The Research Of Vehicle Steering Attitude Based On Kalman Filter

Many active safety control technologies have been designed in oversea auto industry, which mostly aim at some key variables donating vehicle steering attitude. However, it is almost impossible to measure these variables directly at reasonable cost, which bring forward challenge to control arithmetic practicability. Based on analysis in effect factors on vehicle turning or changing lanes abruptly, this dissertation focus on application of state estimation theory on vehicle steering attitude.At first, with assistance of MATLAB/SIMULINK software, full vehicle dynamical simulation models are set up based on mathematical model of vehicle dynamics. Different models' responses to various steering conditions, including velocity, front-wheel corner and friction coefficient of road surface, are analyzed, which explain that the tire nonlinearity between slip angle and cornering force have effects on the simulation results. Meanwhile, the reliability and validity of the model with consideration of tire nonlinearity, lateral mass transfer and axis slip are verified by the steady static circular test and steering wheel angle step input test using a Unique (NJ1020) car.Then based on validated simulation model the factors effecting on steering attitude including Vehicle configuration parameters (rear roll steer coefficients, the position of center of mass, wheelbase, yaw inerter, roll damping, roll stiffness and distance between center of mass to roll center of roll axle) and Wheel alignment parameter-camber angle setting are discussed when vehicle turning or changing lanes abruptly. And for the first time above analysis is carried out under the condition of tire cornering force approach to saturation.Finally a kalman filter is designed to estimate yaw rate, side slip angle and roll angle which are generalized coordinates denoting vehicle steering attitude. Firstly based on 'Magic Formula' a bilinear tire model is set up to correct tire lateral forces, which are calculated from 'bicycle model' with high speed or large turning angle. Meanwhile an adaptive kalman filter is also set up to avoid radiation. This research solves the problem that some state parameters in vehicle dynamic control process are too difficult to measure on-line, and provides a feasible, accuracy and low-cost way for the measurement of vehicle state parameter.