考虑轮胎侧偏刚度在线更新的智能电动汽车状态估计

State Estimation of Intelligent Electric Vehicle Considering Online Updating of Tire Cornering Stiffness

实时准确地估计车辆行驶状态是汽车智能化发展的前提,而现有的研究通常忽略了轮胎侧偏刚度的时变特性,在整车模型中引入线性轮胎模型严重影响到极限工况下车辆状态的估计精度。基于此,提出了一种轮胎侧偏刚度在线更新的智能电动汽车纵向车速、横摆角速度和质心侧偏角的估计算法。基于模糊自适应扩展卡尔曼滤波算法(FAEKF)建立了车辆状态估计模型,采用模糊控制器对扩展卡尔曼滤波中含观测噪声协方差的卡尔曼增益矩阵进行实时调节,达到算法的自适应效果。以带遗忘因子递推最小二乘法(FFRLS)为基础建立了轮胎侧偏刚度估计模型。将两种算法以嵌入式的方式结合提出FAEKF+FFRLS算法,更好地实现了状态与参数联合估计和互相校正,通过Trucksim和MATLAB/Simulink联合仿真对算法进行了验证。结果表明:相比于标准的EKF算法,所提出的状态估计算法具有更高的精度,更好的稳定性和鲁棒性。

The real-time and accurate estimation of vehicle states is the premise of vehicle intelligence development.However,the existing researches usually ignore the time-varying characteristics of tire cormering stiffness,and introducing linear tire model into vehicle model seriously affects the estimation accuracy of vehicle states under extreme conditions.An algorithm for estimating intelligent electric vehicle longitudinal speed,yaw rate and sideslip angle of vehicle mass center with tire comering stiffness updated online is proposed.Based on the fuzzy adaptive extended Kalman filter(FAEKF),the vehicle state estimation model is established.The fuzy controller is used to adjust the Kalman gain matrix including the covariance of observation noise in EKF algorithm in real time to achieve the adaptive effect of the algorithm.Using the forgetting-factor recursive least square method(FFRLS),the estimation model of tire cornering stiffness is established.A new FAEKF+FFRLS algorithm is proposed by combining the two algorithms in an embedded way,which can better realize the joint estimation and mutual correction of states and parameters.The algorithm is verified by co-simulation Trucksim and MATLAB/Simulink.The results show that compared with the standard EKF algorithm,the proposed state estimation algorithm has higher accuracy,better stability and robustness.