An improved connectivity-based boundary detection algorithm in wireless sensor networks

This paper tackles the problem of boundary detection by proposing a simple, distributed and connectivity-based algorithm. Our algorithm examines the 2-hop iso-contour of each node, and outperforms existing algorithms examining iso-contours. Specifically, the proposed algorithm makes a rough decision on a suspected boundary node by examining its 2-hop iso-contour, and then refines the decision based on a heuristic operation, which significantly reduces the size of the suspected boundary node set. More importantly, boundary cycles corresponding to inner and outer boundaries are identified and provide valuable knowledge to various applications. A thorough evaluation shows that our algorithm is applicable to dense as well as sparse WSNs.