Enhanced Integrated Indoor Positioning Algorithm Utilising Wi-Fi Fingerprint Technique

Abstract

This paper describes an integrated positioning algorithm utilizing Wi-Fi fingerprint technique for indoor positioning. The main contribution of this work is the improvement of positioning accuracy for indoor localization even in extreme RSSI fluctuation which leads to variation of positioning error. Several layers of Wi-Fi positioning is proposed, which are based on deterministic techniques, iterative Bayesian estimation, and also Kalman filter to enhance accuracy due to noise presence. Here, accumulated accuracy is introduced where the distribution of location error is determined by estimation at each test point on path movement. The results show that the integrated algorithm enhances the estimation accuracy in several scenarios which are different Wi-Fi chipsets and movement directions. The error distribution shows an achievement of up to 65% for error less than 5m compared to the basic deterministic technique of only 45%.