APs Deployment Optimization for Indoor Fingerprint Positioning with Adaptive Particle Swarm Algorithm

Indoor positioning service gives people much better convenience, but its efficiency is affected by the spatial deployment of access points, APs. We propose an algorithm from adaptive particle swarm, APS, and then apply it in APs deployment optimization for fingerprint based indoor positioning. In our method, solutions of APs placement are taken as individuals of one population. Particle swarm method is improved with adaptive technology to ensure the population diversity and also avoid large number of inferior particles. After evolutions, the optimal result is obtained, corresponding to the best solution of APs deployment. The algorithm works well for both single-objective and multi-objective optimizations. Experiments with deployments of 107 iBeacons have been tested in an underground parking lot. Compared with the existing APs placement methods, our APS algorithm can obtain the least indoor positioning error with fixed APs number, while receive the best integrated evaluation considering both positioning error and APs cost with unfixed APs number. The proposed algorithm is easily popularized to the other kinds of indoor spaces and different types of signal sources.