基于支持向量机的逆动力学模型辨识及应用

建立系统逆动力学过程模型已经成为许多非线性系统控制问题研究与实现中的关键问题。该文应用支持向量机(SVM)回归方法实现了对热工对象的逆动力学过程在线辨识,并通过三个典型的仿真算例对基于SVM的非线性系统逆动力学过程模型的有效性进行了考察。仿真结果表明,基于SVM的逆动力学过程模型不仅具有较高的辨识精度,同时还具有较为理想的泛化性能和在线跟踪能力;利用所建立的系统逆动力学过程模型能够获得恰当的控制作用,保证系统的输出按照给定的轨迹达到设定值。

Identification of Inverse Dynamics Model Based on Support Vector Machine and Its Application

To research the control problem of nonlinear system, it is a crucial problem to establish its inverse dynamic model. The on-line identification for inverse dynamics of thermal object based on support vector machine (SVM) was realized, and the validity of inverse dynamic model of nonlinear system based on SVM was proved according to three typical simulation examples. The result shows that the inverse dynamic model based on support vector machine has not only upper identification precision, but also quite perfect generalization and traceable ability. The right control can be acquired that ensures the output of model to achieve the setting value according scheduled track by utilizing the identification model of inverse dynamics of system.