LMI-Based Design of Distributed Controllers to Achieve Component Swapping Modularity

摘要

Component swapping modularity (CSM) refers to distributed modular control design in networks of swappable smart components. CSM reduces control design effort and complexity in platform-based systems. Existing CSM methods achieve promising results for low order multi-input-multi-output systems. However, the lack of generalization, heavy computational burden, and to a lower extent, designer’s involvement limit the applications of the existing CSM methods. Thus, this brief utilizes linear matrix inequalities (LMIs) to present an almost automatic CSM algorithm which converges rapidly and is easy to generalize to an arbitrary linear system. The LMI-based CSM is designed to maintain both disturbance attenuation and quadratic stability. Also, it is desired to satisfy specific time response criteria. Thus, the proposed algorithm combines$mathcal {H}_{2}$optimization and robust$mathcal {H}_infty $optimization to obtain a distributed control composed of parts with prescribed orders and partially tunable parameters. The designer’s involvement is dramatically reduced to the iterative tuning of two scalar parameters to optimize the fixed part of the controller. The proposed algorithm incorporates reference tracking. Also, stability measures and design criteria are checked numerically at each step. The LMI-based CSM algorithm has been numerically verified using an engine idle speed control example.