Adaptive aggregation of modular control

Modular methodologies are popular in various control designs. We propose a performance-based adaptive aggregation for modular controllers. Conventional modular control techniques, like decentralized control, singular perturbation and hierarchical decomposition, use non-adaptive aggregation and are only applicable to a restricted class of nonlinear systems. It is demonstrated in this paper that adaptive aggregation extends the validity and the performance of a modular design to a wider class of nonlinear systems. Rule-based aggregation, induced by fuzzy logic and rough sets, is proposed for ease of design and utilization of a priori knowledge. The emphasis in this paper is on model-independent design.