The integration of city models and GNSS for the simulation and modelling of multipath and availability : paving the way for new applications

• Main Author: Bradbury, Jody Neil
• Published: University College London (University of London) 2008
• Subjects: 621.3841
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.503109

Positioning in the urban environment, using Global Navigation Satellite Systems (GNSS), is hampered by poor satellite availability due to signal obstruction and disruption caused by both manmade and natural features of the urban environment. In addition, range measurement to satellites for positioning and for navigation, is severely degraded by the presence of reflected and diffracted signals. The arrival and continuous enhancement of computerised geometric environment models, makes it possible to tackle these problems through mathematical modelling. Environment models have been used by architects and town planners for around thirty years. Within the last decade city model quality in terms of coverage, level of detail and geometric accuracy, has increased. Consequently, the number of applications for such models is continuously increasing and this ensures their continued availability and future usage. Currently, accurate prediction of satellite availability, particularly in cities, is extremely problematic. This thesis presents research undertaken into the use of environment models for the accurate prediction of satellite availability. A software simulator has been developed in order to accurately predict satellite availability, taking as input any environment model. Line of sight (LOS), Fresnel zone, diffracted propagation and reflected signal propagation models have been implemented in the simulator. Validation of each signal propagation model has been performed by comparison of observed data with corresponding simulated outputs. Successful validation of the Fresnel and diffracted propagation models, presented in this thesis, clearly indicate an enhancement over the use of conventional LOS propagation models alone. The reflected propagation model provides a potentially significant enhancement to the area of multipath modelling. It is certain that this work is of importance for mobile positioning devices. With the arrival of greater computational resources inside GNSS receivers, this research is likely to provide a means of reducing positioning errors due to multipath.