Resilient State Estimation for 2‐D Systems With Dead‐Zone‐Like Censoring and Multiplicative Noises Under Binary Encoding Scheme

ABSTRACT This article addresses the resilient Tobit Kalman estimation problem for a class of linear shift‐varying two‐dimensional (2‐D) systems with multiplicative noises and dead‐zone‐like measurement censoring. To practically characterize the signal transmission between the sensor and the remote filter, a binary encoding scheme is employed. Under this scheme, the measurement output is first encoded into a finite‐length binary bit string, which is then transmitted through memoryless binary symmetric channels where the bits flip probabilistically. Furthermore, a Tobit model incorporating the multiplicative noises and the binary encoding strategy is developed for the measurement outputs subject to dead‐zone‐like censoring. The objective of this paper is to design an unbiased resilient Tobit Kalman estimator for the considered 2‐D system, where an upper bound for the estimation error variance is derived and further minimized by selecting appropriate estimator gains. Additionally, given the critical role of the binary bit string length in the binary encoding process, its impact on the estimation performance is thoroughly analyzed. Finally, a simulation study is conducted to validate effectiveness of the proposed estimation algorithm.