Particle filter and clustering based robust positioning algorithm in the WiFi weak signal environment

Abstract

To estimate a position in the indoor environment, WiFi positioning system (WPS) is useful technology in view of positioning performance, cost, availability. Especially, particle filter (PF) present a stable performance by integrating WiFi signal and sensor output such as accelerometer and gyroscope. But the initial positioning error remains a problem. In this paper, we propose the pattern matching algorithm and clustering approach to correct the initial positioning error. To make a weak WiFi environment, we utilize only 20% installed access point (AP). We identify that the proposed algorithm makes a stable positioning performance than the conventional PF.