Abstract: Steering wheel angle and vehicle speed are two inputs of vehicle system model for handling stability research. However, the two inputs are hardly to be persistently exciting (PE) at the same time in a single test so that the identification model is incomplete under this condition. Models with varied structure are identified on the basis of a set of pylon course slalom tests in order to analyze properties of vehicle model with constant vehicle speed. According to researching findings, a nonlinear model based Wiener structure including a linear identification model with steering wheel angle as the sole input and a nonlinear function of vehicle speed is proposed. A linear model is identified by subspace method at the lowest test speed. The relationship between system output residual and vehicle speed is obtained by correlation analysis. After that, a set of nonlinear functions for every output are fitted by least square method. The final identification model containing the linear model and a nonlinear function is certified by another slalom test at different vehicle speed as a good predictor. Moreover, the tests for identification are national standard ones so this modeling method has good engineering value.

Key words: Nonlinear vehicle model, Non-persistent exciting input, Subspace method, System identification