基于遗传算法的主动悬架多目标H_2/H_∞鲁棒控制

Investigation on the Multi-objective H_2/H_∞ Robust Control of the Active Suspension Based on Genetic Algorithm

利用多目标遗传算法(μMOGA)并结合线性矩阵不等式(LMI)对四分之一主动悬架的混合H2/H∞控制的保守性进行了研究。利用多目标遗传算法对控制增益进行搜索,H2和H∞范数通过解矩阵不等式得到。将基于μMOGA/LMI优化得到的非支配解与多目标控制工具箱的结果进行比较发现:基于μMOGA/LMI的优化结果明显优于多目标控制箱优化结果。利用μMOGA/LMI对四分之一主动悬架进行设计,得到的主动悬架性能指标仿真结果明显优于被动悬架的性能指标。

Multi--objective genetic algorithm （μMOGA） was used to deal with the conservation problem about the active suspension multi- objective control produced while solving the mixed H2/H∞ multi--objective robust control with linear matrix inequality（LMI） method. After the control gain was randomly sought by μMOGA, the H2 and H∞ norm of the dosed--loop system were calculated by LMI toolbox. The comparison between non- dominated solve calculated by μMOGA/LMI and the multi--objective control toolbox in MATLAB shows clearly the advantages of theμMOGA/LMI method. In the framework of μMOGA/LMI optimization,the mixed H2/H∞ multi -objective robust control active suspensions were designed on a quarter- vehicle model. The comparison of the simulation results between passive and active suspension shows that ride comfort performance is clearly improved while assuring suspension deflection within bounds and convincing a firm contact of wheels to road.