电动车驾驶机器人控制器参数自学习方法研究

Research on Self-Learning Method of Controller Parameters for Electric Vehicle Driving Robot

针对人工驾驶和现有的机械式驾驶机器人在电动车续驶里程试验中存在成本高、耗时长和错误率高的问题,利用电动车结构和控制上的独特性,依托自行设计的电动车转毂驾驶机器人,采用电信号完成对车辆的控制。为了提高系统适应性,采用非线性最小二乘法实现电动车模型参数的在线辨识;并基于粒子群算法对车辆PID控制器参数进行在线整定,从而完成对任意设定工况速度曲线的跟随。实车试验表明,提出的方法能实现加速踏板的平滑控制,且控制重复性高,完全可以代替驾驶员进行电动车续驶里程试验。

In order to solve the problems of high cost,long time-consuming and high error rate in the driving mileage test of electric vehicles for manual driving and mechanical driving robots,the uniqueness of the structure and control of electric vehicles is adopted and the self-designed electric vehicles driving robot is relied to complete the control of the vehicle by electrical signals.In order to improve the system adaptability,nonlinearleast squares method is used to realize the online identification of electric vehicle model parameters.Based on the particle swarm optimization algorithm,the vehicle PID controller parameters are adjusted online,which contributes to completing the tracking of the arbitrarily set operating conditions.The actual vehicle test shows that the proposed method can achieve smooth control of the accelerator pedal and has high control repeatability,which can completely replace the driver to perform the electric vehicle driving mileage test.