支持向量机α阶逆系统控制——连续非线性系统

Support vector machines α th-order inverse control—nonlinear continuous systems

针对传统逆系统方法中逆模型难以建立的问题,提出了连续非线性系统基于最小二乘支持向量机(LS-SVM)α阶的逆系统控制方法.该方法用具有径向基核函数(RBF)的LS-SVM,离线建立被控对象的静态非线性逆模型.由静态非线性逆模型外加若干表征非线性动态特性的积分器,构成了连续非线性系统的α阶逆系统.将得到的LS-SVMα阶逆系统串连在原系统之前,得到基本上线性化的伪线性系统,进而将复杂的非线性问题转化为线性问题.仿真结果表明,在没有被控对象先验知识的情况下,利用该方法能准确地建立连续非线性系统的逆模型.基于SVM的α阶逆系统方法适应于较一般的连续非线性系统,且具有良好的控制性能.

To deal with the squares support vector mac d h fficulties of inverse modelling in the traditional inverse system method, a least nes （LS-SVM）α th-order inverse system method for general nonlinear continuous systems was proposed. LS-SVM with the radial basis function （RBF） kernel was used in the method, and the nonlinear offline static inverse model of the controlled plant was built. Some integrators denoting the nonlinear dynamic characteristics were added to the nonlinear static inverse model, and theα th-order inverse model of the nonlinear continuous system was constructed. After cascading LS-SVM α th-order inverse system before the original system, a pseudo-linear system with basic linearization was formed, and then the complex nonlinear problem was transformed into a linear problem. Simulation results show that the presented method can accurately construct the inverse dynamic model of the continuous nonlinear system even without prior knowledge about the controlled plant. The α th-order inverse system method based on SVM can apply to general nonlinear continuous systems, and has satisfactory control performance.