增程式电动汽车的智能体组网及推理技术

Agent Networking and Reasoning Technology for Extended-range Electric Vehicle

针对传统增程式电动汽车在集成控制和能量分配中存在的问题,提出一种基于多智能体组网技术的增程式电动汽车动力总成控制方法。在JADE(Java agent development framework)软件平台上构建动力部件智能体模型,利用SQL Server建立相应数据库;在接收到整车需求功率后,智能体间进行信息交互,并根据自身状态进行调整,完成对能量的合理分配;最后在MATLAB/Simulink中建立整车动力学模型对所提方法进行仿真验证。结果表明,基于多智能体组网技术的控制方法可以实现增程式电动汽车基本的能量管理,与传统恒温式控制方法相比,在行驶相同里程下,整车燃油消耗降低了3.15%。

Aiming at the problems existing in the integrated control and energy distribution of traditional extendedrange electric vehicle(EREV), a powertrain control method for EREV based on multi-agent networking technology was proposed. The power component agent model was built on JADE(Java agent development framework)software platform, and the corresponding database was built by SQL server. After receiving the required power of whole vehicle, the agents interacted with each other and adjusted according to their own state to complete the reasonable distribution of energy. Finally, the vehicle dynamics model was established in MATLAB/Simulink to verify the proposed method. The results show that the control method based on multi-agent networking technology can realize the basic energy management of EREV,compared with the traditional constant temperature control method, the fuel consumption of whole vehicle is reduced by 3.15% under the same mileage.