Optimal robust non-fragile Kalman-type recursive filtering with finite-step autocorrelated noises and multiple packet dropouts

• Published in: Aerospace science and technology Vol. 15; no. 6; pp. 486 - 494
• Main Authors: Feng, Jianxin, Wang, Zidong, Zeng, Ming
• Format: Journal Article
• Language: English
• Published: Issy-les-Moulineaux Elsevier SAS 01.09.2011; Elsevier
• Subjects: Applied sciences; Computer science; control theory; systems; Computer systems and distributed systems. User interface; Control theory. Systems; Dropouts; Exact sciences and technology; Filtering; Filtration; Finite-step autocorrelated noises; Mathematical models; Modelling and identification; Multiplicative noises; Noise; Non-fragile filtering; Optimization; Packet dropouts; Recursive; Robust Kalman filtering; Software; Uncertainty; Non-fragile filtering; Multiplicative noises; Robust Kalman filtering; Finite-step autocorrelated noises; Packet dropouts; Infinite impulse response filter; Probabilistic approach; Kalman filter; Lossy channel; Modeling; Stochastic system; Recursive filtering; Multiplicative noise; Uncertain system; Minimal variance; System identification; Non deterministic system; Signal to noise ratio
• ISSN: 1270-9638; 1626-3219
• DOI: 10.1016/j.ast.2010.10.006

In this paper, the optimal robust non-fragile Kalman-type recursive filtering problem is studied for a class of uncertain systems with finite-step autocorrelated measurement noises and multiple packet dropouts. The system state, measurement output and filter parameters are all subject to stochastic uncertainties or multiplicative noises, where the measurement noises are finite-step autocorrelated. When there exist multiple packet dropouts in the system output, the original system is converted into an auxiliary stochastic uncertain system by the augmentation of system states and measurements. The process noises and measurement noises of the auxiliary system are shown to be finite-step autocorrelated and cross-correlated. Then, a robust non-fragile Kalman-type recursive filter is designed that is optimal in the minimum-variance sense. The proposed filter is not only robust against the uncertainties in the system model and measurement model, but also non-fragile against the implementation error with the filter parameters. Simulation results are employed to demonstrate the effectiveness of the proposed method.