Three-dimensional non-line-of-sight localisation in an indoor multipath environment

This paper presents the design of a non-line-of-sight (NLOS) localization algorithm that determines the position of a mobile device in a typical multipath environment where the scatterers are aligned parallel or perpendicular to each other. Our proposed algorithm leverages not only on the line-of-sight paths but also NLOS paths which undergo single bounce reflection. Furthermore, it adopts a unidirectional estimation approach in which the time of arrival (TOA) and angle of arrival (AOA) of signals are measured at the reference devices. By resolving the error variance due to noisy measurement into three mutually perpendicular components and applying appropriate weighting factors for each of these components, location error is greatly reduced. Simulation results in a multipath environment indicate that location accuracy is improved by about 42 percent compared to the existing NLOS algorithm even at large distance and angle measurement noises.