J.S.-H. Tsai, S.-J. Pan, W.-C. Liao, S.-M. Guo (Taiwan), and L.-S. Shieh (USA)
Observer/Kalman filter identification, system identification, singular system, digital-redesign, evolutionary programming
Modeling of decentralized linear observers and trackers for the unknown sampled-data interconnected large-scale linear singular system with time delay and closed-loop decoupling property is proposed in this paper. Through the off-line observer / Kalman filter identification (OKID), the appropriate (low-) order decentralized linear observers with direct transmission terms from input to output for the unknown sampled-data interconnected large-scale linear singular system with time delay are determined. Then, the high-gain linear quadratic suboptimal analogue observer and tracker are proposed based on the each decentralized equivalent regular system with a direct transmission term from input to output. Subsequently, the prediction-based digital redesign method is utilized to obtain practically implemental digital observer and tracker for the sample data system. With the high-gain property, the decentralized digital-redesign trackers have the decoupling property for the closed-loop interconnected system. Finally, appropriate weighting of the each decentralized observer can be obtained to improve the performance of observer-based tracker by the evolutionary programming (EP).
Important Links:
Go Back
