RSS-Based Localization and Mobility Evaluation Using a Single NB-IoT Cell

Low Power Wide Area Networks (LPWAN) have the ability to localize a mobile transmitter using signals of opportunity, as a low power and low cost alternative to satellite-based solutions. In this paper, we evaluate the accuracy of three localization approaches based on the Received Signal Strength (RSS). More specifically, the performance of a proximity, range-based and optimized fingerprint-based algorithm is evaluated in a large-scale urban environment using a public Narrowband Internet of Things (NB-IoT) network. The results show a mean location estimation error of 340, 320 and 204 m, respectively. During the measurement campaign, we discovered a mobility issue in NB-IoT. In contrast to other LPWAN and cellular technologies which use multiple gateways or cells to locate a device, only a single cell antenna can be used for RSS-based localization in NB-IoT. Therefore, we address this limitation in the current NB-IoT hardware and software by studying the mobility of the cellular-based 3GPP standard in a localization context. Experimental results show that the lack of handover support leads to increased cell reselection time and poor cell sector reliability, which in turn results in reduced localization performance.