A pointwise approach for synthesizing semi-active control laws for vehicle suspension systems

Active and semi-active vehicle suspensions can provide better performance trade offs compared to passive vehicle suspensions. Semi-active vehicle suspensions have attracted a lot of attention since they only need small amount of power and enable improvement in ride quality comparable to active vehicle suspensions. The desired performance for vehicle suspensions is to isolate passengers from the violation due to road disturbance for increasing ride comfort while satisfying the constraints of road traction and suspension space. A semi-active vehicle suspension utilities variable dampers to adjust the damping rate of the suspension in tune with road disturbance. Mathematically, the synthesis of a semi-active suspension can be formulated as an optimal control problem for a bilinear system with constraint on control input.; Presented in this dissertation is a methodology for synthesizing semi-active control laws for vehicle suspensions. The methodology is based on pointwise optimization. The pointwise approach avoids the difficulty for solving a two-point boundary value problem associated with classical optimal control, while optimal performance can be achieved locally in time. In contrast to the open-loop solution for classical optimal control, the resulting control law has a state feedback structure suitable for real time implementation. A novel design which combines pointwise minimization and an "optimal aim strategy" (OAS) concept was developed to synthesize: (1) a modified optimal aim strategy control, (2) a hybrid control scheme-utilizing continuous and discrete controls, and (3) a model predictive control for semi-active vehicle suspensions.; Another approach which utilizes clipped action coupled with frequency domain control design techniques was also developed. In this approach, (1) clipped {dollar}Hsb{lcub}infty{rcub}{dollar} control, and (2) clipped QFT-type control were synthesized. For comparison, the numerical solutions of classical optimal control were obtained by using a conjugate gradient method.; Computer simulations for a quarter-car suspension model were used to evaluate the performance of synthesized semi-active control laws. The performance is evaluated on the reductions of car-body acceleration, tire deflection and suspension deflection subject to three different road disturbances. The results of simulations confirm the efficacy of the proposed approaches.