Evaluation of Dual-Frequency Multi-Constellation PPP/INS using QZSS Augmentation Data

With the increasing demand for autonomous driving applications, accurate navigation solutions are required. Lane-level navigation needs to be developed to enable robust vehicle guidance. GNSS can provide a user with the precise absolute position, and the solution accuracy is independent of trajectory and time. Precise point positioning (PPP) is used widely because of its high positioning accuracy. In recent years, relatively fast convergence PPP has been realized with an SBAS augmentation signal. In this study, a multi-GNSS advanced orbit and clock augmentation (MADOCA) product was applied to enhance the navigation performance and shorten the PPP convergence time. The augmentation data provided correction messages for the satellite orbit and clock bias in the Radio Technical Commission for Maritime Services (RTCM) state-space representation (SSR) format. The performance of static PPP combining broadcast ephemeris with the MADOCA product was first evaluated between different types of the GNSS constellation, such as GPS, GLONASS, and QZSS. Ionosphere-free combination was also applied to eliminate the ionospheric ranging errors. Furthermore, the vehicle experiment was conducted in Taiwan in a suburban environment to assess the performance of kinematic MADOCA PPP. Along with the root mean square error (RMSE), the average satellite number was compared among different types of constellations as mentioned above. However, standalone MADOCA PPP might pose severe problems because of environmental factors. Lane-level navigation was still difficult to achieve even with augmentation data aiding. Thus, kinematic MADOCA PPP was integrated with tactical-grade IMU by using tightly coupled scheme, which made the solution robust. Accurate velocity and attitude could thus be estimated. The result revealed that an INS-integrated MADOCA PPP could improve the performance in 3D and that lane-level positioning was achieved in a suburban area. All the algorithms proposed in this paper were processed in post-mission, but the application of the real-time satellite orbit and clock correction products enabled the analysis to approach a navigation solution in the real world.