Evaluation of GNSS RF Signal Simulators and Receivers Based on Recorded Multi GNSS Signals in Scenarios of Traffic Telematics

R. Richter, B. Wolf, O. Michler

Abstract:	Evaluation of GNSS RF Signal Simulators and Receivers based on Recorded Multi GNSS Signals in Traffic Telematics Scenarios Global Navigation Satellite Systems (GNSS) such as GPS or GLONASS form a fundamental basis for the optimization of traffic processes over all modes of transport (road, rail, air and water). Reliable location, navigation and timing information of the operating traffic participants serve as a basis for that. In traffic telematics, algorithms to locate and track vehicles, people and objects play an important role. Thereby, both the choice of input data in the form of number and type of data to be fused as well as the functioning of algorithms themselves are of decisive importance. Future system concepts, such as fully automated driving, require the compliance of integrity, continuity and especially positioning requirements of the used traffic ways. For example this means a necessary lane-selective localization of vehicles on the road or high-precision positioning in inland navigation during a lockage. From the operational view field tests for GNSS measurements in rail traffic and inland navigation are difficult in planning, work-intensive in realization and expensive in implementation. Furthermore validation tests regarding receiver hardware and positioning or tracking algorithms require comparable and repeatable signal conditions. In general this task represents a challenge for GNSS measurements because satellite geometry, ionospheric conditions and environmental reflections change steadily. A solution for the above problems is the use of a modular GNSS radio frequency (RF) signal simulator for mobile field trials. It consists of a RF generator unit and a RF recorder unit. On the one hand the generator unit is able to output synthetic protocol-compliant GNSS signals and on the other hand it is able to replay GNSS signals reproducibly, which were recorded during field tests. Both methods offer advantages and disadvantages. The protocol-compliant generation of synthetically generated GNSS signals offers the advantage of knowledge about noise performance and artificially generated disturbances (e.g. multipath propagation). Therefore algorithms for their characterization can be developed, analyzed and assessed specifically to the error sources. The disadvantage in using synthetically generated GNSS signals is that error sources can be only limitedly modeled and integrated. In contrast recorded real measurement data contains all of the above mentioned error sources, but their magnitude and occurrence is not completely known, nor can the error sources be divided. In this context, the paper discusses the achievable reproducibility of GNSS signal generators used to record and replay field measurements in traffic telematics environments. In this regard, problem scenarios in the field tests are identified and classified in the second part of the paper. For this purpose, representative scenarios in the European rail environment, to be concrete in the railway network of the DB RegioNetz Erzgebirgsbahn (Germany) and Matterhorn Gotthard Bahn (Switzerland), were signal technically analyzed and archived in a signal library. This signal library contains problematic areas, such as one or two sided shading because of woods, rock faces, buildings of rail infrastructure or driving under bridges, in tunnels and gallery tunnels. Based on the GNSS signals generated by using this signal library, qualitative investigations were carried out in terms of achievable track sensitive accuracies. For that reason, several GNSS receivers of different performance and price levels were used. Furthermore the paper discusses the search for reasons and errors regarding the reproducibility of the calculated locating solutions. Those errors could be found in the subcomponents RF recorder unit, RF generator unit, GNSS receiver and also in the laboratory test setup. The discussion of statistically evaluated results in accuracy and reproducibility follows in the third part of the paper. To detect the causes of these divergences, the GNSS raw data, recorded during playback and converted to RINEX, was analyzed in terms of reproducibility and compared with the real recorded RINEX data during the field measurements. These investigations show that the distance measurements between the satellites and the receivers already have variances from one playback run to another. Consequently the receiver’s internal locating algorithms are not or not alone accountable for the variances in the locating solutions. Finally the inphase and quadratur (IQ) data, generated by the RF recorder unit during several playback runs, was evaluated to find out whether the RF generator unit inflicts divergences by itself. The paper closes with a conclusion and an outlook on further investigations to check if the used GNSS signal simulator is suitable for testing GNSS receivers and developing tracking algorithms in traffic telematic scenarios.
Published in:	Proceedings of the 26th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2013) September 16 - 20, 2013 Nashville Convention Center, Nashville, Tennessee Nashville, TN
Pages:	1881 - 1889
Cite this article:	Richter, R., Wolf, B., Michler, O., "Evaluation of GNSS RF Signal Simulators and Receivers Based on Recorded Multi GNSS Signals in Scenarios of Traffic Telematics," Proceedings of the 26th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2013), Nashville, TN, September 2013, pp. 1881-1889.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In