Study On Robust Control For DC-motor-based Active Suspension

This master’s thesis was undertaken between2010and2013at ShanghaiJiao Tong University. The DC-motor-based active suspension was treated asthe research objective. Considering the characteristics of the proposedDC-motor actuator, the robust control algorithm was applied to the activecontrol of this suspension. Meanwhile, the improvement on performance ofactive suspension by robust control algorithm was analyzed.Firstly, based on proposed DC-motor actuator, the mathematical modelof this actuator was established, which included models of DC-motor andball-screw mechanism. Then a current-hysteresis-based motor control systemwas proposed, aiming to make the DC-motor actuator track the referencetorque.Then, by linearizing model of DC-motor actuator, the quarter-car activesuspension model was obtained, including characteristics of actuator. Theμ-synthesis controller for active suspension was proposed with regards tosome possible uncertain parameters and high-order unmodeled uncertainty ofmotor output. Simulation results showed that the μ-synthesis control was ableto improve ride comfort of suspension greatly. Meanwhile, analysis of robustperformance also revealed that the robust performance requirement of systemwas fulfilled by μ-synthesis controller.Thirdly, the main/inner-loop control structure was applied for full-carcontrol, in which the main-loop was used for calculation of active controltorque, and the inner-loop was to track this active control torque rapidly andaccurately. According to H∞control algorithm, the design of main-loopcontroller was based on full-car active suspension model. Simulations wereconducted to verify the control performance of the H∞controller and theperformance was compared with a passive counterpart. It showed that theactive suspension saw great improvement in ride comfort and attenuation in heave, pitch and roll motion.Lastly, by applying mixed H₂/H∞control strategy, vehicle acceleration,suspension working space, dynamic tire deflection and control torque werechosen as H₂performance indices. Meanwhile, suspension working space andcontrol torque were treated as H∞performance indices. When solving mixedH₂/H∞controller, reasonable γ value was selected according to therelationship between H₂norm and H∞norm. It was revealed by simulationresults that the proposed controller was able to improve ride comfort greatlyin both random and bump road profiles. In addition, the suspension workingspace was in acceptable ranges.