Abstract：For manipulators with interval uncertain parameters,evolving design methods of robust controllers are proposed based on the genetic algorithm.The encoded parameters of the controller with a given structure are taken as the controller population,and the encoded uncertain parameters of the plant as the plant population.A co-evolution operation is performed on the two populations to obtain a controller with sufficient robustness against the plant interval uncertainties and a plant model that achieves the worst control properties.Results of applications to robust controller design for a two-degree-of-freedom manipulator with parameter uncertainties demonstrate the efficiency of the proposed method.
[1] 解学书,钟宜生.H∞控制理论(H∞ Control Theory)[M].北京;清华大学出版社,1994. [2] 胡跃明.非线性控制系统理论与应用[M].北京;国防工业出版社,2002. [3] Jones A H,Ajlouni N,Kenway S B,et al.Genetic design of robust PID controllers to deal with prescribed plant uncertainties through a process of competitive co-evolution[A].Proceedings of the IEEE International Symposium on Intelligent Control[C].USA;IEEE,1996.372-377. [4] 潘正君,康立山,陈毓屏.演化计算[M].北京;清华大学出版社,1998. [5] Hillis W D.Co-evolving Parasites Improve Simulated Evolution as an Optimization Procedure[M].Redwood City,CA;Addison-Wesley,1991.313-323. [6] Liu M.Decentralized PD and robust nonlinear control for robot manipulators[A].Journal of Intelligent and Robotic Systems,1997,20(2-4);319-332.
