Travel Efficiency Investigation: Unravelling Local and Global Insights via Taxi Trajectory Analysis

Abstract

Transportation issues have a significant impact on people's lives because they spend a significant amount of time commuting for either daily needs or entertainment. These issues can be associated with travel time, longer travel distance, and/or fuel consumption. Due to the global positioning system (GPS) enabled devices installed in these vehicles, enormous amounts of trajectory data have been collected over the last decade from travelling vehicles such as cars, buses, and taxis, among others. This data provides an excellent opportunity to trace vehicle movements in fine spatiotemporal granularity. Moreover, this data tackles many of the traffic problems, including bottleneck identification. Identifying traffic bottlenecks is essential in traffic planning it also aids in the prevention of traffic congestion. Traffic congestion begins with congested road segments in key locations and spreads to other parts of the urban road network, causing additional congestion. The problem investigated in this thesis is analysing the road network travel efficiency locally and globally to reduce travel times, minimising fuel consumption, energy demands, and making better use of existing infrastructure. In much of the current literature, the focus is often on either a global analysis, which identifies the most efficient trip destinations, or a local analysis, which identifies the cause of traffic anomalies or congestion. However, it is necessary to consider both of these scales in order to gain a nuanced understanding. Specifically, it is crucial to quantify the extent to which each individual road segment affects travel efficiency, both at a local and a global scale. In order to provide a comprehensive understanding of urban traffic data, this thesis integrates both local and global analyses. In local analysis, we dive deep into each trajectory, much like deep-sea exploration, to uncover reasons for inefficiencies by examining all combined road segments. Then we extend the analysis globally to understand the behaviour of each part on road networks and how it effects on other road parts. The local analysis of the road network explores the measuring of the travel efficiency for each single trajectory trip across numerous origin-destination (OD) pairs in an entire city. Moreover, the consideration of a low travel efficiency path rises a 1 question of exactly which road segment is causing low efficiency. So, local analysis aims to measure the travel efficiency for each path. Furthermore, the local analysis provides the road segment inside a particular path that is responsible for low travel efficiency. In contrast, a small set of road segments that affect globally in the congestion problem is known as global analysis in this thesis. The global analysis seeks to identify a major source of traffic congestion. The global analysis provides some important opportunities for furthering the understanding of the congestion value for each edge in the road network and provides the top-k congested edges that influence the greatest number of other edges in the road network with the highest influence value recorded. The highest number of the influence values proves evidence of the global congestion effect in the entire road network.