面向数据包丢失补偿的网络化广义预测控制

Networked Generalized Predictive Control for Compensation of Packet Loss

针对网络化控制系统中存在的数据包丢失,考虑了基于状态空间模型的网络化广义预测控制问题;在假设反馈通道和控制通道的数据包丢失过程确定可知的情况下,提出了一种采用预测器和预测控制器分别补偿反馈通道和控制通道的数据包丢失对系统性能影响的方法,通过把广义预测控制问题转化为滚动线性二次型最优跟踪问题,基于动态规划给出了网络化广义预测控制器的设计方法,并基于Riccati差分方程非负定解的单调性,给出了末端加权矩阵保证系统稳定性的充分条件,最后通过仿真验证了所提出方法的有效性。

This paper is concerned with state space model based generalized predictive control for networked control systems with packet loss. Given that packet loss process is certainly known, a new method is proposed to compensate the influences of packet loss in feedback channel and control channel on control performance by a predictor and a predictive controller, respectively. After transforming the generalized predictive control problem into a receding horizon linear quadratic tracking problem, the design method of networked predictive controller is given by dynamic programming approach. Sufficient conditions are given by exploring the monotonicity of the nonnegative definite solution of Riccati difference equation so that the end weighting matrix guarrantees the stability of the closed loop system. Simulation results are given to show the effectiveness of our proposed method.