Human motion modelling for simulation testing of GNSS equipment

• Main Author: Voutsis, Kimon
• Other Authors: Groves, Paul
• Published: University College London (University of London) 2017
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.746758

Pedestrian motion-induced dynamics along the line-of-sight (LOS) between a GNSS receiver and a satellite, may disrupt the nominal operation of GNSS carrier-tracking loops, by introducing cycle slips and/or false frequency locks. In combination with other factors, e.g. multipath interference, weak signal conditions or limited availability of GNSS signals, the receiver could provide a degraded navigation solution or even lose signal lock. This thesis researches firstly how pedestrian motion affects the operation of carrier phase lock loops (PLLs), used by some GNSS receivers, and frequency lock loops (FLLs), used by all GNSS receivers; and secondly, what is the best way to model pedestrian motion in order to simulate the error effects of pedestrian motion-induced dynamics on a GNSS antenna, via a simulated GNSS carrier phase lock loop (PLL). The thesis reviews the relevant literature on human biomechanical modelling, path-finding and inertial/GNSS navigation, to design a custom pedestrian motion model (PMM). The PMM validation is supported by motion capture (MoCap) experiments using an inertial/GNSS sensor held by, or attached, on a pedestrian. The thesis also describes an implementation of simulated GNSS carrier-tracking loops (SGCTLs) in Matlab, to assess the effect of human MoCap profiles and synthetic human motion profiles (from the PMM) on the performance of the SGCTLs. The testing results suggest that GNSS antenna motion dynamics due to typical pedestrian motion can induce excessive cycle slips due to dynamics stress on the simulated PLL and FLL. Therefore, antenna dynamics should be considered when designing GNSS tracking loops and navigation algorithms for pedestrian applications to allow the GNSS receiver track human motion-induced dynamics effectively. The thesis concludes with carrier-tracking bandwidth recommendations for GNSS receiver design, based on the presented evidence. Under good signal conditions (above 40dB-Hz), the minimum recommended bandwidths for PLLs and FLLs are 15Hz and 5Hz, respectively, in order to respond effectively to the dynamic stress induced by typical pedestrian movements. Finally, the results indicate that the PMM can represent the LOS dynamics stress on the SGPLL within an acceptable tolerance. Future work encompasses the analysis of the pedestrian motion effects on real GNSS receivers.