The wireless localisation matching problem and a maximum likelihood based solution

We propose a new approach towards wireless localisation related to scenarios where the device positions are known a priori, however the device IDs are not, and therefore need to be matched using RF methods. We propose a maximum-likelihood based matching algorithm called MLMatch for resolving this problem based on measured RSSI values. Since the search space of node-to-positions permutations grows factorially with the number of target devices, an Integer Linear Programming formulation is therefore utilised to reduce computation time. In addition, we analyse the stability of the algorithm with respect to different fading models and other wireless propagation parameters such as pathloss. Finally, we report on experiments performed indoors and outdoors using up to 33 wireless devices in order to validate our results.