Multiptah Channel Modelling for Satellite Navigation Systems - Mehrwegekanalmodellierung für Satellitennavigationssysteme

Kurzfassung

The first GPS satellite was launched in February 1978. The envisaged goal of the USA was to operate a globally available navigation system with an outstanding accuracy. Controlled by the US military, this accuracy was artificially degraded for civil users at the beginning. But techniques where found to overcome this degradation and even reduce the influence of several other error sources, leading to an accuracy which allows numerous civil applications. For example GPS makes road toll payment or avionic landing systems realizable without any ground based infrastructure today. However the required position accuracy is often not easy to guarantee. One error source still impedes applications where sub-meter accuracy is needed: multipath reception. Thereby signal echoes from objects in the near surrounding of a navigation receiver can cause positioning errors of several tens of meters. So far neither system design improvements nor advanced receiver techniques succeeded in solving this problem. The main emphasis of this thesis is to provide detailed models for the multipath environment in satellite navigation systems. The effect of multipath causing position errors in navigation receivers is well understood. It strongly depends on the power, excess delay and phase of the echo signals. Thereby especially the timely variation of these parameters is relevant. So far existing L-band models are not accurate enough. They are based on channel measurements of insufficient delay resolution and on physical simulations with a too low level of environmental detail. In this thesis two worldwide unique measurement campaigns are presented which are the basis to realistically and accurately model the multipath channel for satellite navigation systems. These models are necessary to simulate the critical environments for the test of new signal structures and for the development of multipath mitigation and estimation techniques for future highly accurate receivers. The first measurement campaign focused on land mobile car and pedestrian applications. A channel sounding signal was transmitted from a hovering Zeppelin towards a measurement van. The routes of the van were chosen in a way to allow a satellite elevation and azimuth dependent modelling of the multipath propagation channel. More than 60 measurements, each lasting for about 15 minutes were recorded in and around Munich, Germany, to gain statistically relevant data for urban, suburban and rural environments. The large measurement bandwidth of 100MHz and the high snapshot rate, as well as highly precise clocks, accurate tracking of Zeppelin and van movement and the synchronization of recorded data and video streams guaranteed most accurate and most reliable results. In a first post-processing step a super resolution algorithm allowed to resolve echoes from objects in the dimension of the wavelength. Further a novel algorithm was used to trace the delay change of these echoes. The interpolation of the measurement data finally allowed to extract the complex signals, thus providing power and phase trends of each echo. Although only a single receive antenna was used, the position of reflecting objects could be estimated from these echo signals, which in addition provided the angle of arrival information. From this data a novel combination of a deterministic-stochastic channel model was developed and implemented. Among the new features are the full dependency on the elevation and azimuth of the satellite and the modelled correlation between coexisting echo signals. Time variant receiver speed and heading input allows to simulate specific dynamics for various applications like for instance the stop of a car in front of a traffic light. The measured influence of houses, gaps, trees and lamp posts on the satellite signal and the changing number of coexisting echoes are taken into account, too. The channel model also comprises a synthetically generated scenery which can be parameterized to fit a given environment. Despite the very high level of detail, the implementation of the channel model is very efficient and for the first time allows the realistic simulation of multipath reception almost in real-time. The second measurement campaign explored the final approach flight phase of aircraft since this is the most critical scenario in aeronautical applications which are a driving factor in the design of Europe’s satellite navigation system GALILEO. Wrong assumptions in existing models became clearly visible, thus a new verified model is presented. The measurement revealed a fuselage reflection with an excess delay of only 1.5 ns, not critical in communications, but together with the reflection from the ground, dominating the error in high precision positioning. The hitherto dreaded wing reflection was found to be much weaker than expected at this wavelength. The new model offers the possibility to choose a large or a small plane type, taking different flight dynamics and different fuselage diameters into account. The deep insight into the propagation effects of space-earth radio links are the cornerstone to significantly reduce the error due to multipath reception. The cognitions obtained from the measurements already led to some promising new approaches for multipath robust receiver architectures which are also presented in this work.