the vehicle, wherein, based on the benchmark test system, determining the vehicle trajectory includes: sending the scene around the vehicle to the benchmark test system; receiving the most suitable path planning algorithm under the scene from the benchmark test system; and using the path planning algorithm to plan the vehicle trajectory from the current position of the vehicle to the target parking position.

[0012] As a supplement or alternative to the above scheme, in the above method, based on the benchmark test system, determining the vehicle trajectory includes: determining the path planning algorithm with the best overall performance in the entire scenario library based on the benchmark test system; and using the path planning algorithm to plan the vehicle trajectory.

[0013] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the medium includes instructions, and the instructions execute the above method when executed.

[0014] According to another aspect of the present application, a computer program product is provided, comprising a computer program, and when the computer program is executed by a processor, the method as described above is implemented.

[0015] The benchmarking system for the automatic parking assistance function of the embodiment of the present application includes: a scenario creation module, a path planning module and a trajectory evaluation module, wherein the scenario creation module can be used to create a parking scene, the path planning module uses one or more path planning algorithms to plan the vehicle trajectory from the current position of the vehicle to the target parking position in the parking scene; the trajectory evaluation module evaluates the vehicle trajectory based on multiple indicators. By evaluating the trajectory, it is possible to determine the best algorithm in the current scene and improve the target algorithm by identifying the gap. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The above and other objects and advantages of the present application will become more fully apparent from the following detailed description in conjunction with the accompanying drawings, wherein the same or similar elements are represented by the same reference numerals.

[0017] Figure 1 A schematic diagram of the structure of a benchmark test system for an automatic parking assistance function according to an embodiment of the present application is shown;

[0018] Figure 2 A schematic flow chart of a method for controlling an automatic parking assistance function according to an embodiment of the present application is shown;

[0019] Figure 3 A schematic flow chart of a method for controlling an automatic parking assistance function according to another embodiment of the present application is shown; and

[0020] Figure 4 A schematic diagram of a parking lot scene is shown. DETAILED DESCRIPTION

[0021] Figure 1 FIG. 1 is a schematic diagram showing a structure of a benchmark test system 1000 for an automatic parking assistance function according to an embodiment of the present application. Figure 1 As shown, the benchmark test system 1000 includes: a scenario creation module 110, a path planning module 120, and a trajectory evaluation module 130. In the benchmark test system 1000, the scenario creation module 110 is used to create a parking scene, and the parking scene at least includes the current position of the vehicle and the target parking position; the path planning module 120 is used to plan the vehicle trajectory from the current position of the vehicle to the target parking position in the parking scene by using one or more path planning algorithms of the automatic parking assistance function; and the trajectory evaluation module 130 is used to evaluate the vehicle trajectory based on multiple indicators.

[0022] In the context of this application, the term "automatic parking assist function" is also referred to as an automatic parking assist system or an automatic parking assist function. As the name implies, the system can automatically help users park their vehicles into parking spaces. The automatic parking assist function (APA) mainly utilizes on-board sensors throughout the vehicle itself and the surrounding environment to measure the relative distance, speed and angle between the vehicle itself and surrounding objects, to achieve perception and recognition of parking spaces and surrounding environments, and then calculates the operation process through the on-board processor / on-board computing platform or cloud computing platform. The controller calculates the appropriate parking path based on the identified information and controls the actuator to implement the steering and acceleration and deceleration of the vehicle, and automatically and correctly completes the parking action into the parking space, so as to realize automatic parking in, out and partial driving functions, and achieve safe and smooth parking of the vehicle.

[0023] The benchmark test system 1000 for the automatic parking assistance function according to the embodiment of the present application can understand the best algorithm for path planning in the automatic parking assistance function in the current scenario through test evaluation, and can improve the target algorithm through the identified gaps.

[0024] In the benchmarking system 1000, the scene creation module 110 is used to create a parking scene, which includes at least the current position of the vehicle and the target parking position. In one embodiment, the parking scene also includes configurable objects or parameters, including vehicle parameters (such as length, width, wheelbase, turning radius), obstacles around the vehicle, and ground signs.

[0025] Figure 4 A typical parking lot scene diagram is shown in Figure 1. Figure 4 As shown, the vehicle 410 is ready to park in a parking space 420. In this parking lot scene, other obstacles are also included, such as other vehicles 424 and 426, and walls / pillars 422 and 428. Based on this scene, since the current position of the vehicle and the target parking position are known, a trajectory 430 from the vehicle's stopping position to the target parking final position can be planned by calling the path planning algorithm in the automatic parking assistance function.

[0026] In one or more embodiments, the scenario creation module 110 may be used to create various scenarios to simulate different parking scenarios. The collection of various scenarios may be stored in a scenario library.

[0027] The path planning module 120 is used to plan a vehicle trajectory in the parking scene created by the scene creation module 110 using one or more path planning algorithms of the automatic parking assistance function. For example, in the parking scene A created by the scene creation module 110, a first vehicle trajectory is planned using a first path planning algorithm; and in the same parking scene A, a second vehicle trajectory is planned using a second path planning algorithm different from the first path planning algorithm. These planned vehicle trajectories can be provided to the trajectory evaluation module 130 for evaluation based on multiple indicators.

[0028] In one or more embodiments, the path planning algorithm may include, but is not limited to, a geometry-based algorithm, a graph search-based algorithm, and a reinforcement learning-based algorithm. In fact, whether it is a geometry-based algorithm, a graph search-based algorithm, or a reinforcement learning-based algorithm, it will have its limitations. Therefore, by generating trajectories in the same scenario and then handing them over to the trajectory evaluation module 130 for evaluation, it can help identify the performance differences of different algorithms in specific scenarios, and improve the target algorithm based on the differences.

[0029] In one or more embodiments, the improvement of the target algorithm can be divided into multiple levels: (1) The simplest level is parameter setting: no matter which algorithm is used, it is to balance hardware resource consumption and performance, so parameters must be set; (2) A little more complicated, some constraints in the algorithm can be modified; (3) Even more complicated, a special algorithm can be introduced for a specific scenario.

[0030] In one embodiment, the path planning module 120 is configured to: receive a user's selection of the path planning algorithm (eg, select from a plurality of candidate path planning algorithms); and call the selected path planning algorithm to plan the vehicle trajectory in the parking lot scenario.

[0031] In one embodiment, the path planning module 120 is further configured to receive an ideal vehicle trajectory determined based on expert knowledge. In this embodiment, the path planning module 120 may provide the ideal vehicle trajectory together with other planned vehicle trajectories to the trajectory evaluation module 130 for evaluation.

[0032] The trajectory evaluation module 130 is used to evaluate the vehicle trajectory based on multiple indicators. The multiple indicators can be predetermined or based on user input. In one embodiment, the trajectory evaluation module 130 is configured to: receive multiple indicators input or selected by the user and their priorities (or weights of various indicators), the indicators including trajectory length, trajectory smoothness, number of gear shifts, and occupied space; and evaluate the vehicle trajectory based on the multiple indicators to determine the most suitable path planning algorithm or vehicle trajectory in the parking lot scenario, and the path planning algorithm that needs to be improved.

[0033] For example, for a certain scenario, algorithm A identifies more parking attempts than algorithm B (algorithms A and B here can also be the trajectories of human drivers). Based on the comprehensive evaluation of the trajectory evaluation module 130, the most suitable path planning algorithm or vehicle trajectory in the scenario and the path planning algorithm that needs to be improved can be determined. For example, the technician can analyze whether the target algorithm for this scenario can be improved by modifying parameters, constraints or algorithms. In addition, the overall performance of each algorithm in the entire scenario library (under different scenarios) can also be compared.

[0034] Figure 2 FIG. 1 is a flow chart showing a method for controlling an automatic parking assistance function according to an embodiment of the present application. Figure 2 As shown, the control method of the automatic parking assist function includes:

[0035] In step S210 , a vehicle trajectory is determined based on the aforementioned benchmark test system (eg, benchmark test system 1000 ); and

[0036] In step S220 , the vehicle is controlled according to the vehicle trajectory so as to perform the automatic parking assistance function.

[0037] In one embodiment, step S210 includes: determining the path planning algorithm with the best overall performance in the entire scene library based on the benchmark test system 1000; and using the path planning algorithm to plan the vehicle trajectory. In other words, the path planning algorithm with the best overall performance in the entire scene library can be determined in advance (for example, offline) using the benchmark test system 1000. In this way, when using the automatic parking assistance function for path planning, the best path planning algorithm will be used. The control method of the automatic parking assistance function of this embodiment is relatively simple to implement, and the existing control scheme of the automatic parking assistance function is slightly modified.

[0038] Figure 3 FIG. 2 shows a flow chart of a control method of an automatic parking assistance function according to another embodiment of the present application. Figure 3 As shown, the control method of the automatic parking assist function includes:

[0039] In step S310, the scene around the vehicle is acquired using an environmental perception sensor;

[0040] In step S320, the scene around the vehicle is sent to a benchmark test system;

[0041] In step S330, receiving the most suitable path planning algorithm in the scenario from the benchmark test system;

[0042] In step S340, the path planning algorithm is used to plan a vehicle trajectory from the current position of the vehicle to the target parking position; and

[0043] In step S350 , the vehicle is controlled according to the vehicle trajectory so as to perform an automatic parking assistance function.

[0044] Different from Figure 2 The control method shown in Figure 3 In the control method of the automatic parking assist function shown, there is no need to determine in advance the path planning algorithm with the best overall performance in the entire scenario library. Instead, for the current specific scenario, the most suitable path planning algorithm (e.g., algorithm name or algorithm number) is provided by the benchmark test from the system. In this way, after understanding the above information, the automatic parking assist function of the vehicle can call the most suitable path planning algorithm according to the current scenario to plan the vehicle trajectory, and then control the vehicle according to the vehicle trajectory. Figure 3 The control method of the automatic parking assist function shown in the figure has increased interaction with the benchmark test system (the process is relatively simple). Figure 2 The control method shown is more complex), but can achieve better parking control performance.

[0045] In addition, it is easy for those skilled in the art to understand that the control method of the automatic parking assistance function provided by the above one or more embodiments of the present application can be implemented by a computer program. For example, the computer program is contained in a computer program product, and when the computer program is executed by a processor, the control method of the automatic parking assistance function of one or more embodiments of the present application is implemented. For another example, when a computer-readable storage medium (such as a USB flash drive) storing the computer program is connected to a computer, running the computer program can execute one or more embodiments of the present application and the control method of the automatic parking assistance function.

[0046] In summary, the benchmark test system for automatic parking assistance function of the embodiment of the present application includes: a scenario creation module, a path planning module and a trajectory evaluation module, wherein the scenario creation module can be used to create a parking scene, the path planning module uses one or more path planning algorithms to plan the vehicle trajectory from the current position of the vehicle to the target parking position in the parking scene; the trajectory evaluation module evaluates the vehicle trajectory based on multiple indicators. By evaluating the trajectory, the best algorithm in the current scenario can be understood and the target algorithm can be improved by identifying the gap.

[0047] Although only some of the embodiments of the present application are described, it should be understood by those skilled in the art that the present application can be implemented in many other forms without departing from its subject matter and scope. Therefore, the examples and embodiments shown are considered to be illustrative rather than restrictive, and the present application may include various modifications and substitutions without departing from the spirit and scope of the present application as defined in the claims.

Claims

1. A benchmarking system for automatic parking assistance functions, It is characterized in that The system comprises: A scene creation module, used to create a parking scene, wherein the parking scene at least includes the current position of the vehicle and the target parking position; a path planning module, configured to plan a vehicle trajectory from the current position of the host vehicle to the target parking position in the parking lot scenario using one or more path planning algorithms of the automatic parking assistance function; and The trajectory evaluation module is used to evaluate the vehicle trajectory based on multiple indicators.

2. The benchmark testing system according to claim 1, in, The parking lot scene also includes configurable objects or parameters, including vehicle parameters, obstacles around the vehicle, and ground signs.

3. The benchmark testing system according to claim 1, in, The path planning algorithms include: a geometry-based algorithm, a graph search-based algorithm, and a reinforcement learning-based algorithm.

4. The benchmark testing system as claimed in claim 3, in, The path planning module is configured as follows: receiving a user's selection of the path planning algorithm; and The selected path planning algorithm is called to plan the vehicle trajectory in the parking lot scenario.

5. The benchmark testing system as claimed in claim 4, in, The path planning module is also configured to: Receive an ideal vehicle trajectory determined based on expert knowledge.

6. The benchmark testing system according to claim 1 or 5, in, The trajectory assessment module is configured to: receiving a plurality of indicators and their priorities input or selected by a user, the indicators including track length, track smoothness, number of gear shifts, and occupied space; and The vehicle trajectory is evaluated based on the multiple indicators to determine the most suitable path planning algorithm or vehicle trajectory in the parking lot scenario and the path planning algorithm that needs to be improved.

7. A control method for an automatic parking assistance function, It is characterized in that The method comprises: Determining a vehicle trajectory based on the benchmark test system according to any one of claims 1 to 6; and The vehicle is controlled according to the vehicle trajectory to perform the automatic parking assist function.

8. The method according to claim 7, further comprising: include: Use environmental perception sensors to obtain the scene around the vehicle. Wherein, based on the benchmark test system, determining the vehicle trajectory comprises: sending a scene around the vehicle to the benchmark test system; receiving from the benchmark system the most suitable path planning algorithm for the scenario; and The path planning algorithm is used to plan the vehicle trajectory from the current position of the vehicle to the target parking position.

9. The method according to claim 7, in, Based on the benchmark test system, determining the vehicle trajectory includes: Determining the path planning algorithm with the best overall performance in the entire scene library based on the benchmarking system; and The vehicle trajectory is planned using the path planning algorithm.

10. A computer-readable storage medium, It is characterized in that The medium includes instructions which, when executed, perform the method of any one of claims 7 to 9.

11. A computer program product comprising a computer program, It is characterized in that When the computer program is executed by a processor, the method according to any one of claims 7 to 9 is implemented.

Images