Coordinates: A resource on positioning, navigation and beyond

The experiment was performed in the GATE Base Mode and Extended Base Mode as well as in Virtual Satellite Mode, where the GATE/Galileo signals are simulated as they were transmitted from orbiting satellites. In fact of course the signals are transmitted by the earth fixed transmitters, so that signal fading and multipath effects, due to building and the landscape, are still present. The viewgraphs in Figures 4 show the position solutions on L1 frequency.
The observation time for the static measurements was about 5 to 10 minutes for each GATE mode. The GATE position accuracies (√2 σ) for these measurements are below 10 m for all three modes.

Dynamic field tests
Several tracking / positioning tests with the GATE system under dynamic conditions with a speed of up to 100 km/h were performed. The standard dynamic tests cover low dynamic conditions with an averaged speed of about 30 to 50 kilometers per hour. A sample track of a test drive in the EBM mode is presented below in the left-hand figure. High dynamic tests can only be performed at a section of the road B20, which passes the eastern part of the GATE area in north/ south direction, where a good visibility of the GATE transmitters is available. The road B20 is the only one in the GATE area where it is allowed to drive at 100 km/h. As starting point a dedicated position at the roadside was selected where all 6 GATE transmitters could be tracked well. After a short time of static positioning with all 6 signal sources – to make sure that the GATE receiver is in a well-defined starting position with stable tracking – the test car was accelerated rapidly up to a velocity of more than 90 km/h.

The sample results of a test drive in the GATE mode VSM are presented in fig. 5.

The dynamic positioning tests of the GATE receiver with Galileo signals in the GATE test area gave proof of the operational capability as well as the performance of the receiver and the whole
system also under dynamic conditions in all three GATE modes. This was evaluated not only for the receiver in uniform motion but also particularly with regardto significant accelerations of the receiver.

The acceleration values during the relevant parts of the test runs were in the range from about 15 to 20 seconds for speeding up from 0 to about 100 km/h. Also under these conditions regular Galileo position updates were obtained. Slight outages in the position solution, as seen in Figures 6 are the result of a conventional epoch by epoch data processing of the GATE User Terminal Software. For estimation of the position solution a standard leastsquares approach is used to get unfiltered solutions for each single epoch. A Kalman- Filter or dead-reckoning algorithm as it is implemented in common low cost GPS receivers would smooth such outliers in the position solution. For the position estimates illustrated in this paper even no carrier smoothing was applied to smooth the pseudo-ranges obtained from the code measurements from only data channels. Hence, any degradation of the measurements’ quality due to e.g. signalshading, as it is the case in the wooded part of the road in the fig. 5, strongly effects the quality of the position solution.

Regarding the illustrated tests in this paper it should be pointed out, that in all operational modes of the GATE system the receiver has at maximum six GATE transmit stations in view. Due to the vegetation and housing in the GATE test area and the low elevation angles of the transmitters in view at the user receiver position, shading of lines of sight to the transmitters occurs very often, while moving through the test area. Position outliers are often caused due to heavily degraded HDOP values, especially in VSM mode, when the remaining (not shaded) lines of sight represent a satellite constellation where the satellites and the user form a polyhedron with a very small volume (e.g. the satellites are situated nearly in one line from the users point of view). To a certain extent such scenarios can be evaded by an elaborate configuration of the virtual satellite constellation to be applied.

However, it is not possible to completely avoid such cases, because this would result in too frequent PRN switches of the transmitters. A PRN switch will decrease the number of potential available measurements at least for the time that is needed to receive the whole navigation message of the “new” satellite (at least 50 seconds for F/ Nav and 15 seconds for I/Nav messages).