Data fusion methods for accuracy improvement in wireless location systems

Abstract

Wireless location methods in wireless communications networks have received significant attention in recent years. Among the network-based approaches, time of arrival (TOA) and time difference of arrival (TDOA) are two major time-based techniques used for location estimation. It is known that the accuracy of location estimation may suffer from poor geometric dilution of precision (GDOP) situation and non-line of sight (NLOS) propagation effect. To improve the accuracy of time-based location techniques in wireless communications networks, data fusion methods using fuzzy logic and Bayes rules are presented. Based on the statistical properties of the raw location estimates, data fusion methods integrate the data obtained from the TOA and TDOA techniques. Computer simulations are conducted to investigate the wireless location problem affected by GDOP and NLOS. Simulation results show that the overall accuracy of location estimation in wireless location systems can be significantly improved when the LOS reconstruction and the data fusion methods are used.