The Sentinels form a family of Earth Observation satellites that are developed and operated by ESA in the framework of the Copernicus program. The Copernicus program aims at providing long-term, comprehensive monitoring of the Earth not just by building individual Earth observation satellites, such as ERS and Envisat in the past, but by providing long-term continuous data services in many areas of Earth observation, such as radar imagery, multispectral optical data, atmospheric measurements, temperature of oceans and atmosphere, or ocean level.
The six different Sentinel missions - logically numbered Sentinel-1 to Sentinel-6 - each consist of pairs of satellites or passenger payloads on other satellites. The pair of satellites fly in tandem orbits, offering overlaps in time and geography of most datasets, and in many cases allowing for a kind of three-dimensional vision in a way similar to a pair of eyes.
Most of the Sentinel missions require precise orbit information in order to maximize the amount and detail of information that can be extracted from the scientific payload. For instance, the accuracy of a radar altimetry mission is directly dependent on the accuracy at which we know the height of the satellite above the ocean or ice mass that it is measuring. For precise orbit determination, the Sentinel satellites typically rely on GNSS observations, but some of the Sentinel satellites also carry additional precise tracking equipment such as Laser Retro-Reflectors or DORIS instruments (Doppler Orbitography and Radio-positioning Integrated by Satellite). Also, a radar altimeter payload instrument can be used effectively as tracking device, notably in the form of cross-over differences in the many thousands of intersection points of the satellite ground track pattern.
The routine precise orbit determination for the Sentinel missions is performed by European industry at dedicated processing facilities for the various Sentinel missions. However, the Navigation Support Office also has full capability for precise orbit determination of all current and future Sentinel satellites, both to support and validate the routine operations by independent means.
Because the low Sentinel satellites are not in constant view of a download ground station, the GNSS data (and other data) of the Sentinel satellites is stored on-board, and downloaded to any of three ground stations as soon as this station is visible. Because these ground contacts are short in duration (a few minutes), the data that can be downloaded has to be carefully selected so that the amount of data remains compatible with the capacity of the communication links. Furthermore, the amount of processing power on-board a satellite is always strictly budgeted to optimize all aspects of the mission in parallel.
For the GNSS observation data this means that the on-board receiver does not form actual GNSS code and carrier pseudorange observations - as ground-based GNSS receivers do - but only collects the so-called intermediate frequency signals, and sends these to the ground station as part if the telemetry download. The signal processing from this raw binary data (level 0) to higher level observation data that can be processed by GNSS software (Level 1) is done on the ground. Again, this is done routinely by the Sentinel ground segments, but ESOC still needs the full capability to do the conversion from intermediate signals to Level-1, both in support of the ground segment and as independent check. We could say that in the case of the Sentinel on-board GNSS receivers, part of the receiver is flying on the satellite, and another part of the receiver is running on the ground.
Once that the Level-1 data has been produced - typically, within 24 hours after real-time - the precise satellite orbits for the Sentinel satellites are computed. For all operational Sentinel satellites this is done by means of conventional batch least squares estimation processes. If available, these processes also use DORIS observations and/or Satellite Laser Ranging observations in order to maximize the accuracy of the orbits. The use of multiple independent datasets also increases the confidence in the overall accuracy of the output orbit products, which has reached levels of around 1 cm for the Sentinel satellites. Reaching this high accuracy requires sophisticated state-of-art models for the satellites, the Earth, the atmosphere and any other detail that affects the tracking data analysis. In the case of GNSS, it also requires highly accurate input information for the orbits and clocks of the GNSS satellites themselves. The ESOC Navigation Support Office generates all these products routinely, at short latency and in a highly reliable operational computer environment.