Sequential Monte Carlo Methods for Localisation in Wireless Networks
Wireless indoor and outdoor localisation systems received a great deal of attention in recent years. This paper surveys first the current state-of-the-art of localisation techniques. Next, it formulates the problem of localisation within Bayesian framework and presents sequential Monte Carlo methods for localisation based on Received Signal Strength Indicators (RSSIs). Multiple model particle filters are developed and their performance is evaluated based on RSSIs by accounting for and without considering the measurement noise time correlation. A Gibbs sampling algorithm is presented for estimating the unknown parameters of the measurement noise which highly increases the accuracy of the localisation process.