3:00 PM - 3:15 PM
[SCG55-29] Realization of high-precision GNSS-A observation using a seaplane-type UAV
Keywords:GNSS-A, Acoustic ranging, Seafloor crustal deformation observation, UAV, Seaplane
There is a method called Global Navigation Satellite System – Acoustic combination (GNSS-A), which combines GNSS and acoustic ranging to observe seafloor crustal deformation. This method repeatedly performs acoustic ranging between a transducer—whose absolute position is determined from GNSS and the attitude information of a sea surface platform—and a seafloor station installed on the seafloor at depths exceeding 1,000 m. This identifies the location of the seafloor station. In particular, the Japan coast guard has deployed a GNSS-A observation network called the Seafloor Geodetic Observation – Array (SGO-A), mainly settled on the Japan Trench and the Nankai Trough, to continuously monitor seafloor crustal deformation.
Although GNSS-A observation data provides valuable information for disaster prevention science, the observations are mainly conducted from vessels, and the high cost of such observations limits their frequency. Consequently, research is underway on GNSS-A observations using a seaplane-type Unmanned Aerial Vehicle (UAV) as a new sea surface platform that can reduce resources and allow for rapid post-earthquake observations. Test observations with a prototype have demonstrated that seafloor positioning can be achieved at the sub-meter level (Yokota et al., 2023). In this study, aiming for centimeter-level seafloor positioning, we evaluated the performance of a UAV-based GNSS-A observation system that was reconstructed based on the insights obtained from the prototype using a new observation system. Furthermore, we conducted GNSS-A observations at an actual SGO-A observation site and evaluated the results.
We evaluated the performance of the GNSS and IMU devices, as well as the acoustic ranging device mounted on the UAV. In particular, the accuracy of the acoustic ranging device was verified by performing acoustic ranging in a test tank where the true distance was known. As a result, the ranging error was kept at the centimeter level. Moreover, it was suggested that an acoustic ranging bias equivalent to a fractional wavelength of the carrier exists, which cannot be corrected by the existing bias correction method “AAR method” (Yokota et al., 2024).
UAV-based GNSS-A observation was conducted in January 2024 at the actual SGO-A observation site “SAGA” off the coast of Ito, Shizuoka Prefecture. Due to engine trouble of the UAV during the observation, the experiment was stopped in the middle of the survey line. As a result, the number of GNSS-A data acquired was only about half of the number of data normally acquired by a survey vessel. GNSS-A analysis was performed using GARPOS v.1.0.1 (Watanabe et al., 2020). The determined seafloor position was compared with seafloor positions obtained by survey vessels. The horizontal positions determined were generally consistent with the results from the survey vessels. The success of centimeter-class observations with a limited amount of data is thought to be due to the fact that UAV has a shorter offset distance between the GNSS antenna position and the sonar than survey vessels, which reduces the effect of attitude information bias on the sonar position.
On the other hand, a discrepancy of 10–15 cm was observed in the vertical direction. This may be due to the possibility that the phase center of the transducer assumed in SGO-A differs from the actual one, or due to an acoustic ranging bias of the survey vessel’s system. It suggests that all the seafloor positions determined by the survey vessel might be offset from the true values.
In conclusion, the UAV-based GNSS-A observation has been shown not only to achieve a reduction in resources but also to possess the potential to rapidly acquire high-quality data. However, new challenges, such as integrating biases across sea surface platforms, have also been revealed.
Acknowledgement: This study was supported by ERI JURP 2024-Y-KOBO12 in Earthquake Research Institute, the University of Tokyo, by SECOM science and technology foundation, and by JSPS KAKENHI Grant Number JP21H05200 in Grant-in-Aid for Transformative Research Areas (A) “Science of Slow-to-Fast Earthquakes.”
Although GNSS-A observation data provides valuable information for disaster prevention science, the observations are mainly conducted from vessels, and the high cost of such observations limits their frequency. Consequently, research is underway on GNSS-A observations using a seaplane-type Unmanned Aerial Vehicle (UAV) as a new sea surface platform that can reduce resources and allow for rapid post-earthquake observations. Test observations with a prototype have demonstrated that seafloor positioning can be achieved at the sub-meter level (Yokota et al., 2023). In this study, aiming for centimeter-level seafloor positioning, we evaluated the performance of a UAV-based GNSS-A observation system that was reconstructed based on the insights obtained from the prototype using a new observation system. Furthermore, we conducted GNSS-A observations at an actual SGO-A observation site and evaluated the results.
We evaluated the performance of the GNSS and IMU devices, as well as the acoustic ranging device mounted on the UAV. In particular, the accuracy of the acoustic ranging device was verified by performing acoustic ranging in a test tank where the true distance was known. As a result, the ranging error was kept at the centimeter level. Moreover, it was suggested that an acoustic ranging bias equivalent to a fractional wavelength of the carrier exists, which cannot be corrected by the existing bias correction method “AAR method” (Yokota et al., 2024).
UAV-based GNSS-A observation was conducted in January 2024 at the actual SGO-A observation site “SAGA” off the coast of Ito, Shizuoka Prefecture. Due to engine trouble of the UAV during the observation, the experiment was stopped in the middle of the survey line. As a result, the number of GNSS-A data acquired was only about half of the number of data normally acquired by a survey vessel. GNSS-A analysis was performed using GARPOS v.1.0.1 (Watanabe et al., 2020). The determined seafloor position was compared with seafloor positions obtained by survey vessels. The horizontal positions determined were generally consistent with the results from the survey vessels. The success of centimeter-class observations with a limited amount of data is thought to be due to the fact that UAV has a shorter offset distance between the GNSS antenna position and the sonar than survey vessels, which reduces the effect of attitude information bias on the sonar position.
On the other hand, a discrepancy of 10–15 cm was observed in the vertical direction. This may be due to the possibility that the phase center of the transducer assumed in SGO-A differs from the actual one, or due to an acoustic ranging bias of the survey vessel’s system. It suggests that all the seafloor positions determined by the survey vessel might be offset from the true values.
In conclusion, the UAV-based GNSS-A observation has been shown not only to achieve a reduction in resources but also to possess the potential to rapidly acquire high-quality data. However, new challenges, such as integrating biases across sea surface platforms, have also been revealed.
Acknowledgement: This study was supported by ERI JURP 2024-Y-KOBO12 in Earthquake Research Institute, the University of Tokyo, by SECOM science and technology foundation, and by JSPS KAKENHI Grant Number JP21H05200 in Grant-in-Aid for Transformative Research Areas (A) “Science of Slow-to-Fast Earthquakes.”