Performance Enhancement of Wi-Fi Fingerprinting-based Indoor Positioning using Truncated Singular Value Decomposition and LSTM Model

Wi-Fi based indoor positioning has been considered as the most promising approach for civil location-based service due to the widespread availability Wi-Fi systems in many buildings. One of the most favorable approaches is to employ received signal strength indicator (RSSI) of Wi-Fi access points as the signals for estimating the mobile object locations. However, developing a solution to obtain high positioning accuracy while reducing system complexity using traditional methods as well as deep learning based methods is still a very challenging task. This paper presents a proposal to combine the Truncated Singular Value Decomposition (SVD) technique with a Long Short -Term Memory (LSTM) model to enhance the performance of indoor positioning system. Experimental results on a public dataset demonstrate that the proposed approach outperforms other state-of-the-art solutions by means of positioning accuracy as well as computational cost.

@article{Nguyen2024,
  title     = {Performance Enhancement of Wi-Fi Fingerprinting-based Indoor Positioning using Truncated Singular Value Decomposition and LSTM Model},
  journal   = {International Journal of Advanced Computer Science and Applications},
  volume    = {15},
  number    = {5},
  year      = {2024},
  publisher = {The Science and Information Organization},
  author    = {Duc Khoi Nguyen and Thi Hang Duong and Le Cuong Nguyen and Manh Kha Hoang},
  doi       = {10.14569/IJACSA.2024.0150529},
  url       = {https://doi.org/10.14569/IJACSA.2024.0150529}
}