Communications and control co-design: a combined dynamic-static scheduling approach

Abstract

This paper investigates the problem of joint design of an output feedback controller and the medium access scheduling policy for networked control systems (NCSs). The capacity of communication networks on sensors side and actuators side are limited so that at any time the sensors and actuators cannot be accommodated simultaneously and that transmission delays are inevitable. The joint design problem is solved in three steps: determining communication sequences (called feasible sequences) for both sides that can preserve controllability and observability, respectively; designing output feedback controller without considering the effect of communication sequences; and setting up a rule that governs the switching between feasible communication sequences. Essentially, the first step is a static scheduling procedure that yields a set of feasible communication sequences for each side. The third step is a dynamic scheduling process resulting in a switching policy in line of a stability criterion of the closed-loop switching delay system. To show the effectiveness of the combined static and dynamic scheduling method, a numerical example is also provided.

Control with sensors/actuators assigned by Markov chains: transition rates partially unknown

“…Definition 1: The system in (9) is said to be stochastically stable if for all finite initial condition φ(t) defined on [−τ ca −τ sc , 0], initial medium access status M σ (0) and M ρ (0), there exists a finite number γ (φ(·), M ρ (0)) > 0 such that (12) holds, where E is the statistical expectation operator.…”

“…These works are restricted to either offline static scheduling or dynamic scheduling. Most recently, an interesting hybrid static-dynamic co-design method was presented in Lu and Guo [12], which combines the advantages of both static and dynamic strategies and solves the control synthesis and scheduling problem without decoupling them.…”

Stochastic Observability of Linear Systems under Access Constraints

“…In this setting, the system's properties need to be discussed under a policy managing the plant-controller (estimator) communication. The communication policies include the "static access scheduling", such as in [3][4][5] where the access of sensors or actuators is determined offline in terms of period communication sequences, and the "dynamic scheduling" [6,7,24,26]. With an extended model whose dynamics incorporate those of the original plant and the access status, Hristu-Varsakelis & Zhang [8,11] discussed linear systems where access sequences consisting of binary-valued integers (0 or 1) are identified to preserve the stability, reachability and observability.…”

“…With an extended model whose dynamics incorporate those of the original plant and the access status, Hristu-Varsakelis & Zhang [8,11] discussed linear systems where access sequences consisting of binary-valued integers (0 or 1) are identified to preserve the stability, reachability and observability. Based on this access description, Guo et al [7,[26][27][28] provided methods of sensor assignment, access-controller co-design, and analytic results of controllability, observability and stability. While, the works on random access, such as access assignment by Markov chain modeling [7,24,29], estimation and fault detection with the access status described as a Bernoulli process [13,14], provided an alternative methodology to deal with access constraints.…”

“…While, the works on random access, such as access assignment by Markov chain modeling [7,24,29], estimation and fault detection with the access status described as a Bernoulli process [13,14], provided an alternative methodology to deal with access constraints. Other related works include the stabilizability and detectability [9,24], estimability [10] and observability [26], linear quadratic Gaussian (LQG) and robust control [29,12]. To the authors' knowledge, the observability of stochastic systems under access constraints is still open.…”