Real-time implementation of the multi-swarm repetitive control algorithm

In this paper, a real-time implementation of the recently developed particle multi-swarm repetitive controller is presented. The proposed solution employs the Particle Swarm Optimizer (PSO) to solve in on-line mode the dynamic optimization problem (DOP). The proposed controller, unlike the classic PSO, has additional mechanisms to track non-stationary optimal solution to the control signal shaping task in the constant-amplitude constant-frequency voltage source inverter with an LC output filter. The challenges of real time implementation come from limited sampling period to evaluate optimal solution. The algorithm is composed of several subswarms, where each subswarm focuses on a part of the control signal along the pass. The implementation procedures of the time-distributed algorithm are described in detail. Simulation results and experimental verification on typical off-the-shelf microcontroller are presented.