Analysis of Vehicle Handling and Stability in Frequency Domain Based on System Identification Method

Abstract

Based on data generated from four national standard tests of the constant speed condition, a set of vehicle system models is identified by subspace method for frequency domain characteristics analysis of vehicle system in full speed range. Both identification algorithm and model structure are validated. Based on the analysis results, steady state gains of all outputs as a function of vehicle speed are studied, which proved to be a benefit supplement to national standard. A new estimation method of understeer gradient based on yaw velocity gain characters is proposed and validated in comparison with the value measured by steady state cornering test. With the help of this method, the understeer gradient varying with vehicle speed is estimated, which is helpful to adjust vehicle chassis structure parameters. Similar to steady state gains, the relationship between the phase delay and vehicle speed is nonlinear, which can be described by a range between two fit curves. The major factor related to the phase delay at the same speed is confirmed as the exciting level of input signals quantified by the power spectral density (PSD). The analysis indicates that: the resonant frequency of lateral acceleration is lower than that of yaw velocity at the same speed, which indicated that the former is more suitable as an engineering evaluation index. It also is suggested that the return ability test is more suitable to measure resonant frequency than other tests due to its high-level exciting input. The analyses in frequency domain based on identification method provide new insights into the steady state handling quality of vehicle test in a wide speed range. This analysis method makes full use of test data obtained from the national standard tests so it is a valuable method in engineering.