Obstacle Constrained Total Area Coverage in Wireless Sensor Networks

This paper deals with the accomplishment of total area coverage of an arbitrary region using sensors with a finite sensing radius of rs. For a given region, we aim to obtain a deterministic placement of sensors which, apart from ensuring that the entire region comes under the purview of at least a single sensor, minimises the number of sensors utilised. We begin by considering regions devoid of obstacles and thus having every location amenable for placement. Herein, we formalise the popular notion that sensors at the centres of the hexagons of a hexagonal tessellation provide the most optimal placement. We then move on to regions which may comprise obstacles of arbitrary size at arbitrary locations. We recognise two distinct classes of obstacles, namely transparent and opaque obstacles, which are distinguished by their ability (or the lack of it) to permit sensing radiation through them. In the real world, transparent obstacles model lakes, ponds and swamps, while the opaque ones stand for, inter alia, hills, trees and walls. We propose a polynomial-time algorithm for achieving optimal placement in the aforesaid scenarios and we prove its convergence.