Statistical Methods for Fast LOS Detection for Ranging and Localization

Received signal strength indication (RSSI) data is often used for ranging and localization algorithms, where the data may be obtained using Bluetooth Low Energy (BLE) radios. In the BLE protocol, when a device is in advertising mode, it is possible to obtain RSSI values. However, these RSSI values are noisy and can often fluctuate due to multipath effects, which reduces the accuracy and reliability of ranging and localization. Additionally, the effectiveness of RSSI ranging degrades when there is an absence of line of sight (LOS) or when devices are in rich scattering environments. Therefore, the detection of LOS plays a very significant role in indoor localization and room reconstruction. In this paper, we present algorithms to detect whether there is LOS present between a transmit-receive pair being used for ranging. Our focus in this paper is fast detection with a minimum number of samples. We use measurements such as the energy distance and Mahalanobis distance, and benchmark our results against the Neyman-Pearson detector. Numerical simulations are used to validate our algorithms.