Cooperative geo-location in underground mines: A novel fingerprint positioning technique exploiting spatio-temporal diversity

Abstract

Underground narrow-vein mines result in complex indoor scenarios which require sophisticated localization techniques to maintain basic security measures. While some traditional localization systems use the triangulation techniques for outdoor channels, fingerprint positioning techniques are mostly used in more complex indoor environments like mines. One of the techniques exploited in the quasi-curvilinear topology of underground mines is the Channel Impulse Response (CIR) based fingerprint positioning combined with Artificial Neural Networks (ANNs). This article innovates a CIR-based positioning technique within a cooperative memory-assisted approach that exploits both the temporal (from different time instances) and spatial (from different space positions) diversities of the collected fingerprints. Introducing memory-type signatures in a cooperative localization technique within the spatial confinements of the tunnel-shaped narrow-vein mines significantly increases the accuracy, precision and robustness of the localization system. The cooperative memory-assisted technique is capable of localizing a transmitter with an accuracy of less than 25 cm 90% of the time.