Concept for hybrid optimization for schedule design in nonlinear networked control

This paper introduces a computational framework which allows to assess a-priori the control performance of a nonlinear networked control system (NCS) for which actuator signal communication is time triggered. The plant system is nonlinear and continuous-time. A discrete nonlinear controller has been designed without the knowledge of the communication network. The integration of the control system under the assumption of time-triggered actuator signal communication requires the choice of a feasible communication policy. It is shown that a continuous-time quadratic infinite horizon cost can be evaluated for the NCS via a sum-of-squares (SOS) approach. The equivalent linear cost computation for the linearized closed-loop is used to minimize computational effort by reducing the number of variables for the cost computation and avoiding SOS-computations where possible. A numerical example proves the feasibility of the approach.