Asymmetric Positioning for NLOS Mitigation

Abstract

Conventional GNSS positioning algorithms rely on the assumption that the pseudo-range error follows a normal distribution, which allows for the use of statistical techniques and probabilistic models to improve the accuracy and reliability of the positioning solution. However, this assumption does not always hold true in practice, especially in urban environments where blocking, reflections, and other factors can significantly impact the quality of the GNSS signals and lead to errors that do not conform to a normal distribution. In this paper, an efficient NLOS mitigation algorithm is proposed to improve positioning performance in cities. It allows conventional least-squares ranging (LSR) and extended Kalman filtering (EKF) to handle asymmetric distributions and to determine an appropriate distribution for each measurement based on its signal strength. This algorithm can be implemented on any GNSS receiver with only a small increase in processing load, and it does not require any additional information or hardware. The experiments were conducted in 13 different locations alongside busy roads in the London Borough of Camden, where two 3-minute sessions of static pedestrian navigation data were collected at each location using a u-blox ZED-F9P GNSS receiver, one for training and the other for testing. The experimental results confirm that the pseudo-range error of the NLOS signal does not conform to a normal distribution. Compared to the conventional approaches, the proposed method was able to reduce the RMS horizontal position error by about 21% and 34% in the single and multi-epoch cases, respectively. The performance of the proposed method was also compared to 3D-mapping-aided (3DMA) GNSS positioning.