|Distributed State Estimation with Bounded Errors over Multiple Access Channels
|Ghassen Zafzouf, Girish N. Nair, University of Melbourne, Australia
|T.2: Information and Control
|Topics in Information Theory
|Click here to download the manuscript
|Click here to watch in the Virtual Symposium
|Although state estimation in networked control systems is a fundamental problem, few efforts have been made to study distributed state estimation via multiple access channels (MAC's). In this article, we give a characterization of the zero-error capacity region of an M-input, single-output MAC at any finite block-length. To this end, nonstochastic information-theoretic tools are used to derive the converse and achievability proofs. Next, a tight condition to be able to achieve uniformly bounded state estimation errors over such a MAC is provided. The obtained condition establishes a connection between the intrinsic topological entropies of the linear systems and the zero-error capacity region of the MAC.