Robust control and scheduling codesign for networked embedded control systems

Robust control and scheduling for networked embedded control systems (NECS) with uncertain but interval-bounded time-varying computation and transmission delay is addressed in this paper. The NECS is described by a set of continuous-time plant models and associated quadratic cost functions. Since the uncertainty of the computation and transmission delay affects the discretized plant models and cost functions in a nonlinear manner, a polytopic overapproximation of the uncertainty utilizing a Taylor series expansion is considered. For the resulting discrete-time switched system model with polytopic uncertainty, a periodic control and online scheduling (PCS_on) strategy is proposed to guarantee stability and performance of the resulting controlled system. The design is based on a periodic parameter-dependent Lyapunov function and exhaustive search. Furthermore, a method for reducing the online complexity of the PCS_on strategy is presented. The effectiveness of modeling and design is evaluated for networked embedded control of a set of inverted pendulums.