Trajectory Analysis at Intersections

In this thesis, we focus on trajectories at different intersections with various regulated types (traffic-light-controlled, priority/yield- controlled, uncontrolled) and test some methods to detect and recognise movement patterns, in terms of their geometrical and spatio-temporal components. That is, in the first case the geometric paths that vehicles follow while driving through the intersections and in the second case the way that they follow these geometric paths. For this scope, Machine learning methods such as clustering techniques are being used, and the performance of some known techniques and algorithms for similarity measurements (DTW, Hausdorff and Fréchet distance) and clustering (Affinity propagation and Agglomerative clustering) are compared based on the clustering results. The movement behaviour observed on different junctions is therefore analysed with clustering techniques, identifying some differences in the the movement patterns (speed and time) according to the regulation type that intersections use for regulating their traffic. These results can be useful for intersection categorisation according to traffic regulations, that can enrich modern maps with additive information (traffic signs, traffic lights, etc.).

Keywords: trajectory analysis; junctions; intersections; clustering; DTW; Hausdorff distance; Fréchet distance; Agglomerative clustering; machine learning; pattern recognition; movement behavior; GPS tracks; traffic regulations; traffic rules; traffic lights; traffic signs.