Abstract
In this thesis, we consider the problem of developing a Gauss-Markov parameter estimator in wireless sensor network. To this purpose, we propose an on-line method (GMPE) to estimate Gauss-Markov parameters with little requirement in terms of communication overhead of messages. We compared by simulations the performance of GMPE prediction model with the instant and the exponential average prediction models. Simulations showed that GMPE prediction model achieves more accurate prediction on the location of the object in Gauss-Markov mobility model as compared to the instant and the exponential average prediction models.