CN-121979045-A - Automobile control method and device based on combined search

CN121979045ACN 121979045 ACN121979045 ACN 121979045ACN-121979045-A

Abstract

The application belongs to the technical field of vehicle-mounted intelligent control, and discloses an automobile control method and device based on combination search, wherein the method comprises the steps of obtaining voice data of a user and carrying out intention analysis to obtain an optimization target and constraint conditions; generating formal task description according to optimization targets and constraint conditions, acquiring user portraits, vehicle environment data and vehicle state data and constructing state space vectors, carrying out combined search in a preset atomic action library according to the formal task description and the state space vectors by utilizing a heuristic graph search algorithm to obtain a plurality of alternative control strategies, calculating comprehensive scores of the alternative control strategies by adopting a multi-target evaluation algorithm, determining target control strategies according to the comprehensive scores or user selection instructions, and executing the target control strategies. The application can generate feasible cross-domain alternative control strategies in the vehicle executable preset atomic action library and the multidimensional environment state space.

Inventors

LEI ZIJIAN
XIONG YU
HOU SHAOYANG
Quan Jianxian
HE HONGBIN
ZHOU MINGHUA
TAN MINGXIANG
Ruan Gaowan
WEI YONGYOU

Assignees

东风柳州汽车有限公司

Dates

Publication Date: 20260505
Application Date: 20260121

Claims (10)

1. An automobile control method based on combination search, comprising: Acquiring voice data of a user and carrying out intention analysis to obtain an optimization target and constraint conditions; Generating formalized task descriptions according to the optimization targets and the constraint conditions; acquiring user portraits, vehicle environment data and vehicle state data and constructing state space vectors; performing combined search in a preset atomic action library according to the formalized task description and the state space vector by using a heuristic graph search algorithm to obtain a plurality of alternative control strategies; Calculating the comprehensive score of each alternative control strategy by adopting a multi-objective evaluation algorithm; and determining a target control strategy according to the comprehensive score or the user selection instruction and executing the target control strategy.
2. The method for controlling an automobile based on a combined search according to claim 1, wherein the steps of obtaining user voice data and performing intention analysis to obtain an optimization target and constraint conditions include: Acquiring the user voice data and converting the text to obtain a text required by a user; And inputting the text required by the user into a natural language understanding model in the vehicle-mounted field to obtain an optimization target and constraint conditions.
3. The method of claim 1, wherein the vehicle environment data includes outside air temperature, GPS location, navigation planned route, and along-road charging station information, and the vehicle status data includes current vehicle speed, motor torque, battery state of charge, and individual consumer status information.
4. The method for controlling an automobile based on a combined search according to claim 2, wherein the preset atomic motion library comprises a plurality of atomic motions including a motion number, a motion description, a state variable, an adjustable parameter range of the state variable, a motion execution time-consuming and a dependent condition.
5. The method for controlling an automobile based on a combined search according to claim 4, wherein the performing a combined search in a preset atomic action library according to the formal task description and the state space vector by using a heuristic graph search algorithm to obtain a plurality of alternative control strategies includes: Taking the state space vector as a root node of a search tree, and searching in the preset atomic action library according to constraint conditions in the formalized task description and the dependency conditions of the atomic actions to obtain compatible atomic actions; Taking the consistent atomic actions as state expansion nodes of the search tree until the search depth reaches a preset depth, the search duration exceeds a preset time limit or the accumulated cost of the search tree meets the optimization target in the formalized task description; And combining all the state expansion nodes in the search tree according to the search sequence to obtain the alternative control strategy.
6. The method for controlling an automobile based on a combined search according to claim 1, wherein the determining a target control strategy according to the composite score or a user selection instruction and executing comprises: Arranging the alternative control strategies in descending order according to the comprehensive scores; and selecting the first alternative control strategy as the target control strategy.
7. The method for controlling an automobile based on a combination search according to claim 5, further comprising: acquiring an execution result of the target control strategy and user feedback information; optimizing the natural language understanding model and the multi-objective evaluation algorithm according to the user feedback information; And optimizing the heuristic graph searching algorithm according to the comparison result of the execution result and the optimization target.
8. An automobile control device based on a combination search, comprising: The intention analysis module is used for acquiring the voice data of the user and carrying out intention analysis to obtain an optimization target and constraint conditions; the task formalization module is used for generating formalized task description according to the optimization target and the constraint condition; the acquisition module is used for acquiring user portraits, vehicle environment data and vehicle state data and constructing a state space vector; The combined search module is used for carrying out combined search in a preset atomic action library according to the formalized task description and the state space vector by utilizing a heuristic graph search algorithm to obtain a plurality of alternative control strategies; The scoring module is used for calculating the comprehensive score of each alternative control strategy by adopting a multi-objective evaluation algorithm; and the execution module is used for determining a target control strategy according to the comprehensive score or the user selection instruction and executing the target control strategy.
9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the combined search based vehicle control method according to any one of claims 1 to 7 when the computer program is executed.
10. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the car control method based on a combined search as claimed in any one of claims 1 to 7.

Description

Automobile control method and device based on combined search Technical Field The application relates to the technical field of vehicle-mounted intelligent control, in particular to an automobile control method and device based on combination search. Background Along with the continuous improvement of the intelligent and networking degree of automobiles, a vehicle-mounted intelligent operating system becomes a core component of modern automobiles. In the prior art, the main functions of the vehicle-mounted intelligent system, such as a vehicle-mounted system developed based on Android Automotive, aliOS, QNX and other platforms, are concentrated in aspects of information entertainment, navigation, voice control and the like. However, the prior art has significant drawbacks in addressing the complex, multi-level interaction intentions of the user and providing an intelligent, proactive, comprehensive solution. The prior art can only provide a single and linear response after understanding the user intention, for example, for the user intention of optimizing the energy consumption strategy, the prior system may only provide a suggestion of switching to the economical driving mode or displaying the nearest charging station, and the real-time state of the vehicle, the external environment and the combination of the atomic actions of the bottom layer executable by the vehicle (such as adjusting the output power curve of the driving motor, coordinating the working duty ratio of the air conditioner compressor, adjusting the regenerative braking intensity, planning the kinetic energy recovery strategy containing the specific gradient road section in advance and the like) cannot be comprehensively considered, so that the intelligent service of the automobile lacks depth and accuracy. Disclosure of Invention The application provides an automobile control method and device based on combined search, which can generate feasible cross-domain alternative control strategies in a preset atomic action library and a multidimensional environment state space which can be executed by an automobile. In a first aspect, an embodiment of the present application provides an automobile control method based on a combined search, including: Acquiring voice data of a user and carrying out intention analysis to obtain an optimization target and constraint conditions; generating formalized task descriptions according to the optimization targets and the constraint conditions; acquiring user portraits, vehicle environment data and vehicle state data and constructing state space vectors; performing combined search in a preset atomic action library according to formalized task description and state space vectors by using a heuristic graph search algorithm to obtain a plurality of alternative control strategies; calculating the comprehensive score of each alternative control strategy by adopting a multi-objective evaluation algorithm; and determining a target control strategy according to the comprehensive score or the user selection instruction and executing the target control strategy. Further, the obtaining the user voice data and performing intent analysis to obtain the optimization target and the constraint condition includes: acquiring user voice data and converting text to obtain a text required by a user; and inputting the text required by the user into a natural language understanding model in the vehicle-mounted field to obtain an optimization target and constraint conditions. Further, the vehicle environment data comprises outside air temperature, GPS position, navigation planning route and along-road charging station information, and the vehicle state data comprises current vehicle speed, motor torque, battery charge state and state information of each electric equipment. Further, the preset atomic action library comprises a plurality of atomic actions, and the atomic actions comprise an action number, an action description, a state variable, an adjustable parameter range of the state variable, action execution time consumption and a dependent condition. Further, the method for searching the heuristic graph in the preset atomic action library according to the formal task description and the state space vector to obtain a plurality of alternative control strategies includes: The state space vector is used as a root node of a search tree, and searching is carried out in a preset atomic action library according to constraint conditions in formal task description and dependent conditions of atomic actions to obtain compatible atomic actions; Taking consistent atomic actions as state expansion nodes of the search tree until the search depth reaches a preset depth, the search duration exceeds a preset time limit or the accumulated cost of the search tree meets the optimization target in formal task description; and combining the state expansion nodes in the search tree according to the search sequence to obtain an alternative control strat