Contents I. Introduction
One of the fundamental characteristics of networked control systems (NCSs) [1] is the existence of an imperfect communication network between computational and physical entities. In such setups, an analytical framework to assess impacts of communication and data-rate limitations on the control performance is strongly required. In this article, we adopt information-theoretic tools to analyze such requirements. Specifically, we consider sequential coding of correlated sources initially introduced by [2] (see also [3]), which is a generalization of the successive refinement source-coding problem [4], [5]. In successive refinement, source coding is performed in (time) stages and the goal is to “refine” the description of the source with every new stage when more information is available. Sequential coding differs from successive refinement in that at the second stage, encoding involves describing a correlated (in time) source as opposed to improving the description of the same source. To accomplish this task, sequential coding encompasses a spatio-temporal coding method. In addition, sequential coding is a temporally zero-delay coding paradigm since both encoding and decoding must occur in real time. The resulting zero-delay coding approach should not be confused with other existing works on zero-delay coding, see, e.g., [6]–[11], because it relies on the use of a spatio-temporal coding approach whereas the aforementioned papers rely solely on temporal coding approaches.
