Session B3: Precise GNSS Positioning Applications

MADOCA-PPP/INS Tightly Coupled Integration Based on RTKLIB for Moving Vehicles
Hideki Yamada, Saya Matsushita, Keito Yoshida, Fuya Ezuka, Tatsuya Nagano and Satoshi Kogure, Japan Aerospace Exploration Agency
Location: Beacon B

1. Introduction and Objective
JAXA developed Multi-GNSS Advanced Demonstration tool for Orbit and Clock Analysis precise GNSS orbit and clock determination tool (MADOCA) [1] as a function of QZSS 1st satellite. Currently, JAXA continues to research and promote MADOCA and its precise point positioning (PPP) (called MADOCA-PPP) with both Internet derived product from JAXA and QZSS L6E signals (called QZSS MADOCA-PPP from the Cabinet Office [2]), which is based on RTKLIB, an open-source program package for GNSS positioning. The RTKLIB has a proven track record as a software for a low-cost receiver [3]. However, to promote MADOCA-PPP navigation for moving vehicles, such as unmanned aerial vehicles (UAVs) and cars, providing robust and highly accurate positioning solutions even with weak GNSS signals has become a significant technical issue. To overcome these issues, the integration with additional IMU sensor not affected by weak GNSS signals have been evaluated in previous studies. There are literatures [4][5] regarding PPP/INS tightly coupled integration which describes INS-derived position and velocity terms in the range and range-rate observation equations. But their literatures have not been evaluated in the MADOCA product and RTKLIB. Other literature [6] has also evaluated MADOCA-PPP/INS, but only achieved loosely coupled integration which describe the equation between the INS-derived position and velocity and the PPP-derived ones. Based on the above, as far as we have investigated, there are currently no papers that discuss MADOCA-PPP tightly coupled integration and its implementation in RTKLIB to evaluate its effectiveness.
As a notable difference from previous studies, the main objective of this study is to establish PPP/INS based on RTKLIB to promote MADOCA-PPP for moving vehicles. Therefore, PPP/INS tightly coupled integration is first implemented in RTKLIB which enables MADOCA-PPP, and its effectiveness is presented in this study. Both MADOCA products obtained via the QZSS MADOCA-PPP signal as well as SP3 format via the Internet are used. In this study, the evaluation regarding the integration of INS with multi-GNSS PPP ambiguity resolution (PPP-AR) is presented. A summary of how to implement PPP/INS tightly coupled integration in RTKLIB is also described.

2. Methodology
In this study, the MADOCA-PPP Library (MALIB) which is based on RTKLIB, is used since it has the function of decoding correction data in MADOCA formats like QZSS L6E CSSR and SP3. In addition, MALIB can execute PPP-AR in contrast to conventional RTKLIB and the PPP/INS integration function was incorporated into MALIB to estimate INS navigation position, velocity, and attitude errors and IMU biases using 3-axis gyro and accelerometer data according to the literature [7]. For the experimental data, we used a UAV and a car as the bodies and installed a Septentrio GNSS receiver with an IMU and two GNSS antennas on the bodies to obtain GNSS/IMU data. The reference solutions for the vehicle’s bodies were calculated using software such as Qinertia and PosPac and were used for comparison with our PPP solution.
3. Results
We conducted a preliminary evaluation of our algorithm using GNSS/IMU data obtained from the UAV and the car. Firstly, MADOCA-PPP/INS could suppress the increase with position errors and close to the reference position during missing epoch of GPS carrier-phase measurement in the UAV navigation. Next, we confirmed the effect of the QZSS MADOCA-PPP-AR/INS tightly coupled integration on PPP-AR without IMU and revealed that the proposed method could reduce position errors by 28 % and achieve approximately 20 cm class position accuracy in the car navigation. To maintain an accurate fixed solution for PPP-AR/INS, it is necessary to suppress the increase in the observation residual due to missing epochs of not only GNSS observation data but IMU data. We believe that a good approach is to exclude the increased observation residual from the PPP-AR estimation process with a threshold value, and this study also discuss its effectiveness.
4. Conclusions
We proposed MADOCA-PPP/INS tightly coupled integration based on RTKLIB and verified the effect of this method for UAV and car. The proposed method was found to suppress the increase position errors of MADOCA-PPP compared to the case without the IMU sensor or during GNSS observation data loss and to maintain position accuracy at decimeter class. We are currently investigating ways to maintain better position accuracy even in cases where the vehicle rotates a lot, and this will be presented in further study. The prototype version and algorithm developed in this research are expected to serve as reference information for promoting PPP/INS navigation.
5. New and/or Innovative about our research
I) The novelty is that MADOCA-PPP/INS tightly coupled integration was first evaluated using QZSS L6E CSSR and SP3 formats.
II) The additional distinguishing point of this research is that the proposed method was first implemented in RTKLIB, and evaluation regarding the integration with INS for several PPP methods (multi-GNSS and PPP-AR) was presented.

[1] Kawate K., et al. (2023). MADOCA: Japanese precise orbit and clock determination tool for GNSS, Advances in Space Research, Volume 71, Issue 10, 3927-3950.
[2] CAO. (2022). IS-QZSS-MDC-001, Quasi-Zenith Satellite System Interface Specification Multi-GNSS Advanced Orbit and Clock Augmentation-Precise Point Positioning.
[3] Takasu T., & Yasuda A. (2009). Development of the low-cost RTK-GPS receiver with an open source program package RTKLIB, International Symposium on GPS/GNSS, International Convention Center Jeju, Korea, November 4–6.
[4] M Abd Rabbou., et al. (2015). Tightly coupled integration of GPS precise point positioning and MEMS-based inertial systems, GPS Solution 19:601–609.
[5] Luguang L., et al. (2023), PPP/INS Tight Integration with BDS?3 PPP?B2b Service in the Urban Environment, sensors, 23, 2652?
[6] Kubo N., et al. (2024). Performance Evaluation of PPP/PPP-RTK in Semi-Urban Areas, Proceedings of the ION 2024 Pacific PNT Meeting, 54-64, April 15 - 18.
[7] Groves, P.D, “Principles of GNSS, inertial, and 11multisensory integrated navigation systems”, 2nd edition, 2015.