作　　者： ; ; ;

机构地区： 湖南大学电气与信息工程学院

出　　处： 《电子测量与仪器学报》 2007年第1期39-43,共5页

摘　　要： 支持向量机(SVM)具有很强的非线性逼近能力与泛化能力,文章研究了基于SVM的非线性系统逆模型辨识,并设计了基于模糊控制补偿的SVM逆控制系统。由SVM辨识的逆模型作为前馈控制器,形成直接逆模型控制器。同时,设计模糊控制器构成反馈补偿控制,克服逆模型的建模误差,提高系统鲁棒稳定性。仿真研究表明,SVM具有优良的逆模型辨识能力,基于模糊控制补偿的支持向量机逆控制系统的动态性能好、跟踪精度高、鲁棒稳定性强。 Support vector machines （SVM） have good nonlinear function approximation ability and generalization capability. In this paper, SVM based inverse model identification of nonlinear system is studied. SVM based inverse control system with fuzzy control controller （ FLC ） compensation was designed. SVM based inverse model of nonlinear system is used as feed - forward controller to form direct inverse model controller. At the same time, fuzzy logic controller is used as feedback controller that compensates for inversion error and rejects disturbances. Simulations demonstrate that SVM has good nonlinear approximation capability for inverse model, and the proposed control system has good dynamic and static performances as well as good robustness.

关 键 词： 非线性系统 机器学习 支持向量机 逆系统 模糊控制

领　　域： [自动化与计算机技术] [自动化与计算机技术]