The invention provides a target network training method for automatic driving, an electronic device and a system, and the method comprises the steps: obtaining training data from a plurality of clients, and the clients are disposed in a vehicle, and are used for constructing the training data through a simulation environment processing module and a first actor network; based on the training data, obtaining a second decision result corresponding to the first state information through a second actor network, and obtaining an advantage value corresponding to the first decision result through a commentator network; and determining a loss value of the first actor network according to a comparison result of the first decision result and the second decision result and the advantage value, and updating parameters of the first actor network according to the loss value until a target network meeting a preset condition is obtained. According to the technology of the invention, massive training data can be obtained from a plurality of clients, abundant training samples are provided for the training of the target network, and the decision performance of the trained target network can be improved.
本申请提出一种用于自动驾驶的目标网络的训练方法、电子设备和系统,其中的方法包括:从多个客户端分别获取训练数据,客户端部署于车辆,用于通过仿真环境处理模块和第一演员网络构建训练数据;基于训练数据,通过第二演员网络得到第一状态信息对应的第二决策结果,以及通过评论家网络得到第一决策结果对应的优势值;根据第一决策结果和第二决策结果的比较结果以及优势值,确定第一演员网络的损失值,并根据损失值对第一演员网络的参数进行更新,直至得到符合预设条件的目标网络。根据本申请的技术,能够从多个客户端获取到海量的训练数据,为目标网络的训练提供了丰富的训练样本,有利于提高训练出的目标网络的决策性能。
Target network training method for automatic driving, electronic equipment and system
用于自动驾驶的目标网络的训练方法、电子设备和系统
23.05.2023
Patent
Elektronische Ressource
Chinesisch
IPC: | G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / B60W CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION , Gemeinsame Steuerung oder Regelung von Fahrzeug-Unteraggregaten verschiedenen Typs oder verschiedener Funktion |
Target network training method, device and system for automatic driving
Europäisches Patentamt | 2023
|Automatic driving equipment and target identification method
Europäisches Patentamt | 2024
|Method and device for training automatic driving model and electronic equipment
Europäisches Patentamt | 2024
|Automatic driving method, related model training method, electronic equipment and storage medium
Europäisches Patentamt | 2022
|Europäisches Patentamt | 2022
|