PID-Based Event-Triggered MPC for Constrained Nonlinear Cyber-Physical Systems: Theory and Application

Model predictive control (MPC), which obtains the control signal via solving an optimal control problem online, has been widely applied in various cyber-physical systems (CPSs) due to its ability to handle the physical constraints on system variables explicitly. To further reduce the communication and computational resource consumption while ensuring the control performance, a novel event-triggered MPC (ET-MPC) framework is proposed for perturbed nonlinear CPSs with input and state constraints. Specifically, rather than utilizing the instantaneous state error to determine the next triggering instant, which may result in residual error and state mutation for the considered system, the developed strategy adopts the proportion, integral, and differential (PID) of the state error within a time horizon simultaneously to ensure the performance of the CPS under the event-based sampling framework. Moreover, the PID-based ET-MPC algorithm is developed, and the theoretical properties, including feasibility and stability, are strictly proved. Finally, the resource saving property, as well as the control performance of the proposed ET-MPC strategy are extensively verified via MATLAB simulation and a real robot platform powered by the robot operating system, respectively.