4:30 PM - 4:45 PM
[MTT37-11] Optimization of a GNSS multipath noise reduction method based on MHM methods
Kinematic GNSS analysis for monitoring crustal deformation with high-rate sampling is a crucial technique, especially for understanding seismic and volcanic phenomena on timescales of one day or less. However, kinematic GNSS analysis exhibits high noise levels in the time domain ranging from several hours to several days, making it challenging to detect small-scale deformations with similar time scales. Multipath noise is one of the causes of this noise and originates from the environment surrounding the GNSS receiver antenna.
Methods for reducing the multipath noise include utilizing the satellite revisit period of GPS satellites, such as the sidereal filter and the Multipath Hemispherical Map (MHM) method, which creates a carrier-phase residual map before positioning analysis and corrects the carrier-phase data to remove multipath noise (e.g., Iwabuchi et al., 2004; Fuhrmann et al., 2015). In particular, the MHM method applies to multi-GNSS positioning analysis.
We have developed the Multi-Sites Stacked Multipath Hemispherical Map (MSS-MHM) method, which generates a MHM from short-baseline double-difference positioning with multiple baselines for a GNSS station targeted for multipath noise reduction, thereby removing multipath noise. Applying this method to a GNSS station operated by SoftBank Corp., which may be affected by multipath, showed an improvement in the stability of the kinematic coordinate time series for multi-GNSS positioning analysis. This method is theoretically less susceptible to tropospheric and ionospheric delays in creating MHM. In addition, it is expected to be independent of positioning analysis software and strategies because it outputs carrier-phase data with multipath correction in RINEX format. However, it is necessary to perform double-difference positioning by selecting multiple reference stations while keeping a suitable baseline length to a rover station where tropospheric and ionospheric delays are negligible.
In this study, we utilized a high spatial density GNSS observation network consisting of the GEONET (operated by Geospatial Information Authority of Japan) and SoftBank Corp.'s original GNSS stations to optimize the MSS-MHM method. This study utilized 30s sampling data from January 1, 2023, to January 30, 2023, acquired from GEONET and SoftBank stations around Miyagi Prefecture. RTKLIB (Takasu, 2013) was used for the analysis, and the final CODE MGEX orbit and clock (Dach et al., 2024) was used.
First, perform double-difference positioning analysis up to the day before the correction target date (1 to 15 days), with the surrounding GEONET as base stations. Second, calculate the median carrier-phase residual from several baselines (number of baselines: 1 to 5) for each satellite within an optimum grid of elevation and azimuth angles, obtaining the carrier-phase residual value. Third, the carrier-phase data in the RINEX observation file for the target station should be corrected. Finally, long-baseline double-difference positioning will be performed using the corrected RINEX observation file, and the standard deviation of the coordinate time series and the ambiguity resolution fix rate (AR-FIX rate) before and after correction will be compared.
Preliminary application of this method to multi-GNSS analysis using GPS, GLONASS, and QZSS has confirmed a reduction in the standard deviation of the coordinate time series and an improvement of AR-FIX rate when the number of days used for MHM is longer than the satellite revisit period, and the number of baselines is more significant. We will comprehensively analyze and optimize the parameters with a high multipath noise reduction effect.
Acknowledgments: The SoftBank's GNSS observation data used in this study was provided by SoftBank Corp. and ALES Corp. through the framework of the "Consortium to utilize the SoftBank original reference sites for Earth and Space Science".
Methods for reducing the multipath noise include utilizing the satellite revisit period of GPS satellites, such as the sidereal filter and the Multipath Hemispherical Map (MHM) method, which creates a carrier-phase residual map before positioning analysis and corrects the carrier-phase data to remove multipath noise (e.g., Iwabuchi et al., 2004; Fuhrmann et al., 2015). In particular, the MHM method applies to multi-GNSS positioning analysis.
We have developed the Multi-Sites Stacked Multipath Hemispherical Map (MSS-MHM) method, which generates a MHM from short-baseline double-difference positioning with multiple baselines for a GNSS station targeted for multipath noise reduction, thereby removing multipath noise. Applying this method to a GNSS station operated by SoftBank Corp., which may be affected by multipath, showed an improvement in the stability of the kinematic coordinate time series for multi-GNSS positioning analysis. This method is theoretically less susceptible to tropospheric and ionospheric delays in creating MHM. In addition, it is expected to be independent of positioning analysis software and strategies because it outputs carrier-phase data with multipath correction in RINEX format. However, it is necessary to perform double-difference positioning by selecting multiple reference stations while keeping a suitable baseline length to a rover station where tropospheric and ionospheric delays are negligible.
In this study, we utilized a high spatial density GNSS observation network consisting of the GEONET (operated by Geospatial Information Authority of Japan) and SoftBank Corp.'s original GNSS stations to optimize the MSS-MHM method. This study utilized 30s sampling data from January 1, 2023, to January 30, 2023, acquired from GEONET and SoftBank stations around Miyagi Prefecture. RTKLIB (Takasu, 2013) was used for the analysis, and the final CODE MGEX orbit and clock (Dach et al., 2024) was used.
First, perform double-difference positioning analysis up to the day before the correction target date (1 to 15 days), with the surrounding GEONET as base stations. Second, calculate the median carrier-phase residual from several baselines (number of baselines: 1 to 5) for each satellite within an optimum grid of elevation and azimuth angles, obtaining the carrier-phase residual value. Third, the carrier-phase data in the RINEX observation file for the target station should be corrected. Finally, long-baseline double-difference positioning will be performed using the corrected RINEX observation file, and the standard deviation of the coordinate time series and the ambiguity resolution fix rate (AR-FIX rate) before and after correction will be compared.
Preliminary application of this method to multi-GNSS analysis using GPS, GLONASS, and QZSS has confirmed a reduction in the standard deviation of the coordinate time series and an improvement of AR-FIX rate when the number of days used for MHM is longer than the satellite revisit period, and the number of baselines is more significant. We will comprehensively analyze and optimize the parameters with a high multipath noise reduction effect.
Acknowledgments: The SoftBank's GNSS observation data used in this study was provided by SoftBank Corp. and ALES Corp. through the framework of the "Consortium to utilize the SoftBank original reference sites for Earth and Space Science".