基于LS-SVM/PID复合逆系统的预测控制

摘要

In order to improve the control performance of predictive control algorithm, a new predictive control algorithm of nonlinear systems based on the least squares support vector machine and PID inverse control was proposed. Basis on PID control,the algorithm uses LS-SVM to offline establish the nonlinear system inverse model of controlled object as a feed-forward controller, forms direct inverse control, and overcomes robustness of direct inverse system method, and with the original system series constitutes a pseudo linear system, then, realizes predictive control of system combined with predictive control algorithm. The simulation results show that the algorithm has the better tracking performance and the anti-interference ability.%为了提高预测控制算法的控制性能,提出一种基于最小二乘支持向量机(LS-SVM)/PID复合逆系统的预测控制算法.该算法在PID控制的基础上,利用LS-SVM离线建立被控对象的非线性逆模型作为前馈控制器,形成直接逆控制,其克服了逆系统方法鲁棒性不强的缺陷,并与原系统串联构成一个伪线性系统；然后,结合预测控制算法实现系统的预测控制.仿真结果表明,该算法具有较好的跟踪性能和抗干扰能力.