Indoor localization is now an important part of intelligent environments, including smart buildings, retail spaces, hospitals, and industrial spaces. Bluetooth Low Energy (BLE) has become a potential technology in indoor positioning with low power consumption and ubiquitous nature. Nevertheless, traditional RSSI-based BLE localization schemes, such as fingerprinting and trilateration, usually experience multipath fading, environmental non-uniformity, calibration cost, and complexity of computation. The current paper describes the experimental validation of the design of a zone-based, calibration-free BLE indoor localization system with complete implementation on a low-power ESP32. Rather than calculating exact coordinates, the proposed method does the similar task of classifying zones proximity-wise by the use of lightweight sliding-window RSSI filtering and hysteresis-based zone stabilization. Signal processing and localization decisions are run directly on the embedded device without cloud dependency. Experimental analysis in a real indoor setting illustrates that the accuracy of zone detection of RSSI filtering increases on average, between 74% unfiltered and 91% filtered, at low computational complexity and infrastructure. The proposed architecture offers a scalable and efficient resource-based solution that can be used in real-life usage of proximity-sensitive types of applications.
S et al. (Thu,) studied this question.