Performance Analysis of CGS, a k-Coverage Algorithm Based on One-Hop Neighboring Knowledge

Abstract

Abstract-Controlled Greedy Sleep (CGS) is a recent kcoverage algorithm based on one-hop neighboring knowledge. This paper presents the Controlled Greedy Sleep algorithm and highlights its best use cases thanks to performance analysis. Studies have been performed within a ring topology. This speci£c topology enables CGS validation and the design of an energetic model. They are used to compare CGS performance results with the maximal performances that could be expected. The ring topology allows us to determine analytically the minimum, maximum and relative values of our performances criteria which are: the quality, the cost and the duration of the k-coverage. According to the previous criteria we show in this paper that CGS is before anything designed to guarantee the maximum quality of coverage. We prove the CGS robustness when there are message losses. We show that CGS relative performances are improved when the overdosing and the communication cost increase. This is mostly because sensors just need to send at most three CGS messages per period and only need to know their one-hop neighbors.

Deployment of a mobile wireless sensor network with k-coverage constraint: a cellular learning automata approach

“…Notwithstanding the centralized approaches [38][39][40][41][42][43] in providing k-coverage in wireless sensor networks, most of the researches in this area are based on some sort of sleep/ wake scheduling of sensor nodes [44][45][46][47][48][49][50][51][52][53][54][55]. Slijepcevic et al in [44] introduced the ''set k-cover'' problem, which is to organize mutually exclusive sensor nodes into a number of covers or sets each of which can fully cover the network area.…”

“…[49] Proposes CGS algorithm, which makes use of the notion of bipartite graphs for selecting a subset of sensor nodes as active nodes so that every region of the network area is covered by at least k active sensor node. Performance of the CGS algorithm is analyzed in [50]. [51] Proposes two greedy algorithms (one centralized and one distributed) for selecting a subset of sensors from the set of deployed sensors so that the selected subset k-cover the entire area of the network.…”