Smart spatio-temporal fingerprinting for cooperative ANN-based wireless localization in underground narrow-vein mines

One of the main concerns in the mining industry is ensuring the safety and security of miners and their equipment. Being aware of the real-time position of personnel in such harsh environments within a special quasi-curvilinear topology is challenging and requires a sophisticated localization system. While traditional triangulation techniques fail to accurately localize in such indoor scenarios, new approaches that rely on fingerprints extracted from the Channel Impulse Response (CIR) succeed to localize with high accuracy using Artificial Neural Networks (ANNs) for fingerprint-location matching. Signatures collected from different locations in space, at different instances in time, are concatenated to form spatio-temporal fingerprints for improved localization accuracy. In this paper, we overview these novel and very promising localization techniques then investigate the impact of the spatial sampling grid's resolution in fingerprint collection on their accuracy in underground narrow-vein mines. We show by simulations that the significant accuracy gains reaped from the new exploitation of spatio-temporal diversity, if not needed in some applications, can be alternatively traded for remarkable and extremely useful cost reductions in the fingerprint collection step.