Intelligent Semi-Active Control For Suspension Vibration Of A Full-Vehicle Model

Recent years have witnessed considerable attention paid to the semi-active vibration control based on Magneto-Rheological dampers because of their quick response, low energy consumption and high reliability in applications. Numerous studies have focused on the development of this technique to the vehicle suspensions on the basis of either a quarter-car model or a half-car model. However, few publications have been available for the semi-active vibration control of full-vehicle suspensions.The primary objective of this dissertation is to study the intelligent control, such as the neuro-fuzzy control, for the suspension vibration of a full-vehicle model with four MR dampers installed between the sprung-mass and the four un-sprung masses, respectively. The main contributions of the dissertation are as follows.In Chapter 2, the dissertation gives the dynamic equation for a full-vehicle suspension with four MR dampers installed between the sprung-mass and the four un-sprung masses, respectively. Then, Chapter 2 shows how to generate the stationary random excitation of the road roughness from the given spectrum amplitude of the road roughness with the random phase. For the vehicle running at a variable speed on a road with stationary random roughness, the dissertation shows how to derive the non-stationary random excitation in the time domain.Chapter 3 presents the intelligent identification approaches systematically for the model of an MR damper, which is simple but accurate for the description of the hysteretic nonlinearity of the MR damper. In this chapter, a fuzzy identification technique, called simple ANFIS, is developed to identify the inverse model of an MR damper. Then, the simple ANFIS is extended to the hierarchical fuzzy identification, i.e., the hierarchical ANFIS, which is again used to identify the direct model and inverse model of the MR damper. A new method to assign the initial values of various ANFIS is also presented. In addition, the moment method is introduced to form the Moment ANFIS. These intelligent identification approaches can also be used to establish the models of other MR dampers.The dissertation focuses on the semi-active vibration control for the linear suspension of the full-vehicle model in Chapter 4. It develops a two-stage control strategy, dividing the full-vehicle suspension and the four MR dampers into two different parts. The upper stage controls the full-vehicle suspension by using the LQG with all the control objectives taken into account and determines the desired restoring...