Improving the combined GNSS/INS positioning by using tightly integrated RTK

• Published in: GPS solutions 2022-10, Vol.26 (4), Article 144
• Main Authors: Li, Bofeng, Chen, Guang’e
• Format: Article
• Language: English
• Subjects: Algorithms; Atmospheric Sciences; Automotive Engineering; Earth and Environmental Science; Earth Sciences; Electrical Engineering; Field tests; Geophysics/Geodesy; Global navigation satellite system; Horizontal orientation; Kinematics; Occlusion; Original Article; Performance enhancement; Position errors; Satellites; Space Exploration and Astronautics; Space Sciences (including Extraterrestrial Physics
• ISSN: 1080-5370; 1521-1886
• DOI: 10.1007/s10291-022-01331-2

GNSS/INS combined technology is nowadays one of the most widely used high-precision positioning methods for outdoor users. However, the performance of the GNSS/INS combined system will be degraded when the satellite signals are subject to long-term obstruction or interference. A common way to overcome this limitation is to integrate the additional sensors, which increases additional expenses. Motivated by our previous work that the Tightly integrated Real-Time Kinematic (TRTK), which makes use of the additional inter-system differential observations of overlapping frequencies and can significantly improve the performance of the Loosely integrated RTK (LRTK) which directly stacks the intra-system differential observations and is currently used in the GNSS/INS combined system, we further study the enhanced GNSS/INS combined technique to improve the positioning performance by using TRTK, particularly in harsh environments. First, the difference between TRTK and LRTK are analytically addressed, and the uniform expression for three TRTK models specified by differential inter-system biases, i.e., DISB-float, -constant and -fixed models, are presented. Then, the GNSS/INS combined equations are derived based on the DISB-fixed TRTK model. Both semi-simulated and real field tests are carried out in different environments to show the improvements of GNSS/INS with TRTK compared to that with LRTK. The result shows that the TRTK-specified GNSS/INS combined model can significantly improve positioning accuracy, especially when the satellite signals suffer severe occlusion. Thanks to the four-system dual-frequency observations we used, at most 6 additional redundant observations will be introduced at each epoch in TRTK. Compared to the traditional algorithm, our proposed algorithm on average reduces the averaged horizontal position error by 53.2% for the two tests, and even by 73.1% in a situation where the number of visible satellites is less than about 7.