Akademik Veri Yönetim Sistemi
FIG Congress 2022, Warszawa, Polonya, 11 - 15 Eylül 2022, cilt.1, sa.11, ss.1-12
Global Navigation Satellite Systems (GNSS) can effectively provide semi-static and dynamic
displacements by kinematic data processing of high rate observables without any numerical
integration process unlike classical seismological sensors such as accelerometers. However, the
noise amplitude of the high rate GNSS (HR-GNSS) derived displacements is larger than
accelerometers due to various error sources mainly originating from residual atmospheric
propagation effects and multipath caused by the receiver set-up environment. Recent advances
showed that the integration of geodetic grade HR-GNSS displacements with high-quality
accelerometers can provide accurate motion tracking for seismogeodetic applications.
However, geodetic grade HR-GNSS and high-quality accelerometers are expensive tools
together, and seismogeodesy with these high-grade sensors is not cost-effective. In addition to
this, in order to perform HR-GNSS/accelerometer loosely-coupled integration, coordinate
frames of these sensors must be aligned. In this study, a low-cost Inertial Measurement Unit
(IMU) in the Xiaomi Mi8 smartphone was used in the integration of geodetic grade Trimble
NetR9 Precise Point Positioning (PPP) HR-GNSS displacements through a Multi-rate Kalman
Filter. The 1999 Düzce Earthquake (Mw = 7.2) in Türkiye and the 1989 Loma-Prieta Earthquake
(Mw = 6.9) in the United States of America were simulated on a single axis shake table. In order
to provide a cost-effective static alignment for the transformation of the smartphone IMU
coordinate frame into the local North East Down frame, Singular Value Decomposition (SVD)
was used as combining the smartphone accelerometer and the magnetometer sensors. In
addition to this, the same procedure was applied for Xiaomi Mi8 dual-frequency GNSS
displacements and accelerometer integration. The Root Mean Square (RMS) values of Trimble
NetR9 GNSS/smartphone accelerometer for the Düzce and the Loma-Prieta Earthquake
experiments are 4.7 mm and 5.4 mm, respectively. The RMS values of Xiaomi Mi8 dualfrequency GNSS/accelerometer integration are 279 mm for the Düzce Earthquake and 96.7 mm
for the Loma-Prieta Earthquake experiments. Results indicate that the integration of geodetic
grade HR-GNSS with low-cost IMU is suitable for seismogeodetic applications, whereas
Xiaomi Mi8 GNSS/accelerometer integration produces large-scale errors due to the high
uncertainty of its dual-frequency GNSS solutions.