Research On Vehicle Sideslip Angle Estimation Based On Adaptive Kalman Filter

With the car accident caused by loss of stability at high speed gradually increase, people's attention on study for the stability control system gradually improved. Stability control systems is a vehicle active safety technology, which inhibit oversteering and the serious shortage of automobile understeer tendency to improve vehicle handling and stability, to reduce traffic accidents. Sideslip angle is the key control parameters to stability control system. Since the production car hasn't currently direct measurement of sideslip angle, the estimation of sideslip angle is a key technology in the stability control system.Firstly, this paper proposed a estimation method used for linear phase of sideslip angle, estimated the tire cornering stiffness and sideslip angle separately and then evaluate the advantages and disadvantages of this method.Secondly propose an improvement program, which is proposed for non-Linear phase sideslip angle estimation based on adaptive extended kalman filter. Then the road adhesion coefficient, sideslip angle, longitudinal speed, lateral speed, yaw rate, roll angle and roll angular velocity are estimated. Finally, combined commercial simulation software with stability control control algorithms, verified estimated effect of offline simulation experiments for high adhesion and low surface adhesion surface. Full contents include:1.Analysis the significance of the Vehicle Stability Control System and the need to estimate the sideslip angle of the vehicle. Summarized the research current and trends for the sideslip angle of vehicle. Because of the greater impact on the sideslip angle estimation, summarized research current for Tire cornering stiffness and road adhesion coefficient estimates.2.Based on the classical theory kalman filter and the extended kalman filter, according to filtering divergence problem of the kalman filtering process, proposed an adaptive extended kalman filter solution. Do a simple summary for the kalman filter estimates in the field of application to car state estimation3.According to the vehicle parameter uncertainty in the estimation process, based on the algorithm of the classic two degrees of freedom vehicle model for sideslip angle estimation, integrated the least squares method to estimate the tire cornering stiffness and constitute a new estimation algorithm. Do simulation experienment to verify the algorithm in double lane change working conditions. The results show that the algorithm is with high estimated precision in the tire linear stage, but in the nonlinear stage the estimation accuracy greatly decreased and it is need for further improvement algorithm.4.In the non-linear phase the estimation accuracy is declined. So it is proposed adaptive extended kalman filter algorithm. Establish the vehicle model including the movement of vehicle longitudinal motion, lateral motion, yaw motion and roll motion , selected the magic formula tire model as a tire model for the estimation algorithm. The magic formula tire model was normalized and the normalization calculation of the tire model parameters were studied. Finally, using the adaptive extended kalman filter algorithm, write this vehicle model and tire model in state space form required by the kalman filter and discretization of continuous equations. Build the estimation algorithm in Matlab/Simulink environment .5.Do off-line simulation to verify estimation effect for the adaptive extended kalman filter for vehicle state and road adhesion coefficient. Using vehicle models of the commercial simulation software veDYNA, and stability control algorithms of laboratory-developed, modify the connection interface control algorithms with the vehicle model. To determine the control algorithm can work on the vehicle model, and finally verify the proximity of the estimate state and the actual state . Simulation results show that: in the high attachment and low attachment surface conditions, sideslip angle estimation accuracy has been further improved, the road friction coefficient is estimated to meet the requirements of the stability control system, its accuracy needs to be further improved.