This paper presents an implementation of a postprocess kinematic positioning technique similar to PPP (Precise Point Positioning) but based on a network of stations. This technique is independent of precise clock information because it estimates satellite clock offsets 'on-the-fly', and thence it only needs reasonable accurate orbits information (for instance, IGS precise, rapid or even predicted products) to work. Moreover, with this approach the solution rate is only limited by data rate, and not by the availability of precise satellite clock data rate as it is the case with classical kinematic PPP techniques. This procedure is referred to in this paper as 'Precise Orbits Positioning' (POP). Similar methods already exist in the literature, such as 'network-based clocks' and 'phase interpolation'; however, with POP method both predicted and rapid IGS orbits may also be used (not only precise products) without noticeable degradation of positioning results, which is an important advantage. The POP procedure involves multiple stations separated hundreds of kilometers and there are a great number of unknowns of several kinds. In order to ease the implementation of such system some GPSTk-provided facilities were used, including a run-time programmable general solver. A contribution of this work is that a reference implementation is freely available at the development version of the GPSTk source code, facilitating its use and modification by other GNSS researchers. The POP results were very similar (as expected) to the standard kinematic PPP strategy, but yielding a higher positioning rate. This higher positioning rate opens the way for post-process kinematic positioning of vehicles that usually operate very far from reference stations, such as aircraft. Also, our experiments with this network-based processing strategy show additional robustness in their results, even for receivers outside the network area with long baselines.
