Real-time cascading PPP-WAR based on Kalman filter considering time-correlation

Precise point positioning (PPP) with wide-lane ambiguity resolution (PPP-WAR) can quickly achieve decimeter-level positioning accuracy, which is an essential improvement for real-time PPP in terms of reliability and accuracy. However, current implementations of PPP-WAR are based only on the standard Kalman filter (KF) under the assumption of uncorrelated measurements. As Global Navigation Satellite System measurements are often contaminated by colored noises, a KF considering the time-correlation of measurements (CKF) has the potential to improve the performance of PPP-WAR. In this paper, we proposed a real-time cascading PPP-WAR technique based on the CKF model. A real-time kinematic vehicle experiment is conducted to assess the performance of the proposed real-time PPP-WAR method. On the one hand, we analyze and compare the efficiency of CKF-based cascading PPP-WAR with that of KF-based PPP-WAR. The filter output analysis shows that the proposed CKF method is, overall more efficient than the standard KF method in terms of performance indicators, such as wide-lane ambiguity fix-rate, time-to-first-fix, ambiguity residual statistics and correctness of the filter's stochastic model. On the other hand, we evaluate the convergence time and positioning accuracy of CKF-based PPP-WAR by comparing it with KF-based PPP-WAR. The results show that the fix-rate of the CKF method increased by 26% compared with the KF-based WAR method. Moreover, the CKF-PPP-WAR can achieve 0.5 m (sub-meter) level with 1-4 s faster, 0.2 m level with 2-20 s faster, 0.1 m level with 5-60 s faster compared with float PPP. On average, CKF-PPP-WAR can achieve decimeter-level accuracy by reducing the convergence time 20% than KF-PPP-WAR, and 25% than float PPP. After convergence, the average positioning accuracy of CKF-PPP-WAR is 2-6 cm better with 25% improvement over float PPP and 1-4 cm better with a 23% improvement over KF-PPP-WAR.