Wi-Fi Network-Based Fingerprinting Algorithm for Localization in Coal Mine Tunnel

Abstract

Precise localization and target tracking in coal mine tunnel is crucial for miners’ safety protection. Due to the special environment of coal mine tunnel, the conventional positioning systems can not determine the specific location of underground personnel in real-time and with high positioning accuracy. In this paper, an improved fingerprinting algorithm based on underground Wi-Fi network is proposed to increase positioning accuracy. In our localization scheme, Received Signal Strength Indication (RSSI) from wireless Access Point (AP) and Support Vector Machine (SVM) based classifier are employed for position analysis. Specifically, the outliers were excluded by data preprocessing using k nearest neighbor (kNN) rule in the training phase, and results correction was utilized in the positioning stage. The positioning performance in coal mine tunnel environment demonstrates that the proposed improved fingerprinting algorithm can improve the positioning accuracy and the location of the miner can be computed in less time compared with traditional approach.