GNSS Pseudorange Based Multisensor Positioning Incorporating a Multipath Error Model

Abstract

Pedestrian navigation in challenging areas where Global Navigation Satellite Systems (GNSSs) lack the required accuracy and availability such as indoor and urban canyons is becoming increasingly important. Multisensor navigation is one of the techniques that has shown promising positioning accuracy in such environments. Being able to appropriately fuse GNSS pseudoranges in a multisensor positioning system is advantageous as it allows to make use of fewer than four detected satellites signals and to incorporate more accurate estimation models. In this paper a Particle Filter based multisensor pedestrian positioning system is enhanced to use pseudorange measurements of a GNSS receiver instead of direct position solutions. Using correctly modeled pseudoranges, a single visible satellite over a certain time period might improve positioning accuracy and robustness when combined with measurements from other positioning sensors like a foot mounted inertial measurement unit. Additionally in this paper, a statistical multipath error model for pseudoranges is integrated. Promising results have been achieved in a challenging indoor environment where blockage of satellites' signals and multipath are common issues.