Decentralized event-based scheduling for shared-resource Networked Control Systems

Abstract

Abstract-In this paper, we introduce a decentralized eventtriggered scheduling scheme for multi-loop Networked Control Systems (NCSs) in which the individual control loops are coupled through a shared communication channel. The proposed scheduling design combines deterministic and probabilistic attributes to efficiently allocate the limited communication resource among the control loops in an event-based fashion. Based on local error thresholds, each control loop determines whether to compete for the channel access. As a result, control loops with higher transmission priorities are more likely to utilize the channel, which in turn leads to more efficient usage of the limited resource. Each eligible subsystem then attempts to transmit at times specified by local waiting times, which are randomly distributed and local-error dependent. In this manner, the probability of data collisions in the communication channel is reduced. If the channel capacity is reached at some time instants, data packets are dropped. We demonstrate stochastic stability of such NCSs in terms of Lyapunov Stability in Probability (LSP). The numerical results show that our approach improves resource utilization and reduces the networked-induced error variance in comparison with timetriggered, random access, and centralized scheduling policies.

Optimal Scheduling of Downlink Communication for a Multi-Agent System With a Central Observation Post

“…Moreover, in a fully distributed setting, the proposed approach can be used in combination with the approach of [4]: the offline deployment of our framework can give indications on how to select the tuning parameters of that algorithm, i.e. the quadratic cost the relative threshold.…”

Decentralized event-triggered medium access control for networked control systems

“…It should however be noticed that having N transmission chances does not at all guarantee that all N sub-systems transmit. In what follows, we will show that (15) can be guaranteed over an operational interval of length N , i.e. starting from time-step k, we take the interval [k, k+N ].…”

Event-based data scheduling for a class of interconnected networked control systems

“…We assume that the scheduler is updated with error norms e i k 2 Qi ∈ R from all nodes i at every time k. This can be realized by considering a low-capacity channel between the scheduler and sub-systems to send those N real numbers to the scheduler. The randomization in the last line of (8) allows to implement it approximately in decentralized fashion [13].…”

“…Each node is characterized by only one of the above cases during transition from k to k+N −1. We apply Theorem 1 to the cases c 1 − c 3 employing the drift (13) and Lyapunov function (12). Remember that we consider local error of each node.…”

“…Alternatively, probabilistic policies have been investigated that consider stochastic NCSs [9], [10]. These methods are better equipped to deal with noise and collisions but also with model uncertainties [11]- [13]. The aforementioned results deal with NCSs with isolated subsystems, i.e.…”