Neural Network-Based Accuracy Enhancement Method for WLAN Indoor Positioning

Abstract

As the need for location-based services (LBS) in indoor environments increases, high accuracy positioning technologies are required, which makes fingerprinting-based positioning methods using wireless local area network (WLAN) develop from single fingerprinting algorithm into multi-algorithm integration. A neural network-based accuracy enhancement (NNAE) method for indoor positioning using WLAN is proposed in this paper. The method takes the advantages of the fingerprinting algorithms based on pattern matching and distance dependence. It uses a neural network-based pattern matching algorithm to estimate the positioning errors and then the estimated positioning errors are used to correct the positioning results calculated by a distance dependent algorithm. The experimental results show that the proposed NNAE method outperforms classical fingerprinting algorithms and effectively enhances the positioning accuracy.