Hybrid FCM/ANN indoor location method in WLAN environment

Abstract

WLAN hybrid indoor location method based on fuzzy c-mean clustering (FCM) and artificial neural network (ANN) is proposed in this paper. For any pattern matching based algorithm in WLAN environment, characteristics of received signal strength (RSS) or signal to noise ratio (SNR) to multiple access points (APs) are utilized to establish radio map in the off-line phase, and in the on-line phase, actual two or three dimensional coordinates of the mobile terminals (MTs) are estimated based on the comparison between the new recorded RSS or SNR and fingerprints stored in radio map. Although the feed-forward ANN with three layers is sufficient to approximate any continuous functions to a desired accuracy and optimize any mapping relationship between training inputs and targets, generalization ability is difficult to be guaranteed. So in order to bridge this gap, FCM method is proposed to select the reference points (RPs) affected by the multi-path effect. Based on this method, RSS or SNR recorded at these RPs is modified by the linear regression. Feasibility and effectiveness of this hybrid FCM and ANN method are verified according to the experimental comparison with ANN method without FCM modification.