CN-122009164-A - Vehicle avoidance control method, device, vehicle, medium, program product and chip system

Abstract

The invention relates to a vehicle avoidance control method, a device, a vehicle, a medium, a program product and a chip system in the technical field of auxiliary driving, which comprise the steps of acquiring state data in multiple dimensions in the running process of the vehicle, determining the type of a running scene, classifying and determining the type of the running scene according to shadow state data in the multiple dimensions when the vehicle is triggered to execute avoidance, determining a target active avoidance strategy according to the state data, the type of the running scene and a mapping association relation, wherein the mapping association relation is the association relation between a state represented by the shadow state data and the corresponding avoidance strategy, and controlling the vehicle to execute active avoidance according to the target active avoidance strategy. According to the shadow state data in the avoidance scene, the driving scene type and the mapping association relation are determined, the avoidance behavior is quantified from the state data of multiple dimensions, the self-adaptive regulation and control of the active avoidance strategy is realized, the accuracy of the active avoidance of the vehicle is improved, and the driving safety of the vehicle is improved.

Inventors

• Du Leihao

Assignees

• 小米汽车科技有限公司

Dates

Publication Date

20260512

Application Date

20260330

Claims (20)

1. 1. A vehicle avoidance control method, characterized by comprising: Acquiring state data in multiple dimensions in the running process of a vehicle, and determining the running scene type of a running scene where the vehicle is located, wherein the running scene type is determined by classifying the scene according to shadow feedback state data, and the shadow feedback state data is shadow state data in the multiple dimensions when triggering the vehicle to execute avoidance; Determining a target active avoidance strategy according to the state data, the driving scene type and a mapping association relation, wherein the mapping association relation is the association relation between the state represented by the shadow state data and the corresponding avoidance strategy; and controlling the vehicle to execute active avoidance according to the target active avoidance strategy.
2. 2. The method of claim 1, wherein the determining the target active avoidance strategy according to the status data, the driving scenario type, and the mapping association relationship comprises: determining an initial active avoidance strategy according to the state data, the driving scene type and the mapping association relation; According to avoidance object data in the state data, determining avoidance information aiming at the avoidance object; And determining the target active avoidance strategy according to the initial active avoidance strategy, the avoidance information and the perceived disturbance corresponding to the driving scene type.
3. 3. The method of claim 2, wherein the determining the target active avoidance strategy according to the initial active avoidance strategy, the avoidance information, and the perceived disturbance information corresponding to the driving scenario type comprises: determining an avoidance reference track according to the initial active avoidance strategy, the avoidance information and the perceived disturbance corresponding to the driving scene type; And determining the target active avoidance strategy according to the avoidance reference track and the state data.
4. 4. The method of claim 3, wherein the determining the target active avoidance strategy based on the avoidance reference trajectory and the state data comprises: Dividing the area corresponding to the avoidance reference track into a plurality of avoidance areas according to the position of the vehicle in the state data; Determining avoidance feasibility information of the plurality of avoidance areas; and determining the target active avoidance strategy according to the avoidance feasibility information of the plurality of avoidance areas, the avoidance reference track and the state data.
5. 5. The method of claim 4, wherein the determining the target active avoidance strategy based on the avoidance feasibility information for the plurality of avoidance regions, the avoidance reference trajectory, and the status data comprises: Under the condition that the avoidance feasibility information of a plurality of avoidance areas represents that active avoidance can be executed, determining an avoidance track to be optimized according to the avoidance reference track, the position information of the avoidance object in the state data, the vehicle state information and the road environment information; And determining the target active avoidance strategy according to the avoidance track to be optimized.
6. 6. The method of claim 5, wherein the determining the active avoidance trajectory to be optimized based on the avoidance reference trajectory, the distribution information of the avoidance objects characterized by the state data, the vehicle state information, and the road environment information comprises: Determining a plurality of candidate avoidance trajectories according to the avoidance reference trajectories and the distribution information of the avoidance objects represented by the state data; And screening the candidate avoidance strategies according to the vehicle state information and the road environment information represented by the state data, and determining an active avoidance track to be optimized.
7. 7. The method of claim 5, wherein the determining the target active avoidance strategy according to the avoidance trajectory to be optimized comprises: performing iterative optimization on the avoidance trajectory to be optimized according to an optimization constraint condition to generate the target active avoidance strategy, wherein the optimization constraint condition comprises one or more of the following: smooth avoidance trajectory, ride comfort, and safety margin.
8. 8. The method of claim 4, wherein the determining the avoidance feasibility information for the plurality of avoidance regions comprises: determining weights of the avoidance areas according to the perceived stability of the avoidance areas, wherein the perceived stability is related to the positional relationship of the avoidance areas relative to the vehicle; According to the weight, determining effective avoidance spaces of the plurality of avoidance areas; And determining avoidance feasibility information of the plurality of avoidance areas according to critical collision conditions and the effective avoidance spaces of the plurality of avoidance areas.
9. 9. The method of claim 3, wherein the initial active avoidance strategy comprises an initial trigger time and an initial avoidance magnitude, the avoidance information comprises a position and a buffer distance of an avoidance object, and determining the avoidance reference trajectory according to the initial active avoidance strategy, the avoidance information, and the perceived disturbances corresponding to the driving scene type comprises: according to the perceived disturbance corresponding to the driving scene type, adjusting the buffer distance to determine a target buffer distance; And determining the avoidance reference track according to the target buffer distance, the position of the avoidance object, the initial trigger time and the initial avoidance amplitude.
10. 10. The method according to any one of claims 1 to 9, wherein the mapping association is constructed by: Determining trigger time reference information and avoidance operation reference information corresponding to each driving scene type according to trigger time and corresponding avoidance operation information in the shadow state data; And determining the mapping association relation according to the trigger time reference information and the avoidance operation reference information corresponding to each driving scene type.
11. 11. The method according to any one of claims 1 to 9, wherein the driving scene type is determined by scene classification by: Determining a plurality of vehicle speed sections according to the vehicle speed in the shadow state data; according to avoidance object data in the shadow state data, determining a plurality of avoidance risk levels for the avoidance object; determining a plurality of perceived disturbance levels according to perceived disturbance information in the shadow state data; And obtaining a plurality of driving scene types according to the plurality of vehicle speed intervals, the plurality of avoidance risk levels and the plurality of perception disturbance levels.
12. 12. The method of claim 11, wherein the determining a plurality of avoidance risk levels from avoidance object data in the shadow state data comprises: according to the avoidance object data in the shadow state data, determining the motion state type and motion attribute characteristics of the avoidance object; And determining a plurality of avoidance risk levels for the avoidance object according to the motion state type and the motion attribute characteristics.
13. 13. The method according to any one of claims 1 to 9, further comprising: Acquiring execution feedback data aiming at the active avoidance strategy; According to the execution feedback data and the execution feedback conditions, determining matching information of the target active avoidance strategy and the driving scene type of the driving scene where the vehicle is located; And under the condition that the matching information represents mismatching, returning the execution feedback data serving as shadow state data, wherein the execution feedback data is used for correcting the driving scene type and the mapping association relation.
14. 14. The method of claim 13, wherein the execution feedback condition comprises one or more of: the driving track of the active avoidance meets the preset track requirement, and the triggering time of the vehicle for executing the active avoidance by human intervention meets the preset triggering requirement.
15. 15. The method according to any one of claims 1 to 9, further comprising: acquiring real-time state data in the plurality of dimensions, wherein the real-time state data is state data in the process that the vehicle executes the target active avoidance strategy; And updating the target active avoidance track in the target active avoidance strategy according to the real-time state data.
16. 16. The method of claim 15, wherein updating the target active avoidance trajectory in the target active avoidance strategy based on the real-time status data comprises: under the condition that the real-time state data represents the speed change of the vehicle and/or the motion state change of the avoidance object, the avoidance track parameters in the target active avoidance strategy are updated in real time; and updating the target active avoidance track in the target active avoidance strategy in real time according to the updated avoidance track parameters.
17. 17. A vehicle avoidance control device, characterized by comprising: The system comprises a scene type determining module, a shadow feedback state data processing module and a scene type determining module, wherein the scene type determining module is configured to acquire state data in multiple dimensions in the running process of a vehicle, determine the running scene type of a running scene where the vehicle is located, and the running scene type is determined by classifying the scene according to the shadow feedback state data, wherein the shadow feedback state data is shadow state data in the multiple dimensions when the vehicle is triggered to execute avoidance; the strategy determining module is configured to determine a target active avoidance strategy according to the state data, the driving scene type and a mapping association relation, wherein the mapping association relation is the association relation between the state represented by the shadow state data and the corresponding avoidance strategy; The control module is configured to control the vehicle to execute active avoidance according to the target active avoidance strategy.
18. 18. The apparatus of claim 17, wherein the policy determination module is configured to include: The initial strategy determination submodule is configured to determine an initial active avoidance strategy according to the state data, the driving scene type and the mapping association relation; The avoidance information determination submodule is configured to determine avoidance information for an avoidance object according to the avoidance object data in the state data; The target strategy determination submodule is configured to determine the target active avoidance strategy according to the initial active avoidance strategy, the avoidance information and the perceived disturbance corresponding to the driving scene type.
19. 19. The apparatus of claim 18, wherein the target policy determination submodule is configured to: determining an avoidance reference track according to the initial active avoidance strategy, the avoidance information and the perceived disturbance corresponding to the driving scene type; And determining the target active avoidance strategy according to the avoidance reference track and the state data.
20. 20. The apparatus of claim 19, wherein the target policy determination submodule is configured to: Dividing the area corresponding to the avoidance reference track into a plurality of avoidance areas according to the position of the vehicle in the state data; Determining avoidance feasibility information of the plurality of avoidance areas; and determining the target active avoidance strategy according to the avoidance feasibility information of the plurality of avoidance areas, the avoidance reference track and the state data.

Description

Vehicle avoidance control method, device, vehicle, medium, program product and chip system Technical Field The disclosure relates to the technical field of assisted driving, and in particular relates to a vehicle avoidance control method, a device, a vehicle, a medium, a program product and a chip system. Background Active safety is taken as a basic function of auxiliary driving, is a guarantee of high-order auxiliary driving safety, and the auxiliary driving function of the vehicle triggers active avoidance through a fixed threshold in the related art, however, due to complex working conditions of the vehicle driving environment, the suitability of triggering active avoidance through the fixed threshold is poor, and the active avoidance is difficult to accurately perform. Disclosure of Invention In order to overcome the problems in the related art, the present disclosure provides a vehicle avoidance control method, apparatus, vehicle, medium, program product, and chip system. According to a first aspect of an embodiment of the present disclosure, there is provided a vehicle avoidance control method, including: Acquiring state data in multiple dimensions in the running process of a vehicle, and determining the running scene type of a running scene where the vehicle is located, wherein the running scene type is determined by classifying the scene according to shadow feedback state data, and the shadow feedback state data is shadow state data in the multiple dimensions when triggering the vehicle to execute avoidance; Determining a target active avoidance strategy according to the state data, the driving scene type and a mapping association relation, wherein the mapping association relation is the association relation between the state represented by the shadow state data and the corresponding avoidance strategy; and controlling the vehicle to execute active avoidance according to the target active avoidance strategy. According to the technical scheme, the driving scene type and the mapping association relation are determined according to the shadow state data in the avoidance scene, and the avoidance behavior is quantified from the state data of multiple dimensions, so that the self-adaptive regulation and control of the active avoidance strategy can be realized, the planning efficiency and accuracy of determining the active avoidance track of the vehicle can be improved, the accuracy of the active avoidance of the vehicle is improved, and the driving safety of the vehicle is improved. In some possible implementations, the determining the target active avoidance policy according to the state data, the driving scene type, and the mapping association relation includes: determining an initial active avoidance strategy according to the state data, the driving scene type and the mapping association relation; According to avoidance object data in the state data, determining avoidance information aiming at the avoidance object; And determining the target active avoidance strategy according to the initial active avoidance strategy, the avoidance information and the perceived disturbance corresponding to the driving scene type. According to the technical scheme, an initial active avoidance strategy is determined according to the state data, the scene type and the mapping association relation. And then the initial active avoidance strategy is optimized according to the avoidance information and the perceived disturbance determined by the avoidance object, so that the adaptability of the active avoidance strategy and the complex and changeable running environment is improved, the vehicle can make reasonable and effective active avoidance actions according to different conditions, the capability of the vehicle for coping with various road conditions is enhanced, and the running safety and stability of the vehicle are effectively improved. In some possible implementations, the determining the target active avoidance strategy according to the initial active avoidance strategy, the avoidance information, and the perceived disturbance information corresponding to the driving scene type includes: determining an avoidance reference track according to the initial active avoidance strategy, the avoidance information and the perceived disturbance corresponding to the driving scene type; And determining the target active avoidance strategy according to the avoidance reference track and the state data. According to the technical scheme, the avoidance reference track is planned according to the initial active avoidance strategy, the avoidance information and the perceived disturbance, the degree of fit between the avoidance reference track and the actual avoidance of a driver is improved, and the capability of the active avoidance to adapt to a complex environment is improved. And then, the real-time state data of the vehicle is combined to finely adjust the avoidance reference track, so that the executability of the target active