Conventional system to deep learning based indoor positioning system

Abstract

This review article presents the key fundamentals of Indoor Positioning System (IPS) and its progressing footprints. The need of IPS and work done with methodology adopted to implement IPS for various applications have been discussed. The evolution from conventional to Deep Learning (DL) has been presented, addressing various challenges existing in conventional IPS like poor localization, improper accuracy, non-line-of-sight problems, instability of signal due to fading, requirements of large infrastructure, data-set and labour, high cost, and their existing solutions have been disclosed. Further in order to compute the indoor positioning with acute precision various advanced positioning technologies including sensor fusion, Artificial Intelligence (AI), and hybrid technologies have been explored. The issues and challenges existing in current IPS technology have been presented and future insights to work in this direction have also been provided.