Probabilistic 5G Indoor Positioning Proof of Concept with Outlier Rejection
10 February 2022
The continuously increasing bandwidth and antenna aperture available in wireless networks has posed the ground for last years' research and development of competitive positioning solutions relying on communications standards and hardware. However, the challenges brought by non-line of sight (NLOS) scenarios are still posing many questions to the research community, due to the outlier measurements generated and their drastic impact on performance. In this work, we present an iterative method to weight each time difference of arrival (TDoA) and angle of arrival (AoA) measurement by a set of locators, allowing to efficiently removing outliers from the equation. Differently from prior art, we do not have the drawback of relying on a single reference locator for TDoA measures, preventing the catastrophic performance in case of outlier measurement at that locator. This allows to efficiently initialize the gradient-based search for estimating the position with maximum likelihood techniques. Our proposal is validated with an experimental setup at 3.75 GHz with 5G numerology in a realistic indoor factory scenario, allowing to achieve less than 50 cm error in 95% of the measures. To the best of our knowledge, this paper describes the first proof of concept for 5G-based joint TDoA and AoA localization.
