Extended Kalman Filter combined with fuzzy rules for localization using wireless transceivers

Estimating position of moving objects have been an area of research for several decades, but challenges still remain as many of technologies or computational methods are either too costly, computationally intensive, or simply not possible to apply due to environmental, economical, or other types of constraints. In this paper we investigate a novel method for improving position estimation of a moving object using fuzzy rules in combination with Extended Kalman Filter (EKF) and Received Signal Strength Indicator (RSSI) measures. The EKF provides a recursive method for estimating internal states of nonlinear system from measured observations. The estimation performance of EKF is highly dependent on the dynamics of internal system model, which is not always available, as could be the case for the hidden locations of humans, robots, and animals moving according to their own rules. Preliminary results have shown evidence that when fuzzy rules are considered to represent the dynamical system, a reduction of error occurs, and as a result, a less conservative estimation of position is obtained when compared to the traditional weighted least-square estimate in the context of recursive approach.