Fault-Tolerant and Constrained Relay Node Placement in Wireless Sensor Networks

In this paper we focus on wireless sensor networks deployed to cover some given Points of Interest (Pots), achieve connectivity with the sink and be robust against link and node failures. The Relay Node Placement problem (RNP) consists in minimizing the number of relays needed and the maximum length of the paths connecting each PoI with the sink. We propose a solution that determines the positions of relay nodes based on the virtual grid computed by the optimal deployment for full area coverage. We compare our solution with two different solutions based respectively on 1) the straight line that builds the shortest path between each PoI and the sink, 2) the Steiner point that connects Pots together. We then extend these algorithms to achieve k-connectivity. Our solution outperforms the Steiner points solution in terms of maximum path length on the one hand, and the straight line solution in terms of total number of relay nodes deployed on the other hand. We also apply our solution in an area containing obstacles and show that it provides very good performances.