Objectives PPP-RTK (precise point positioning real-time kinematic) is an effective positioning method, for it eliminates the limitations of regional service coverage in RTK(real-time kinematic) and long convergence time in PPP(precise point positioning). However, it is difficult for PPP-RTK to provide continuous and precise positioning results in urban areas for the signals are blocked. Inertial navigation system (INS) and visual navigation can provide continuous navigation information but suffers from error drifts.So, we proposed semi-tightly coupled multi-GNSS (global navigation satellite system) PPP-RTK/VIO (vision inertial odometry) model to complement each other.
Methods In order to evaluate the performance of the integration model, firstly, we collected data in the campus of Wuhan University. Secondly, we used the data to test the integration model and compared the result with other solutions.
Results The result shows that semi-tightly coupled multi-GNSS PPP-RTK/VIO can improve the positioning accuracy by more than 30% in comparison to GPS (global positioning system) +BDS (BeiDou navigation satellite system) PPP-RTK. Its accuracy is 0.58 m and 1.12 m in horizontal and vertical direction, respectively. Moreover, it also performs well in velocity and pose estimation. The accuracy of velocity estimation is 0.04 m/s, 0.04 m/s and 0.02 m/s, respectively, and the accuracy of pose estimation are 0.10, 0.06 and 0.17, respectively.
Conclusions Semi-tightly coupled multi-GNSS PPP-RTK/VIO can make use of their own advantages, improve the positioning performance of PPP-RTK and realize better velocity and pose estimation.