Cloud-Based GNSS Solution Agnostic to any Telecom Network

Guillaume Carrie and Thierry Torlotin

Abstract: Objectives IoT use cases are very demanding in terms of low power and low consumption for design of the objects. The value of a solution is not only given by the raw data reported (pressure, temperature, ...) but more and more by the big data analysis, meaning the combination of all information reported by the sensors and their evolution over time. This leads, for example, to valuable information to predict meteorological phenomenon or the progress of refurbishment of road infrastructure. To that end, the location feature is now commonly embedded in any device whether they are in motion or static. Thus, it is taken for granted that the sensors data come with their position. In most of the case, the object does not require to deal with its position, this information is sent back to a back office that will process it. So, the computation of the position with low power profile and at low cost is a key feature. This can be done by the telecommunication networks with various level of accuracy. The Object can be located with the position of the antenna that received the message and augmented by measurement linked to the power (Radio Signal Strength Information) of the signal received or the time of arrival (TOA) when available. Advanced technics use trilateration to combine the Time difference Of Arrival (TDOA) from several antennas. These implementations are standardized by ETSI 3Gpp. Nonetheless, in terms of accuracy and precision GNSS remains the preferred technology. (It is the one selected to locate emergency call 911 in the USA) Cloud based GNSS computation has blossomed in various ways to get rid of the cold start effect especially because the common use cases require only several acquisitions per days and do not withstand to set the receiver in tracking mode. The implementations come under the name “snapshot GNSS receivers” or “cloud offloaded receivers”. A white paper published by GSA in 2020 describes the architectures taking benefits of the chipset or even using only an RF front end to cut costs. The most efficient solution has the objective to use only a snapshot of few tens of milliseconds of RF samples (see APOLLO research project funded by GSA ) without any dependency on the data contained in the navigation message or on the network. To reach this objective, a method must be developed to identify a coarse position of the object to resolve the ambiguity of the location of the code whether the object is in motion or not. The paper proposed by syntony explains the structure of the reception chain, the issues met and the final implementation with lab test results. Actual results The principle of the algorithm is to restrict to the bare minimum the signal processing on the object. The actual implementation uses as input a limited set of I&Q samples. Typical snapshot duration for GPS L1 C/A is 20 milliseconds. From it, a search in doppler and in PRN code phase is done and the estimated code phase for each PRN found are sent to the server with an associated coarse time tag T. The algorithm in the server identifies a set of candidates ‘nodes’ on earth having in visibility the satellites PRNs found at T by the object. The server then computes the difference between a reference code phase and all the code phases estimated by the connected object. A function F computes, for each candidate node, an error metric gathering for all the visible satellites the gap between on the one hand the theoretical code phase differences corresponding to this node and the list of visible satellites, and on the other hand the code phase differences computed from the RF snapshots. the node corresponding to the optimum of the function F gives the coarse position C. From test samples, first results on ‘roof antenna’ validate the process described above. With 7 to 8 satellites in visibility, a test run on 1000 snapshots taken over a period of 24 hours achieves a coarse position accuracy of 3,8 km for 99% of the points. This coarse position is then the starting point for a fine positioning algorithm able to achieve a metric accuracy. The proposed paper focuses on the coarse positioning which is the key enabler for Syntony solution, making it compatible with any telecommunication network, the fine positioning is thus out of scope of this contribution. Anticipated results Roof top antennas is a ‘best case’ scenario that matches some uses case. The work on going is to refine the algorithm to adapt to more harsh conditions as heavy masking (Urban canyon) while still using the same few amount of data (20 ms snapshot). There are several solutions either to keep a high number of PRN or to add other raw data from the signal processing (e.g. Doppler or geographical exclusion). Our early findings show that one can anticipate reaching a coarse position around 50 km at 99% that still meets the requirements to get a fix. Conclusion For the Iot market, the GNSS is still considered as the best technics should it matches the objectives of power consumption. The most effective solution is to take a snapshot of RF samples, process them to minimize the data stream over the air and compute the PVT in the ‘cloud’. To that end, the PVT algorithm needs a coarse position to solve the ambiguity on each PRN code phase. Although the coarse position could be given by the Cell tower of the network, the more network agnostic the solution is, the more scalable it will be. This study shows that the coarse position can be extracted from the few data sent by the connected Object. Key innovative steps These are : - Minimal amount of data processed by the object (20ms), - Determination of the coarse position without dependency on the given telecommunication network, - Robustness of the algorithm even with harsh condition (5 PRN detected)
Published in: Proceedings of the 34th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2021)
September 20 - 24, 2021
Union Station Hotel
St. Louis, Missouri
Pages: 1618 - 1646
Cite this article: Carrie, Guillaume, Torlotin, Thierry, "Cloud-Based GNSS Solution Agnostic to any Telecom Network," Proceedings of the 34th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2021), St. Louis, Missouri, September 2021, pp. 1618-1646.
Presentation Slides: ION Members/Non-Members: 1 Download Credit
Sign In