On the observation quality of robot-based GNSS antenna calibration for determining codephase corrections

authored by: Yannick Breva
supervised by: Franz Rottensteiner
Abstract: Global Navigation Satellite System (GNSS) signals are received at the electric phase center of GNSS antennas, which varies based on the direction of the incoming signals. In highly precise GNSS applications, phase center corrections (PCC) for carrierphase signals must be taken into account. Additionally, codephase center corrections (CPC) for codephase signals exist, which become more important in aeronautic navigation or in code- and carrierphase linear combinations. Both, CPC and PCC, describe the difference between the actual receiving point of the signal at the antenna and the antenna reference point (ARP), which is the last point of the antenna that can be mechanically accessed for height measurements. The challenge of estimating multi- GNSS multi-frequency CPC with a robot in the field lies in the high observation noise relative to the pattern magnitude itself. This thesis focuses on reducing the observation noise of codephase signals while preserving the important pattern information within the observations during an absolute robot-based antenna calibration, in order to improve the repeatability and accuracy of the estimated CPC. A Monte-Carlo simulation is performed to study the impact of observation noise on the estimated CPC. The simulation shows that white noise and signal strength dependent noise with a standard deviation of the same magnitude as the pattern’s peak-to-peak value results in a degradation of 15% to 25% for various comparison metrics. The weighted average standard deviation of the actual observables ranges from 0.476m to 0.620m for the analysed GNSS antennas, which is nearly equivalent to the magnitude of the pattern’s peak-to-peak values. To reduce the noise, the receiver tracking loop parameters are adjusted based on an experiment using a software receiver during a calibration. The best performance is achieved with a delay lock loop (DLL) bandwidth of 0.5 Hz, using a loop filter order of 1 in a carrier aided DLL. This acquired knowledge is adapted to hardware receivers, significantly improving the weighted average standard deviation of the observables by 42% resulting in a more repeatable and accurate CPC compared to using the manufacturer’s default receiver settings. Additionally, multipath effects during calibration are thoroughly investigated, with the balustrade and the astronomical domes in the robot’s surrounding identified as the most significant sources of multipath effects. The studies demonstrate that using time-differenced observations in combination with a dynamic elevation mask and multipath maps can effectively eliminate almost all multipath affected observations. Furthermore, the time differenced multipath linear combination (ΔMPLC) is introduced as an input for the estimation approach, resulting in reduced observation noise compared to the time differenced receiver-to-receiver single differences (ΔSD) approach, as one differencing step is avoided. The estimated CPC, using ΔMPLC, with optimized receiver settings shows similar repeatability. Additionally, applying CPC in a single point positioning (SPP) leads to a 70% improvement in the estimated Up component compared to when no CPC is applied. In the observation domain, by calculating SD, the long-period trend can be reliably represented by the estimated CPC.
Organisation(s): Institute of Photogrammetry and GeoInformation (IPI)
Type: Doctoral thesis
No. of pages: 172
Publication date: 23.05.2025
Publication status: Published
Electronic version(s): https://doi.org/10.15488/19053 (Access: Open)
: Details in the research portal "Research@Leibniz University"

BibTeX

@phdthesis{2c32cc83e1ef40a08dc4abbd8842896b,
title = "On the observation quality of robot-based GNSS antenna calibration for determining codephase corrections",
abstract = "Global Navigation Satellite System (GNSS) signals are received at the electric phase center of GNSS antennas, which varies based on the direction of the incoming signals. In highly precise GNSS applications, phase center corrections (PCC) for carrierphase signals must be taken into account. Additionally, codephase center corrections (CPC) for codephase signals exist, which become more important in aeronautic navigation or in code- and carrierphase linear combinations. Both, CPC and PCC, describe the difference between the actual receiving point of the signal at the antenna and the antenna reference point (ARP), which is the last point of the antenna that can be mechanically accessed for height measurements. The challenge of estimating multi- GNSS multi-frequency CPC with a robot in the field lies in the high observation noise relative to the pattern magnitude itself. This thesis focuses on reducing the observation noise of codephase signals while preserving the important pattern information within the observations during an absolute robot-based antenna calibration, in order to improve the repeatability and accuracy of the estimated CPC. A Monte-Carlo simulation is performed to study the impact of observation noise on the estimated CPC. The simulation shows that white noise and signal strength dependent noise with a standard deviation of the same magnitude as the pattern{\textquoteright}s peak-to-peak value results in a degradation of 15% to 25% for various comparison metrics. The weighted average standard deviation of the actual observables ranges from 0.476m to 0.620m for the analysed GNSS antennas, which is nearly equivalent to the magnitude of the pattern{\textquoteright}s peak-to-peak values. To reduce the noise, the receiver tracking loop parameters are adjusted based on an experiment using a software receiver during a calibration. The best performance is achieved with a delay lock loop (DLL) bandwidth of 0.5 Hz, using a loop filter order of 1 in a carrier aided DLL. This acquired knowledge is adapted to hardware receivers, significantly improving the weighted average standard deviation of the observables by 42% resulting in a more repeatable and accurate CPC compared to using the manufacturer{\textquoteright}s default receiver settings. Additionally, multipath effects during calibration are thoroughly investigated, with the balustrade and the astronomical domes in the robot{\textquoteright}s surrounding identified as the most significant sources of multipath effects. The studies demonstrate that using time-differenced observations in combination with a dynamic elevation mask and multipath maps can effectively eliminate almost all multipath affected observations. Furthermore, the time differenced multipath linear combination (ΔMPLC) is introduced as an input for the estimation approach, resulting in reduced observation noise compared to the time differenced receiver-to-receiver single differences (ΔSD) approach, as one differencing step is avoided. The estimated CPC, using ΔMPLC, with optimized receiver settings shows similar repeatability. Additionally, applying CPC in a single point positioning (SPP) leads to a 70% improvement in the estimated Up component compared to when no CPC is applied. In the observation domain, by calculating SD, the long-period trend can be reliably represented by the estimated CPC.",
author = "Yannick Breva",
year = "2025",
month = may,
day = "23",
doi = "10.15488/19053",
language = "English",
series = "Wissenschaftliche Arbeiten der Fachrichtung Geod{\"a}sie und Geoinformatik der Leibniz-Universit{\"a}t Hannover",
publisher = "Leibniz Universit{\"a}t Hannover",
school = "Leibniz University Hannover",
}