On optimal anchor placement for area-based localisation in wireless sensor networks

We consider the problem of optimal anchor placement for area-based localisation algorithms with the goal of providing cost-effective, simple, and robust positioning in wireless sensor networks. Due to the high complexity of the problem, we propose two placement algorithms based on heuristics. The first, called genetic algorithm anchors placement (GAAP), is based on genetic algorithms meta-heuristic, and the second, called local search anchors placement (LSAP), is based on an intuitive heuristic inspired from search techniques used in quad-trees. For the evaluation of these algorithms, we built a simulation framework, which we made publicly available for the community, and compared their performance against a Brute force (BF) algorithm, and against RND, a random walk-inspired algorithm. Obtained results show that GAAP provides anchor placements that lead to a very high accuracy while keeping execution time drastically smaller compared to LSAP, BF, and RND.