Adaptive robust control based on immune optimization and LS-SVRM for nonlinear systems
结合免疫优化和LS-SVRM观测器的非线性系统自适应鲁棒控制

An adaptive robust control algorithm is proposed for a class of SISO uncertain nonlinear control systems. Because the problem of the least squares support vector regression machines(LS-SVRM) is transformed to a quadratic programming problem with linear constraints and the ultimate solution is not a local minimum, the system state vector can be estimated by using an observer based on LS-SVRM, when the system state vector is not completely available. Meanwhile, both the norm of the difference between the optimal approximation parameter vector and the nominal parameter vector, and the bounds of the approximation errors are unknown hypothetically; therefore, we can improve the robustness of the systems by adjusting the estimations of the unknown bounds in the algorithm. Considering the effect of parameters of LS-SVRM upon the performance, we present a new immune algorithm for optimizing the parameters of LS-SVRM to improve the approximation ability of LS-SVRM. The theoretical analysis and a simulation example demonstrate the feasibility and validity of the proposed approach.