Modeling and dynamic gain scheduling for networked systems with bounded packet losses

In this paper, the modeling and dynamic gain scheduling design of the Networked Control System(NCS) with packet losses are investigated. Instead of finding statically boundary condition for stability of NCS, we develop a linear quadratic regulation(LQR) gain scheduler method to dynamically compensate for degradation of the NCS real-time performance caused by lossy network. The NCS with packet losses is modeled as a jump linear system(JLS) by multi-rate sampling approach. A bank of packet-loss-dependant LQR gains is obtained via linear quadratic optimal control theory. One-on-one relationship between packet loss value and LQR gain is described through a lookup table. Simulation results of a networked robotic arm show the proposed algorithm can guarantee system stable and fast convergence rate of system states.