基于MPC与EKF算法的电动汽车状态估计与横向稳定性控制  

State estimation and lateral stability control for electric vehicles based on EKF and MPC algorithm

作　　者：邓涛 罗俊林 汪明明 DENG Tao;LUO Junlin;WANG Mingming(School of Mechatronics & Vehicle Engineering, Chongqing Jiaotong University, Chongqing 400074, China)

机构地区：[1]重庆交通大学,机电与车辆工程学院

出　　处：《汽车安全与节能学报》2017年第3期287-295,共9页Journal of Automotive Safety and Energy

基　　金：国家自然科学基金资助项目(51305473);中国博士后科学基金资助项目(2014M552317);重庆市博士后研究人员特别资助项目(xm2014032)

摘　　要：为对电动汽车状态进行精确估计,提高车辆的安全性能,建立了车辆横向稳定性预测控制模型。采用模型预测控制算法(MPC),基于双轨二自由度整车模型,选取整车质心侧偏角、横摆角速度及轮胎滑移率作为状态变量,四轮分配扭矩作为控制变量。考虑驾驶员操作负荷和转向与驱/制动的平滑性,建立模型代价函数。设计扩展Kalman滤波器(EKF)预估和评价车辆当前状态,并利用二次规划(QP)算法进行滚动时域优化求解,得到4个车轮的最优扭矩。搭建Simulink和Carsim联合仿真平台,选取转向盘正弦输入工况,分别对高低附着路面进行仿真分析。结果表明:该预测控制模型能够有效保持车辆的横向稳定性。The model predictive control(MPC)algorithm was used to establish the predictive control model of vehicle lateral stability to estimate the electric vehicle state accurately and improve the safety performance of vehicle,based on two degrees of freedom model.The vehicle sideslip angle,yaw angular and tire slip ratio were selected as state variables and the four-wheel torque distribution was treated as control variables.Considering the driver's operation load and the smooth of steering and driving/braking operations,the cost function of model was built.The extended Kalman filtering(EKF)estimator was designed to estimate and evaluate the current state of the vehicle.And the rolling horizon optimization was solved using Quadratic Program(QP)algorithm to obtain the optimal torque of the four wheels.The co-simulation platform of Simulink and Carsim was constructed and simulated under the high and low attached road surface respectively on the premise that the steering wheel was sinusoidal input.The results show that the MPC medel can effectively maintain the vehicle's lateral stability.

关 键 词：电动汽车 横向稳定性控制 联合仿真 模型预测控制(MPC) 扩展Kalman滤波器(EKF) 

分 类 号：U469.722[机械工程—车辆工程]