Over the past decades, there has been some interest in performing data wipe-off for processing GNSS signals (e.g. GPS L1 C/A) as a pilot signal. Knowledge of the navigation data allows a receiver to preemptively remove it, transforming the data signal into a pilot signal, and thereby allowing the receiver to perform arbitrarily long coherent integration. The benefits of this include weak-signal processing and, in some cases, multi-path mitigation.
Techniques to do this currently include prediction of large parts of the navigation data, given knowledge of the data structure and observed patterns in the repetition of data (there are many patents on this technique); or the more simple approach of using a reference receiver with a clear view of the sky to observe and relay the data bits.
Both of these approaches have limitations: the prediction of navigation data might not be perfect, and there are always epochs where the data is either (predictably) unknown, or is guessed in error; while the use of a reference receiver incurs a latency.
One approach to circumvent both of these problems would be to listen to the navigation data while it is in uplink from the ground-segment. This would simultaneously avoid the errors incurred by guessing the navigation data; while also avoiding the delay associated with a reference receiver gathering the data.
Given the link budget for uplink stations, it is likely that sufficient power might be observed in the side-lobes of the uplink station, for a nearby ground-based observer to readily demodulate the data.
Can any researchers out there confirm whether this is possible, or if it is being done? Are there any obvious flaws in this approach – perhaps some that I am missing?
An exciting initiative but I think one important limitation would be the area that could be covered by such a service. GNSS satellites are sent up high to provide good coverage for the globe. Ground stations signals can only reach limited regions, I guess.
Thanks Malek, agreed - the data would need to be widely disseminated. The idea would be to capture the uplink and preprocess into the GNSS broadcast format (pages/words/subframes/etc..) and broadcast through a terrestrial network (even the internet) as a real-time service. Based purely on time-of-flight, the ground system would have at least a 120ms 'head-start' as the GNSS data need to travel to space, and then back. Assuming some other latencies on board the satellite, it might be possible to have the data at the receiver well before it is needed.
sounds interesting. but, I think this would be useful not because of the signal propagation time (120-130 ms) but because it takes so long (i.ie. few minutes ) to download the message via the satellite signal which might be obtained at much higher rate through internet or other networks.
I agree Malek, although the data required by a receiver to use GNSS positioning is not necessarily the same as the data required by the receiver to perform data wipe-off. For positioning, a receiver needs valid ephemeris (satellite orbit). This does not need to be the exact same data (bit-for-bit) as is broadcast by the satellite.
In fact, many assistance services have existed for years and they provide ephemeris over the network - often faster (e.g. GSM RRLP), more accurate (precise IGS products), or longer lasting (LTE). Practically speaking, if a network connection is available then there is no need to recover the navigation data from the satellite itself. Indeed, anyone can observe GNSS satellites (given enough ground stations) and produce their own orbit estimates – and many companies already do this.
However, if are receiver wants to perform data wipe-off to achieve improved sensitivity then it needs the EXACT same data as was broadcast by the satellite (even if it doesn’t use this data for orbit prediction). This exact data can only be sourced directly from the system – hence the interest in eavesdropping...
- Badges
- Science topic
- Similar topics
- Correction