A Study On Filtering Algorithm For Wireless Localization

Wireless location technology uses information among base stations and mobile station, localization algorithms are applied as well. Besides, localization algorithms are key factors to wireless location performance. Non-line-of-sight errors contribute to the performance differences among localization algorithms. Thus this paper focuses on solving the non-line-of-sight errors in order to improve the performance of localization algorithms.As is known to all, Kalman filter can filter noise by observable information and given predicted information. Kalman filter not only filters noise but also has estimation function.Based on that, Kalman filter is proposed to be applied in localization algorithms to suppress non-line-of-sight errors. Then Kalman filter is used in line-of-sight environment to eliminate measurement noises. After that Kalman filter can be used in non-line-of-sight environment. However, since the difference between line-of-sight environment and non-line-of- sight environment, federal Kalman filter is utilized. Furthermore, smoother is used to improve federal Kalman filter. Simulations prove that the performance of localization algorithms can be improved by Kalman filter about 20.5 meters, 25.2 meters and 28.1 meters under Gauss distribution, Uniform distribution and Exponential distribution.