Collaborative-Prediction-Based Recursive Filtering for Nonlinear Systems Subject to Low-Duty-Cycle Scheduling

• Published in: IEEE transactions on signal processing Vol. 72; pp. 1 - 13
• Main Authors: Gao, Hongyu, Yu, Lindong, Hou, Nan, Song, Jinbo, Li, Yue
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.01.2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Collaboration; Collaborative prediction algorithm; Covariance; Covariance matrices; Filtration; low-duty-cycle scheduling; Nonlinear systems; nonlinear time-varying systems; Performance degradation; Petroleum; Prediction algorithms; recursive filtering; Resource scheduling; Scheduling; Signal processing algorithms; Time measurement; Time-varying systems; Wireless sensor networks
• ISSN: 1053-587X; 1941-0476
• DOI: 10.1109/TSP.2023.3343558

In this paper, the recursive filtering problem is investigated for the nonlinear time-varying systems with the collaborative prediction algorithm (CPA) under a low-duty-cycle wireless transmission mechanism. To significantly save energy and bandwidth resources, low-duty-cycle scheduling (LDCS) is employed in practical engineering. Under this communication scheduling, the sensor nodes are allowed to remain in dormant states for a relatively long period of time. The objective of this study is to design a filtering scheme that can ensure the filtering performance for the nonlinear systems under the LDCS. To solve the problem of filtering performance degradation due to high data sparsity caused by the low duty cycle, the CPA combined with the zero-order holder (ZOH) is introduced into the filtering scheme. The desired gain matrix is first computed recursively by minimizing the obtained filtering error covariance upper matrix. Next, the boundedness of the filtering error covariance is discussed. Finally, the developed filtering approach based on the CPA and ZOH under the low-duty-cycle scheduling is verified by a simulation case for its effectiveness.