Investigation On The Multi-Objective H₂/H_∞ Rubust Control Of The Active Suspension Based On Genetic Algorithm

The tendency of employing active suspensions system in the future automobile has been drawing increasingly more attention and study. However, the design of an appropriate controlling strategy and a controller are required to give a full play to the vibration deduction function of the active suspension. The controlling strategy is the key to studying the active suspension and this paper studied the critical controlling strategy of active suspension.As to the controlling matter of the active suspension, the most difficult part is to achieve a multi-goal optimization of smoothness and operation stability. As the most important evaluation index of the design of automobile suspension system, they are contradictory. The deduction of the spring stiffness of suspension system will bring satisfactory comfort but result in a larger suspension travel and body range which will do no good to the operation stability of the automobile body. On the contrary, higher spring stiffness will lower down comfort satisfactory. There are a lot of approaches based on different theories to this dilemma, mainly focusing on integrating functions of vertical acceleration, suspension travel, static and dynamic load ratio and etc. as the goal of optimization, deciding on a weight and optimizing them in an object function. But, this will end up with unsatisfactory body acceleration.To address the problems above, this paper set up a mathematic model of 2 DOF quarter car active suspension systems, designing a constrained H∞controller, employing the LMI toolbox of MATLAB to seek a solution, and simulate the result to compare with.passive suspension.Next, based on the strengths of H∞controller and H2 controller, active suspensions of the H2/H∞hybrid controller were designed so as to achieve satisfactory acceleration in a time domain and the Robust stability at the same time. A solution was also done by means of LMI and a comparison of the stimulation result and constraint H∞controller was made.At last, this paper sought a solution towards H2/H∞hybrid controller by means of Genetic Algorithm. The stimulation showed that though LMI is efficient and effective it is conservative towards the solution for the multi-goal H2/H∞hybrid controller. Whereas Genetic Algorithm has extraordinary optimization capability in addressing formidable non-linear multi-goal problems. Therefore, H2/H∞hybrid controller based on Genetic Algorithm can have an effective control of active suspensions, which can not only greatly improve the automobile smoothness and ensure operation stability, but can also has strong robust.