Memories of PPP Beginnings

On 19 June 2017, Pierre Héroux officially retired from the Canadian Geodetic Survey of Natural Resources Canada (NRCan) after spending nearly 35 years with this organization. Over these years, Pierre has been a key player in the GNSS community, from testing the first commercial GPS receivers to being an early instigator of precise point positioning (PPP). Pierre kindly accepted my invitation to write a recollection of the early days of PPP and how this positioning methodology evolved within the IGS.


“As I retire, I am pleased to respond to Simon's request to look back at PPP beginnings and remind users of its early history and the key role of IGS collaboration and open products.


When I look back at IGS beginnings in the early 1990s, I recall that most GPS experts at NRCan were focused on improving core IGS products, in particular precise satellite orbits and Earth orientation parameters. Back then, these products were key to connect the inertial and terrestrial reference frames and support double-differenced positioning algorithms. Less attention was given to satellite clock parameters as a number of analysis centres did not estimate them. Therefore, efforts in improving their quality and agreeing to a standard exchange format were somewhat delayed. Fortunately, a few agencies had implemented undifferenced precise orbit determination models which provided developers the opportunity to experiment with PPP methods. While PPP existed as a special case of a more general undifferenced network solution, the implementation of a standalone single-point application still had to be done. This separate application also needed to tap into GPS orbit and clock products that were not yet available at high rate. The PPP approach also had to show that it offered significant benefits if it was to receive widespread acceptance.


As the 1990s was a time when selective availability (SA) degraded the precision of stand-alone positioning to a few tens of metres, the prospect of sub-metre positioning without access to base station data was quite attractive. PPP efforts at NRCan began in 1994 and were mainly focused on finding a method of interpolating IGS satellite clocks to estimate the effects of SA. For practical reasons, this was essential as IGS orbit products were only available at 15-minute intervals, a period of time during which the random-walk signature of SA introduced errors of several tens of metres in range measurements. While satellite coordinates and clock offsets were estimated by some IGS analysis centres with centimetre precision, their states at intermediate epochs during that 15-minute period were unknown. This situation had some impact on convergence of static positioning, but was a serious limitation in terms of usefulness for any fast-static or kinematic application. In order to fully exploit the potential of IGS satellite clocks at 15-minute intervals, some clock interpolation method needed to be devised and applied.


The approach implemented at NRCan was to constrain the linear drift of the hydrogen maser clocks operated at selected IGS sites during 15-minute time periods. This was done by back-substituting IGS station, satellite and tropospheric parameters into a single-point positioning solution. This process enabled the recovery of satellite clocks from the station observations at the measurement sampling interval. Using the variation of the ionosphere-free carrier-phase combination, this method effectively interpolated SA between the 15-minute satellite clock epochs. Combining the SA estimates for all satellites in common view from a global network of stations, global high-rate satellite clock files were then generated. This was the first step at enabling a PPP solution that would receive some interest from the user community.


Once high-rate global satellite clocks at 30-second intervals were produced daily at NRCan, attention was given to implementing a PPP solution for end users. Initially, a dual-frequency code-based solution was developed to offer both static and kinematic positioning for users acquiring GPS data at any sampling interval. This initial PPP release required the optimization of satellite clock interpolation, as satellite clocks were only available every 30 seconds, until IGS high-rate data acquisition started in the late 1990s. The code-based PPP solution was also made a priority at the time as it was still unclear whether or not the quality of IGS orbits and 30-second clocks would be sufficient to support the full precision of the carrier-phase measurements. Initial results of the NRCan PPP implementation were first presented in 1995 in a paper entitled “GPS Precise Point Positioning with a Difference” (Héroux and Kouba 1995).


As the precision and update rate of IGS orbit and clock products increased over the following years, it became apparent that they would be suitable to exploit the millimetre resolution of the carrier-phase observable. This was confirmed by Zumberge et al (1997) in their paper “Precise point positioning for the efficient and robust analysis of GPS data from large networks.” This proven capability and improved internet communication led NRCan to also develop a phase-based online solution, known as CSRS-PPP.  While first results were published by Kouba and Héroux (2001), the operational service was only released in October of 2003. Since then, it has been updated to support GLONASS satellites and multi-day processing for time transfer. To date, CSRS-PPP has processed over a million datasets submitted by end users operating worldwide as detailed in an article submitted by Klatt et al to Geomatica.


Today, a number of post-mission online PPP services are available from public agencies. Commercial providers have also implemented a number of real-time services delivering PPP solutions worldwide. While a couple decades have passed since PPP beginnings, the search for instantaneous ambiguity resolution in a multi-constellation environment still makes PPP an interesting topic for research and development. For users, it continues to offer a relatively simple and efficient way of accessing the global reference frame at the highest accuracy level.”


Thank you Pierre for sharing this fascinating story and for your invaluable contribution to the world of GNSS. Your wisdom will be greatly missed at NRCan and in the GNSS community in general. Enjoy your well-deserved retirement and rest assured that we will always listen to any bits of advice you may still want to share!




Héroux P, Kouba J (1995) GPS precise point positioning with a difference. Presented at Geomatics ’95, Ottawa, Ontario, Canada, June 13-15. (Download)


Kouba J, Héroux P (2001) Precise point positioning using IGS orbit and clock products. GPS Solutions 5(2):12-28 doi:10.1007/PL00012883


Zumberge JF, Heflin MB, Jefferson DC, Watkins MM, Webb FH (1997) Precise point positioning for the efficient and robust analysis of GPS data from large networks. J Geophysical Research 102(B3):5005-5017


Write a comment

Comments: 0