34
14

Persistent Monitoring of Dynamically Changing Environments Using an Unmanned Vehicle

Abstract

We consider the problem of planning a closed walk W\mathcal W for a UAV to persistently monitor a finite number of stationary targets with equal priorities and dynamically changing properties. A UAV must physically visit the targets in order to monitor them and collect information therein. The frequency of monitoring any given target is specified by a target revisit time, i.e.i.e., the maximum allowable time between any two successive visits to the target. The problem considered in this paper is the following: Given nn targets and knk \geq n allowed visits to them, find an optimal closed walk W(k)\mathcal W^*(k) so that every target is visited at least once and the maximum revisit time over all the targets, R(W(k))\mathcal R(\mathcal W(k)), is minimized. We prove the following: If kn2nk \geq n^2-n, R(W(k))\mathcal R(\mathcal W^*(k)) (or simply, R(k)\mathcal R^*(k)) takes only two values: R(n)\mathcal R^*(n) when kk is an integral multiple of nn, and R(n+1)\mathcal R^*(n+1) otherwise. This result suggests significant computational savings - one only needs to determine W(n)\mathcal W^*(n) and W(n+1)\mathcal W^*(n+1) to construct an optimal solution W(k)\mathcal W^*(k). We provide MILP formulations for computing W(n)\mathcal W^*(n) and W(n+1)\mathcal W^*(n+1). Furthermore, for {\it any} given kk, we prove that R(k)R(k+n)\mathcal R^*(k) \geq \mathcal R^*(k+n).

View on arXiv
Comments on this paper