State Observers for Suspension Systems with Interacting Multiple Model Unscented Kalman Filter Subject to Markovian Switching
Zhang, Zipeng,Xu, Nan,Chen, Hong,Wang, Zhenfeng,Li, Fei,Wang, Xinyu
DOI: https://doi.org/10.1007/s12239-021-0126-z
2021-01-01
International Journal of Automotive Technology
Abstract:This paper presents a novel model-based observer algorithm to address issues associated with nonlinear suspension system state estimation using interacting multiple model unscented Kalman Filters (IMMUKF) under various road excitation. Due to the fact that practical working condition is complex for the suspension system, e.g. additional load. Meanwhile, the changed sprung mass parameter will induce model changed of suspension system, and it can lead to state transition between various models. To tackle the mentioned issue, the models of road profile and suspension system are first established to describe the nonlinear suspension dynamics. Then, considering the variation of sprung mass under various movement conditions, an unscented Kalman Filter (UKF) algorithm is proposed to identify the sprung mass. Based on the interacting multiple model (IMM) and Markov Chain Monte Carlo (MCMC) theory, a novel IMMUKF observer is developed to estimate the movement state of suspension system. The stability conditions for the proposed observer is calculated using the stochastic stability theory. Finally, simulations and validations are performed on a quarter vehicle suspension system under various ISO road excitations, to validate the UKF and IMMUKF algorithms for acquiring suspension system states, and results illustrate that the maximum root mean square error of state estimation for the proposed algorithm is less than 7.5 %.