A1265
Title: Tobit filtering for nonlinear systems under random access protocol: Dealing with multirate measurements
Authors: Shuo Yang - Harbin University of Science and Technology (China)
Jun Hu - Harbin University of Science and Technology (China)
Raquel Caballero-Aguila - Universidad de Jaen (Spain)
Antonia Oya - Universidad de Jaen (Spain) [presenting]
Abstract: Over the past years, growing research interest in filtering for networked systems has been driven by a number of crucial challenges. These include limited communication resources, heterogeneous sampling rates, and imperfect measurements. In many practical scenarios, sensor data may be censored as a direct result of resolution constraints or energy-saving strategies. This issue can be further complicated in large-scale systems, where a variety of heterogeneous devices naturally lead to the coexistence of multiple sampling periods. Moreover, when numerous sensors attempt to transmit data simultaneously, random access protocols are often adopted to alleviate communication congestion and reduce collisions. In this context, a Tobit filtering approach is proposed for multirate nonlinear systems with censored measurements operating under a random access protocol. More specifically, an innovative filtering scheme is developed, where a minimized upper bound of the filtering error covariance matrix is obtained, and a recursive procedure is established to determine the filter gain matrix. Furthermore, a sufficient condition for the boundedness of the filtering error is provided from an analytic perspective. Finally, some experimental comparisons are conducted to validate the effectiveness and highlight the advantages of the proposed Tobit filtering algorithm.
