Characterizing the urban freight system and the supporting infrastructure network

Abstract

The urban freight system is essential to today's economy and cities' livability. In the last decade, densification, the growth of e-commerce, and the changing mobility ecosystem have amplified commercial vehicles' challenges in navigating the city streets or finding adequate parking space. Much of the current research and transportation planning efforts at the urban scale have focused on passenger mobility, giving little attention to commercial vehicle flows and their parking behavior. Furthermore, collecting freight data at the urban scale is challenging as this sector is fragmented, and its operations are complex, fast-changing, and heterogeneous. This results in local governments having limited insight into urban commercial operations patterns when developing appropriate and data-driven initiatives and policy measures. In response to this urban challenge, this research focuses on the need for cities and researchers to collect comprehensive and high-quality data; and develop evidence-based knowledge about urban commercial operations and their supporting infrastructure. This dissertation combines empirical case studies and analytical research focusing on two main aspects of urban commercial vehicle operations: on-street parking and traffic flow. First, it documents and analyzes commercial vehicles' parking patterns around five prototype buildings in the Greater Downtown area. Second, it develops and implements a new comprehensive vehicle classification system for collecting urban traffic flow data, focusing on the urban commercial fleet's heterogeneity. Third, it examines the effectiveness of the clustering technique, i.e., K-Means and Hierarchical Clustering, for identifying subgroups of CV traffic daily profiles and helps evaluate what vehicle and direction features may influence the CV Daily Flow patterns. Time-related features have the largest influence on daily temporal variations. Then, the resulting clusters are displayed and evaluated spatially to perform a spatial interpretation of commercial vehicle traffic patterns to evaluate the underlying feature relations among road network attributes and typical traffic flow patterns. As local conditions can significantly affect traffic patterns, Seattle's specific cluster results are then translated to "Typical" CV traffic patterns that can help draw insights into the variations of urban CV traffic flows. This effort will pave the way to further studies in Seattle and other cities aiming to compare and validate the results.