基于多Agent的电动汽车底盘智能控制系统框架

为了解决传统集成控制框架灵活性和可扩展性不足以及传统汽车底盘控制系统不适用于新型分布式轮毂电机驱动纯电动汽车的问题,提出了基于多Agent的四轮独立驱动纯电动汽车的底盘智能动态综合控制系统框架,分析了框架各个层次的功能和相互关系。以底层控制层中直接横摆力矩控制Agent为例,搭建控制器Agent模型。在MATLAB/Simulink和Carsim联合仿真环境中对前轮转向角阶跃输入工况进行仿真试验。仿真结果表明,搭建的控制器Agent较好地实现了预期的动力学控制目标,有效改善了车辆的横向操纵稳定性能,奠定了控制框架的基础。     

[分布式驱动电动汽车动力学控制]基于多Agent的电动汽车底盘智能控制系统框架

To solve the problems that flexibility and scalability limitations of the traditional integrated chassis control frameworks and the traditional vehicles chassis control system was not suitable for distributed in-wheel motors drive electric vehicles, an intelligent chassis dynamic control system framework was proposed based on multi-Agent for four-wheel independent driving electric vehicles. And the functions and interrelations of layers were analyzed. Taking direct yaw-moment control Agent of underlying control layer as an example, the controller Agent model was designed. In the co-simulation environment including MATLAB/Simulink and Carsim, simulation tests were conducted with the conditions of front wheel steering angle step input. Simulation results show that the controller Agent model may achieve the desired objectives of dynamic control and improve the performances of lateral handling stability effectively, which lays the foundation for the proposed control framework.     

[分布式驱动电动汽车动力学控制]四轮驱动电动汽车多Agent能量管理体系研究

In order to explore the energy flow relationship of each electricity system of 4WD electric vehicles, a framework of energy managements for 4WD electric vehicles was structured based on multi-Agent system, and details of this framework were described. A mathematical model of multi-Agent system for energy managements was built with a decision-planning layer based on state transition and transferable belief model. The models of multi-Agent system for energy managements were built in Netlogo, and the energy configuration status was simulated under the rapid acceleration working conditions of electric vehicles using Netlogo. It indicates that the framework of the built energy management multi-Agent system may adjust the energy distribution weights of each Agent dynamically, and the energy allocation of each Agent maybe coordinated reasonably on the premise of satisfying dynamics requirements.     

四轮驱动电动汽车多Agent能量管理体系研究

为探究四轮驱动电动汽车各用电系统的能量流动关系,基于多Agent系统思想构建四轮驱动电动汽车能量管理体系框架,并对该体系框架进行细部描述。对能量管理体系进行多Agent系统数学建模,基于状态转移和可传递置信模型搭建了决策规划层。利用Netlogo仿真软件建立了能量管理多Agent体系模型,并对电动汽车急加速工况下的能量配置状况进行了动态模拟。结果表明,所建立的能量管理多Agent模型能够动态调节各Agent的能量分配权重,在满足车辆系统动力性要求的前提下协调各Agent的能量配给。