基于反馈线性化的整车振动状态观测算法

Full Vehicle Vibration State Estimation Algorithm Based on Feedback Linearization

针对整车振动状态观测器设计中的整车悬架系统高维非线性特性,提出了反馈线性化卡尔曼滤波算法.基于微分几何理论,通过求解坐标变换,将车辆非线性振动模型变换成一个可观测的标准型,实现系统的精确反馈线性化,进而采用线性卡尔曼滤波算法,针对变换后的线性系统设计观测器,最后通过坐标逆变换获得原非线性系统的状态观测值.仿真结果表明,相比扩展卡尔曼滤波算法,该算法能够提高车辆振动状态观测精度和运算效率.

Aiming at the high-dimensional nonlinearity of full vehicle vibration system in the design of vehicle vibration state observer,a feedback linearization Kalman filter algorithm was proposed.Based on differential geometry theory,a nonlinear vehicle vibration model was transformed into a certain observable normal form via change of state coordinates.Based on the obtained linear system,the observer was designed by using Kalman filter algorithm.Finally the estimated states of nonlinear system were obtained through inverse transformation.Simulation results show that compared with extended Kalman observer,the proposed algorithm can improve the observation accuracy and operation efficiency of vehicle vibration states.