A Research on Positioning Algorithm Based on RPCA in Sparse Fingerprint Environment

Yaqin Xie (Nanjing University of Information Science and Technology)
Md Emadur Rahman Ekra (Nanjing University of Information Science and Technology)
Tianyuan Gu (Nanjing University of Information Science and Technology)
Xiaoli Wang (Nanjing University of Information Science and Technology)

International Journal of Informatics, Information System and Computer Engineering (INJIISCOM)
Vol. 6 No. 2 (2025): INJIISCOM: VOLUME 6, ISSUE 2, DECEMBER 2025

Wi-Fi fingerprinting for indoor positioning is cost-effective but struggles with environmental noise and requires extensive data collection for high accuracy. To address these challenges in sparse fingerprint environments, this paper proposes a positioning technique using a Robust Principal Component Analysis (RPCA) algorithm. First, gathered signals are purified using measurement weights to mitigate outlier noise, saving the refined fingerprints in a database. Second, to reduce collection costs, virtual fingerprints are generated near reference points using a transmission loss model and stored offline. Finally, adaptive K-value fingerprint matching is applied to estimate the user's location. Results demonstrate that the proposed RPCA-based algorithm significantly improves positioning accuracy in sparse indoor environments.