Fusion of GNSS Pseudoranges with UWB Ranges Based on Clustering and Weighted Least Squares

Abstract

Global navigation satellite systems (GNSSs) and ultra-wideband (UWB) ranging are two central research topics in the field of {positioning} and navigation. In this study, a GNSS/UWB {fusion} method {is investigated} in {GNSS-challenged} environments or for the transition between out{door and} indoor {environments}. UWB {augments} the GNSS positioning solution {in these environments}. GNSS stop-and-go measurements were carried out simultaneously to UWB range observations within the network of grid points used for testing. The influence {of UWB} range measurements {on the GNSS solution is} examined {with three} weighted least squares (WLS) approaches. The first {WLS} variant {relies} solely on the UWB range measurements. The second approach includes a {measurement} model that {utilizes} GNSS {only}. The third model {fuses} both approaches into a single {multi-sensor} model. As part of the {raw data evaluation}, static GNSS observations processed with precise {ephemerides} were used to define the ground truth. In order to extract the grid test points from the collected raw data in the measured network, clustering methods were applied. A self-developed clustering approach extending density-based spatial clustering of applications with noise {(DBSCAN)} was employed for this purpose. The results of the GNSS/UWB fusion approach {show} an improvement in positioning {performance} compared to the UWB-only approach{,} in the range of {a} few centimeter{s} to the decimeter {level} {when} grid points {were} placed within the area enclosed by the UWB anchor points. However, grid points outside this area indicated a decrease {in} accuracy in the range of about 90~cm. The precision generally remained within 5~cm for points located within the anchor points.