稳定的非线性系统广义预测控制

Stable generalized predictive control for nonlinear system

针对广义预测控制(GPC)算法稳定性分析困难,对参数未知非线性系统提出一种稳定广义预测控制(DGPC)方法。该方法首先将非线性系统转换为时变线性系统,然后利用三次样条基函数逼近时变系统中的系数,通过带时变遗忘因子的递推最小二乘算法辨识系数获得对象模型。基于模型通过性能指标中的前馈增益设计来保证控制系统稳定,仿真结果验证了该方法的有效性。

This paper presented the stable generalized predictive control（ GPC） for a class of nonlinear system with unknown parameters to untie a difficulty of stability analysis of GPC. This method transformed the nonlinear system into the time-varying linear system,and used a group of cubic spline functions to approach the time-varying coefficients,then usd least square method with variable forgetting factors to get the plant model. Based on the linear model the design of feed forward gain made the control system stable. Simulation results demonstrate the effectiveness of this method.