出力フィードバック制御則による分散制御問題の準最適解

概要

論文の詳細を見る
In the optimal control problem, all results are based upon the assumption of centrality, that is, all the information available about the system and decision making based upon this information arises at a single input. The design leads to systems in which every sensor output affects every actuator input. In the modelling and design of multi player control problems, the objectives in the optimal choice may be met by formulating the control problem as a differential game. The purpose of this work is to determine the decentralized control strategies with output feedback form which guarantee certain values obtained from the performance indices. We determine the suboptimal control strategies such that obtained performance indices are bounded by certain values of suboptimality degree γ_i where the bounded values are defined as γ_iJ^^-_i^* for the linear closed-loop Nash game. Thus, the suboptimal control strategies are not equilibrium solutions. The control strategies with output feedback form require neither state estimation nor reconstruction. In this work, the control strategies guarantee the trace criteria associated with each player which are interpreted as expected values of standard quadratic criteria. Sufficient conditions for suboptimal control strategies with output feedback form are given for the infinite problem.
一般社団法人日本機械学会の論文
1996-07-25