Wu, Fan (University of Alberta), Chen, Huiyu (University of Alberta), Qiu, Tony (University of Alberta)

Spatiotemporal Decomposition and Analysis of Vehicular Arrival on Green in A Signalized Arterial Using Trajectories

Scheduled for presentation during the Poster Session "Traffic prediction and estimation III" (ThAT13), Thursday, September 26, 2024, 10:30−12:30, Foyer

2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC), September 24- 27, 2024, Edmonton, Canada

This information is tentative and subject to change. Compiled on December 26, 2024

Abstract

Signal coordination is vital for smooth vehicle flow in corridors. Evaluating arterial progression typically relies on performance indices like travel time and delay, as well as tools such as the Purdue Coordination Diagram (PCD). However, these methods often overlook trajectory details and fail to capture continuous travel patterns. To quantify and visualize the arterial progression quality, this paper identifies spatially distinct vehicle groups using trajectory data and investigates the temporal travel patterns of each group. A velocity matrix of vehicles' travel along a corridor is extracted from the trajectories. Then a non-negative matrix factorization (NMF) is utilized to reduce the dimensionality of the velocity matrix, which is decomposed into two matrices representing vehicle travel patterns and their weights. The resulting eigentravel pattern matrix indicates traffic flow smoothness and identifies co-occurrences of arrivals on green (AOG) or red (AOR) signals. Additionally, a novel metric called the Proportion of Arrival on Green (P-AOG) evaluates progression quality under different traffic control systems. The trajectory data are generated by the simulation software SUMO, and the simulation is based on six consecutive intersections of 23rd Avenue in Edmonton, Canada. A case study involving a before-and-after comparison of two signal control strategies is used for demonstration.