Control Strategy for Semi-Active Suspension Based on Suspension Parameter Estimation

Abstract

<div class="section abstract"><div class="htmlview paragraph">This paper presents an adaptive <i>H</i>₂/<i>H</i>_∞ control strategy for a semi-active suspension system with unknown suspension parameters. The proposed strategy takes into account the damping force characteristics of continuous damping control (CDC) damper. Initially, the external characteristics of CDC damper were measured, and a forward model and a back propagation (BP) neural network inverse model of CDC damper were proposed using the measured data. Subsequently, a seven-degree-of-freedom vehicle with semi-active suspension system and <i>H</i>₂/<i>H</i>_∞ controller was designed. Multiple feedback control matrices corresponding to different sprung mass parameter values were determined by analyzing time and frequency domain performance. Finally, a dual observer system combining suspension state and parameter estimation based on the Kalman filter algorithm was established. The estimated parameter was used to determine feedback control matrix, while the observed states were used to calculate the desired damping force of CDC damper. Simulation results show that the proposed adaptive <i>H</i>₂/<i>H</i>_∞ control strategy can estimate sprung mass value in real time, enabling the switching of feedback control matrix based on the estimated results. Consequently, vehicle ride comfort is enhanced while handling stability is not excessively deteriorated. Compared to real-time calculation of control parameters method, the proposed control strategy reduces computations and ensures robustness of semi-active suspension systems.</div></div>