Performance Analysis of Tightly Coupled Precise Point Positioning with INS in Urban Areas

In this paper, we focus on kinematic Precise Point Positioning (PPP) combined with inertial sensor data from a tactical grade IMU by "Tightly Coupled Integration" techniques. During the last years Precise Point Positioning has become a practical GNSS processing technology with the advantage that no local base station is required for decimetre accuracy on any place on earth. The main disadvantage of pure PPP post-processing is the long convergence time compared to a RTK system, especially if signal outages are present, for example in urban areas. To overcome the convergence issues after signal outages or during bad satellite geometry, a Tightly Coupled approach is proposed so that short outages can be bridged and the convergence time does not increase again. Our Tightly Coupled PPP/INS Integration is implemented as post processing software based on an Error State Kalman Filter and is capable of using IGS products to correct orbit and satellite clock observations. Other error sources in PPP addressed with models and observations in the navigation filter. Finally an RTS smoother is implemented to gain optimal results especially during GNSS outages. The paper discusses the robustness and accuracy of Tightly Coupled PPP-GPS/INS Integration in urban areas compared with open sky tests. We present car test runs in urban areas as well as open sky tests to compare the results. The ground truth is provided by our RTK system.