Qualitative analysis of a one-step finite-horizon boundary for event-driven controllers

Abstract

Performance optimization for networked and embedded control systems refers to the ability of minimizing controllers’ resource utilization and/or improving control performance. Event-driven control has been shown to be a promising technique for minimizing controllers’ computational demands. However, optimization of control performance for event-driven control has not been fully addressed. For LTI plants, this paper presents a boundary for event-driven controllers that determines at each job execution when the next job execution should occur in order to minimize a continuous-time quadratic cost function while minimizing controllers’ computational demand. Simulation results illustrate the qualitative shape of this boundary.