A percolation algorithm for directional sensor networks

Abstract

In many applications of directional sensor networks especially border/barrier surveillance and intrusion detection applications, checking the establishment of the spanning clump/cluster of collaborated sensors that spans all over the region of interest is an essential task. Two sensors are collaborating sensors if their sensing ranges have been overlapped. When the spanning clump of collaborating sensors occurs, nothing could pass the region unless it will be detected by the network. In static networks (when orientation of sensor nodes is fixed), one time checking is needed. But in dynamic networks (when sensor nodes can adjust their orientations by using an algorithm) or hybrid networks (when some nodes are static and other nodes are dynamic), checking must be done periodically upon network topology changes. Therefore, it is very important to have an algorithm to test the establishment of spanning clump. In this paper, we present a distributed algorithm based on cellular automata to find out the establishment of spanning clump (percolation) in the region of interest by directional sensor networks consist of sensors with field-of-view angles between and π. The proposed algorithm is distributed and works locally which makes it suitable for using in sensor networks. Moreover, the algorithm could be in other technical problems of sensor networks such as scheduling and topology control.