A feasibility study of responsive indoor localisation system - A case study in CLP

Abstract

Internet of Things (IoT) is a hot topic under smart city that leads to the rise of Low-Power Wide-Area Network (LPWAN) technologies. The implementation of large scale sensor networks becomes easier after applying such technologies, especially for localization. In this project, localization, as a primary function in different automated systems, is the major focus. Of all wireless communication traits, received strength signal (RSS) fingerprint is examined to be the best practice for indoor localization algorithm. However, there are several glaring problems of the traditional RSS fingerprint approaches, which include low accuracy caused by time-variant fingerprinting, long latency due to high computing complexity and small coverage of the selected wireless protocols. Hence, to overcome these challenges, the methodology of indoor localization is being studied. Various kinds of LPWAN protocols are being reviewed and LoRa is decided to be the protocol building the sensor network for indoor localization. In addition, a novel low complexity RSSI coefficient-based localization algorithm is developed, with a test bed established in the office of China Light and Power(CLP). The experimental results showed that the proposed algorithm can achieve an accuracy of 97% with an acceptable error set to be 10 meters from the positioned object.