Model and Data Driven Partial Ambiguity Resolution for Multi-Constellation GNSS

The goal of carrier-phase ambiguity resolution is to exploit that the carrier-phase observations start to act as very precise pseudoranges. With the development of modern GNSS (GPS, BDS, Galileo, Glonass), more than 30 satellites are visible, however, it might be impossible to reliably fix all the ambiguities due to the computation time. Additionally, due to high measurement noise or residual atmosphere delays in case of longer baselines, the observation model is not strong enough, which makes it impossible to fix all the ambiguities. Therefore Partial Ambiguity Resolution (PAR) becomes more and more essential for real-time precise positioning. In this contribution, a Model and Data driven PAR (MD-PAR) strategy is proposed, and implemented in two different ways. The performance of MD-PAR is assessed using a simulation study by the probability of correct subset fixing, the subset size, and the Root Mean Square (RMS) of the baseline solution. Furthermore, MD-PAR is compared with the classic strategy, which uses only model information. The analysis and simulation results both suggest that the new strategies have better performance than the classic strategy.