Performance Analysis of Localization Techniques with Generalized Prior Distributions

Low error of indoor localization of wireless transmitters is difficult to achieve with only the received signal strength. This work includes an investigation of the probabilistic method for estimating the position of an unknown node in an indoor environment, Improved Prior Measurement Comparison (PMC+). PMC+ relies on a prior distribution for the unknown node position that emphasizes the probability of a node being in a specific room using a comparison to prior measurements made in that room. We expand the method of calculating this distribution to include arbitrary room shape and sensor placement. This method is then applied to a simulation and measurement campaign in an actual building at the University of New Hampshire, and the results of PMC+ are compared to other localization techniques. PMC+ significantly increases the likelihood of estimating the correct room of the unknown node. Copyright © 2015 IFSA Publishing, S. L.