Research On Vehicle State Estimation With Steer-By-Wire System Based On Nonlinear Estimation Method

In order to solve the traffic safety problem, automobile industry began to constantly introduce new technology to improve the vehicle handling stability and active safety with the development of the computer control technology and intelligent information processing technology from the1990s. Automobile steering system performance directly affects the vehicle handling stability. Compared with the traditional mechanical steering system, vehicle with steer-by-wire system (SBW) has more flexible steering characteristics. Through the steer-by-wire control strategy, vehicle can achieve the active steering control and electronic stability control. Before the realization of the control strategies, vehicle driving state must be acquired accurately and in real-time. Considering the complexity of vehicle dynamics control, test level of the existing vehicle sensor, test cost etc, some key vehicle state can’t be measured directly or in low cost. While complexity and integration level of vehicle chassis control system increase, there will exist some phenomenon, for instance, difference among signal rates is big, and sensor’s sampling periods differ. Therefore, the present single rate nonlinear Kalman filter has some limitations in the estimation of vehicle state. Aiming to solve the problem of vehicle state estimation with steer-by-wire system, nonlinear estimation methods and multirate sampling control scheme are merged together and the following aspects are discussed and studied:Firstly, based on investigation and analysis of SBW’s structure, principle, control strategy and characteristics, the components in SBW system are modelling. Moreover, an11degree of freedom vehicle dynamics model is given including the R.Hess driver and Pacejka tire model, and a driver-vehicle close loop system is established. According to the ISO test, virtual vehicle handling stability tests are carried out on Carsim and Matlab/Simulink co-simulation platform using SBW’s control strategy. The results prove validity of the vehicle model.Secondly, Extended Kalman Filter (EKF) is an effective nonlinear estimation algorithm of vehicle states. But it has two major defects:(1) bad robustness to the variation of vehicle parameters;(2) bad tracking ability to the abrupt change of states. To overcome these disadvantages, the Strong Track Filter (STF) is introduced to estimate the vehicle states, which can improve the estimation performance of abrupt change states and the robustness of variable parameters. Besides, the mode of multiple sub-optimal fading factors in STF will destroy the symmetry and positive definiteness of the prediction error covariance matrix, which results in the filter divergence. An improve STF (ISTF) based on the Cholesky decomposition of the multiple sub-optimal fading factors matrix is proposed and applied to vehicle states estimation. The simulation results illustrate that ISTF is more stable and better than STF in the robustness of the initial fading factor’s value selection.Thirdly, KF has obtained a good effect in vehicle state estimation, but its performance is influenced by the precision of the model and the availability of noise statistics. Therefore, a nonlinear adaptive Kalman filter based on fictitious noise compensating technique and an unscented Kalman filter (UKF) are used to estimate vehicle states. Considering an idea of STF, a Suboptimal Multiple Fading Adaptive Extended Kalman Filter (SMFAEKF) is proposed which can estimate noise characteristics and has stronger tracking ability to the suddenly changing states simultaneously. The simulation results show that UKF has a slightly higher estimation precision than EKF under the same conditions, and SMFAEKF is better than EKF on the estimating accuracy, tracking speed and restraining noise.Finally, considering complexity and integration level of vehicle chassis control system increase, there will exist some phenomenon, for instance, difference among signal rates is big, and sensor’s sampling periods differ. So the single sampling rate digital control strategy is no longer suitable for this vehicle chassis control system. Vehicle state space modeling method based on input and output multirate digital control strategy is derived, which can effectively expand the number of input and output vector in the system, can obtain more system information, and improve the SBW’s control quality. Combined UKF with input multirate vehicle state space model, an input multirate unscented Kalman filter (IMRUKF) is proposed and applied to estimate yaw rate, slip angle and longitudinal velocity. The results prove that IMRUKF improves the ability of state estimator by fusing multi-input information, which has better stability and the estimation accuracy increases37.8%-65.3%than single rate Kalman algorithm.