Abstract: | Recently a routine study was made to determine the suitability of currently available RTK systems for the precise positioning of slow moving platforms. The study consisted of two closure tests over known baselines over a number of separate days on a large open playing field. The repeatability and absolute accuracy of the positioning systems were under assessment for various times of the day and differing satellite constellation configurations. During this feasibility study, an unexpected event occurred in one test which caused the computed position output to deviate to produce a 90m error in height for a period of 100 epochs. When reanalysing the data with another RTK system, the error excursion was merely 2m in height. The objective of this paper is to determine the cause of this unexpected excursion in computed output position and to compare the manner in which two commercially available RTK packages handle this erroneous data. The two systems under consideration were Novatel’s float solution RT20 system and GEONAV’S P-RTK integer ambiguity resolution system. The raw satellite observables were analysed and for approximately 100 epochs, 7 out of 8 range measurements were unavailable on the narrow correlator receiver. The standard deviations of the code phase measurements typically jumped from a steady 0.163m to 137.002m for the 4 epochs immediately after this erroneous event. Similarly for the carrier phase pseudorange whose standard deviations jumped from 0.008 cycles to 0.018 cycles. The event lasted another 100 epochs before the standard deviations in the range measurements fell to levels comparable to that before the excursion. It was seen that the maximum and first occurring excursion in the data was for PRNl5. By observing the satellite position at the time of the event it was noticed that PRNlS’s position corresponded exactly in azimuth to the metal hand rail underneath the roving GPS antenna, with a low elevation of 13”. Using the code minus carrier technique to observe multipath we saw that 6m error in range was experienced on the range measurement for PRN15 just before the signal become unavailable. Numerical modelling of this scenario shows that the signal strength of the field at the antenna for this satellite with the ground reflection is 20 times the incident field. It is thus thought that a multipath event has caused a constructive peak in power at the GPS antenna that has jammed the RP front end of the narrow correlator receiver. The performance of the two RTK positioning systems to this event were then compared. There was a 6 epoch delay from where the excursion was seen in the raw satellite observables to when the effect was experienced in the RT20 position output whence there is a 9Om offset in height. The PRTK system produces no position output from the epoch following the excursion, and then there is 15 epoch absence of position output. When the PRTK system comes back on-line it produces a position solution with a 2m offset in height on float ambiguities. The conclusion drawn during this analysis is that a very strong multipath reflection, from PRN15 and reflecting from the metal handrail has effectively jammed the GPS receiver. This produced erroneous range measurements on 7 out of 8 channels at the roving GPS receiver. The techniques with which the RT20 system and the PRTK system handle this multipath event were compared and contrasted. Recommendations are made as to which system is most suited to the precise positioning of slow kinematic/static platforms in harsh environments. Optimal failure modes for GPS positioning software operating under these adverse conditions are also discussed. The conclusion is that RTK software customised for the end user application is essential.Recently a routine study was made to determine the |
Published in: |
Proceedings of the 1997 National Technical Meeting of The Institute of Navigation January 14 - 16, 1997 Loews Santa Monica Hotel Santa Monica, CA |
Pages: | 261 - 265 |
Cite this article: | Walker, Rodney A., Kubik, Kurt, "Comparison of RTK Failure Modes for Receiver Jamming by Multipath," Proceedings of the 1997 National Technical Meeting of The Institute of Navigation, Santa Monica, CA, January 1997, pp. 261-265. |
Full Paper: |
ION Members/Non-Members: 1 Download Credit
Sign In |