The invention discloses a dual-motor energy management method and system based on deep reinforcement learning, and belongs to the technical field of intelligent control, and the method comprises the steps: obtaining a gear mode of a target electric vehicle; constructing a vehicle speed prediction network model; constructing a joint model with an Actor network and a plurality of Critic networks; wherein the Actor network is connected with the Critic network; real-time data of the target electric vehicle in different gear modes are collected, and the real-time data comprise the current gear, the current vehicle speed, the required torque, the SOC value and the SOH value; taking the predicted vehicle speed as a first observation variable, taking the real-time data as a second observation variable, inputting the first observation variable and the second observation variable into a joint model, and outputting a plurality of predicted gears corresponding to the first observation variable; carrying out discrete unification on the predicted gear, and outputting a target predicted gear; and taking the target predicted gear as the gear of the target vehicle at the next moment. And the operation stability of the electric vehicle is improved, and energy consumption is reduced.
本发明公开了一种基于深度强化学习的双电机能量管理方法和系统,属于智能控制技术领域,方法包括:获取目标电车的档位模式;构建车速预测网络模型;构建具有Actor网络和多个Critic网络的联合模型;其中,Actor网络和Critic网络连接;采集目标电车在不同档位模式下的实时数据,其中,实时数据包括当前档位、当前车速、需求扭矩、SOC值和SOH值;将预测车速作为第一观察变量、将实时数据作为第二观察变量,将第一观察变量和第二观察变量均输入至联合模型,输出第一观察变量相应的多个预测档位;将预测档位进行离散唯一化,输出目标预测档位;以目标预测档位作为下一时刻目标车辆的档位。进而提升电动汽车的运行稳定性,降低能源消耗。
Dual-motor energy management method and system based on deep reinforcement learning
一种基于深度强化学习的双电机能量管理方法和系统
2024-01-19
Patent
Electronic Resource
Chinese
Near space aircraft energy system management method based on deep reinforcement learning
European Patent Office | 2024
|HEV energy management method based on deep reinforcement learning A3C algorithm
European Patent Office | 2020
|Deep Reinforcement Learning-Based Vehicle Energy Efficiency Autonomous Learning System
British Library Conference Proceedings | 2017
|