Iteration improvement of Taylor-series estimation using hyperbolic systems for FM-radio source localization in Bangkok

Abstract

Taylor-series estimation is considered a useful technique for estimating the location of a source based on intersections of hyperbolic curves of the time differences of arrival of signals received by several sensors. One of the common challenges in estimating the location of a source emitter is to select a starting point in each iterative calculation. Different starting points can significantly impact search performance and the accuracy of the estimated signal emitter location. This study proposes an initial position for solving Taylor-series expanded nonlinear equations. Our recommended initial position choices (approximately graphical mode or AGM) use a simple approximately hyperbolic estimator to calculate the starting point in each iterative search for the source location. Our study shows that AGM estimator can help improve the calculation speed and produce more accurate results. We compared our method with the other two common choices, first Rx mode and center of sensors mode (CPM). The results show that our AGM method performs up to 4.64% more accuracy.