The invention discloses a method and a device for training a vehicle driving decision model. According to the invention, virtual environment information (i.e., sample environment information) similarto the actual environment information is generated through a pre-trained first GAN; and the sample environment information is inputted to a pre-trained second GAN to generate a virtual driving decision (i.e., a sample driving decision) matched with the sample environment information, and then the sample environment information and the sample driving decision corresponding to the sample environmentinformation are used to train a vehicle driving decision model. According to the training mode, the dependence of the vehicle driving decision model on historical data is reduced, when the vehicle isin an environment with complex traffic conditions, the driving decision adapting to the environment where the vehicle is located and the driving state of the vehicle can also be output, and the generalization of the vehicle driving decision model is high.
本说明书公开了一种车辆行驶决策模型的训练方法及装置。本说明书将实际环境信息通过预先训练的第一GAN生成类似于实际环境信息的虚拟环境信息(即,样本环境信息),并将该样本环境信息输入到预先训练的第二GAN生成与该样本环境信息相匹配的虚拟行驶决策(即,样本行驶决策),然后用该样本环境信息、与该样本环境信息对应的样本行驶决策训练车辆行驶决策模型。该种训练方式降低了车辆行驶决策模型对历史数据的依赖,当车辆处于交通状况较为复杂的环境中时,也能输出适应于车辆所处的环境和车辆行驶状态的行驶决策,使得车辆行驶决策模型的泛化性较强。
Method and device for training vehicle driving decision model
一种车辆行驶决策模型的训练方法及装置
2020-04-10
Patent
Electronic Resource
Chinese
Training method of automatic driving decision model and vehicle control method and device
European Patent Office | 2023
|Training method and device of automatic driving decision model
European Patent Office | 2022
|European Patent Office | 2024
|Driving intention prediction model training method, vehicle driving method and device
European Patent Office | 2024
|