Objectives Precise point positioining (PPP) based on integer least squares (ILS) requires the integer testing procedure during which any missed or false test result may lead to incorrect ambiguity fixing. While adopting the best integer equivariant (BIE) estimation scheme, it does not require integer testing anymore and can obtain ambiguity solutions with the property of minimum mean square error in all integer estimation classes.
Methods Using the observation data of 100 international global navigation satellite system service stations distributed globally, the simulated kinematic PPP with ambiguity resolution (PPP-AR) performance is evaluated and compared under different processing strategies, namely BIE estimation, full ambiguity resolution, and partial ambiguity resolution (PAR).
Results The results show that in the stage where the strength of the PPP ambiguity estimation model is relatively weak and the convergence precision of estimated ambiguities is limited, traditional ILS algorithm will bring a higher risk of false ambiguity fixing, while the minimum mean square error solution of BIE based on the weighted fusion of ambiguity integer candidate solutions is more stable and can effectively suppress the PPP solution jump caused by abnormal estimation result.
Conclusions Compared with the ILS-PAR solution strategy, BIE can significantly improve the convergence speed of PPP-AR. Under the 70% and 90% percentiles, the convergence time of horizontal and vertical components is further reduced by 18.8%, 13.8%, and 24.3%, 15.9% respectively. Moreover, BIE clearly outperforms ILS-FAR improving the positioning accuracy by 43.3% and 15.3% in the horizontal and vertical components respectively (under 90% quantile), which is only slightly better than ILS-PAR.
DownLoad: