Active Suspension Based on Genetic Algorithm Multiobjective H <, 2> / H <, ∞> robust control

Active suspension as a future major trends in automobile suspension systems, more and more attention and research. However, to fully active suspension damping performance, the need for suitable control strategies and design a suitable controller, it can be said of active suspension control technology is the key. The key paper for active suspension control technology to carry out the research. For solving the problem of active suspension control, the most difficult is the ride comfort and handling stability of the multi-objective optimization. As the vehicle suspension system design evaluation of the most important indicator of the conflict between them is contradictory. Reduce the stiffness of the spring suspension system will get a better comfort, but this time the suspension of the dynamic travel will increase, the magnitude of the larger body is not conducive to the operation of the vehicle stability. The increased spring stiffness, ride comfort will decrease. Active Suspension for solving multi-objective optimization problems, different theories have been based on a large number of methods that most of the body vertical acceleration, suspension travel, with the static load as the optimal ratio of performance objectives, the selected weights, then unified into a target function to optimize. However, this body acceleration may be less than ideal results. For the above problem, the paper first establish two degree of freedom quarter car active suspension system model, then the constraint H ∞ controller design, and application of MATLAB to solve the LMI toolbox, and the results for solving the simulation compared with the passive suspension. Next, the paper combines the H ∞ control and the merits of H2 control design of the H2 / H ∞ mixed-active suspension controller, making the body acceleration in the time domain have a good effect, and the system has some robust stability . Solved by LMI, and simulation results with constrained H ∞ controller were compared. Finally, the paper Genetic algorithm based on H2 / H ∞ controller for solving the hybrid simulation results show that solving the LMI approach although efficient and effective, but for the H2 / H ∞ multi-objective hybrid controller problem solving, there must be conservative , and genetic algorithms for dealing with difficult to solve the nonlinear multi-objective problems, there is a strong optimization. Therefore, based on genetic algorithm H2 / H ∞ controller for active suspension mixture can be effectively controlled, not only greatly improved the ride comfort, to ensure the stability of operation, and has strong robustness. 

Key words: active suspension, LMI, H ∞ control, H2 / H ∞ mixed control, genetic algorithms