Distributed Area Coverage by Connected Set Cover Partitioning in Wireless Sensor Networks

Abstract

Assuming a random uniform distribution of n sensor nodes over a virtual grid, this paper addresses the problem of finding the maximum number of connected set covers each ensuring 100% coverage of the query region. The connected sets remain active one after another in a round robin fashion such that if there are P such set covers, it can enhance the network lifetime P-fold. From graph-theoretic point of view, a centralized O(n3) heuristic is proposed here to maximize P. Next, for large self-organized sensor networks, a distributed algorithm is developed. The proposed algorithm is to be executed just once, during the initialization of the network. In case of failure, a distributed recovery algorithm is executed to rearrange the partitions. Simulation studies show that the performance of the proposed distributed algorithm is comparable with that of the centralized algorithm in terms of number of partitions. Also, comparison with earlier works shows significant improvement in terms of number of partitions, message complexity and network lifetime.