GNSS positioning in dense urban areas is poor due to blockage and reflection of signals by the surrounding buildings. This project is investigating the use of 3D city models to improve the positioning accuracy. A new positioning technique, known as shadow matching, has been conceived and tested. This determines position by comparing measured and predicted satellite visibility and is used alongside conventional techniques. Using shadow matching, you can determine which side of a street you are on when conventional positioning is out by tens of metres. Shadow matching has been demonstrated using both recorded GNSS data and in real time on an Android smartphone. Current research is focused on improving the shadow matching algorithm, assessing its performance under a wide range of conditions, and extending it to multi-epoch positioning. Future work will combine shadow matching with other positioning techniques. Project duration: 2010-2014; Sponsor: UCL and CSC; Researcher: Lei Wang.