Finite-Time Distributed State Estimation over Time-Varying Graphs: Exploiting the Age-of-Information

Abstract

We study the problem of designing a distributed observer for an LTI system over a time-varying communication graph. The limited existing work on this topic imposes various restrictions either on the observation model or on the sequence of communication graphs. In contrast, we propose a single-timescale distributed observer that works under mild assumptions. Specifically, our communication model only requires strongconnectivity to be preserved over non-overlapping, contiguous intervals that are even allowed to grow unbounded over time. We show that under suitable conditions that bound the growth of such intervals, joint observability is sufficient to track the state of any discrete-time LTI system exponentially fast, at any desired rate. In fact, we also establish finite-time convergence based on our approach. Finally, we develop a variant of our algorithm that is provably robust to worst-case adversarial attacks, provided the sequence of graphs is sufficiently connected over time. The key to our approach is the notion of a "freshness-index" that keeps track of the age-of-information being diffused across the network. Such indices enable nodes to reject stale estimates of the state, and, in turn, contribute to stability of the error dynamics.1 Such approaches, referred to as two-time-scale approaches, incur high communication cost, and might hence be prohibitive for real-time applications. arXiv:2001.07006v1 [eess.SY]

A simple finite-time distributed observer design for linear time-invariant systems

“…This will happen eventually in a finite time, if the observers for this part have been designed according to Section 3.1. However, until then it has to be ensured that the non-vanishing E o does not destabilize (9). For this, we can consider E o as an external input (since its dynamics is independent of E u ) and show that (9) is input-to-stable [19], which will also directly imply its global asymptotic stability for E o = 0.…”

“…However, it is on the expense of a more complex design compared to the observability decomposition with a scalar gain used in this work. There are other methods which allow selecting the poles in the discrete time-setting and therefore to accomplish finite-time convergence [9] (akin to a dead-beat observer in the centralized setting). Pole placement for continuous-time systems does also open up the possibility to design finite-time distributed observers using generalized homogeneity concepts [10], nonetheless the approach of the present paper using just standard observability decompositions offers a design which is comparatively simple.…”

Remote Short Blocklength Process Monitoring: Trade-off Between Resolution and Data Freshness

“…In order to obtain the MSE as described in (25), we need the expected value of the term L(r, s, r ′ ) above. This can be derived using (17) as follows:…”

Design of a distributed finite-time observer using observability decompositions

“…Additionally, in a mobile environment the communication topology can be changed at some timeinstants, making it more important that time-constraints are met. Time-varying communication graphs and finite-time convergence have also been examined in [16], however for a discrete-time system akin to a dead-beat observer.…”