Key points are not available for this paper at this time.
Conventional localization methods like GNSS face limitations in accuracy, power consumption, and cost within indoor environments. Bluetooth Low Energy (BLE) technology and the fingerprint method have emerged as popular options for short-range networks. Although fingerprint-based localization offers accurate position estimation, it suffers from drawbacks such as the requirement for a complete map and frequent map updates. To address the issue of incomplete signal maps, this study proposes a novel approach that incorporates fuzzy clustering to generate candidate locations and weighted interpolation to estimate the final location by considering the impact of walls on the environment's signal. A Neural Network (NN) is employed to approximate the object's location relative to walls, providing precise weights for candidate positions and enhancing estimation accuracy. Evaluation results obtained from an office environment demonstrate a remarkable 15% improvement in positioning accuracy, affirming the potential of our approach for real-world industry-driven applications.
Pasandi et al. (Fri,) studied this question.