A major limitation of precise point positioning (PPP) is that it needs a long convergence period to obtain the steady-state float
ambiguities. This slow convergence is largely caused by the slow change of satellite geometry and the remaining un-modeled errors.
In contrast, with combining GPS and GLONASS observations, the more visible satellites can improve the poor satellite geometry
distribution, and then a reduction in the convergence period can be achieved. By mitigating the multipath and noise in pseudoranges,
the float ambiguity estimates can be more accurate rapidly, and therefore the PPP convergence period can be reduced. This study
introduces the pseudorange multipath and noise mitigation to a combined GPS/GLONASS PPP model, aiming to shorten the PPP
convergence period. TEQC multipath linear combination or observable is used to compute the pseudorange multipath and noise
corrections. Pseudorange multipath and noise are mitigated using two methods: firstly, direct multipath observable corrections in the
PPP functional model; and secondly, weights derived from the multipath observable in the PPP stochastic model. 7-day data sets
from 108 globally distributed IGS stations were utilized. The results indicated that combining GLONASS observations with GPSonly PPP can reduce the convergence period by 32 % in case of equal- weighted pseudoranges for each system. In addition, the
weights derived from TEQC multipath observable can reduce the convergence period of combined GPS/GLONASS-PPP by 25 %. |
