A National Disaggregate Transportation Demand Model for the Analysis of Autonomous Taxi Systems

Abstract

Around the world every day, nearly one billion personal automobiles and other motorized vehicles travel millions of miles of roadway. The world's drivers spend trillions of dollars on these vehicles, which generally utilize automotive systems and technologies that were developed more than 50 years ago. The additional indirect costs of traffic congestion, pollution, and accidents add many tens of billions of dollars in expenses to the societal bill for personal transportation freedom. As with many other industries, new technologies have emerged with the potential to revolutionize old models and deliver greater individual and collective good at lower cost. These novel technologies can facilitate the creation of fleets of autonomously driven cars to serve individual transportation demand and replace the personal automobile. One of the more vexing challenges of the adoption of self-driving cars is not the technological solution, but the determination of the likely personal demand and expected patterns of use required to design the optimal self-driving automotive system. To that end, modeling existing personal travel behavior is fundamental to the creation of self- driving automotive systems. This thesis develops an existing disaggregate travel demand model, constrained for use within the state of New Jersey, into a daily transportation demand model for the entire United States. Data drawn from the U.S. 2010 Census, as well as other sources, are used to simulate 308,745,538 synthetic individuals with specific personal attributes and create the 1,009,322,835 automotive trips these individuals take across the U.S. on a typical work day. With precise spatial and temporal attributes that mimic actual personal travel behavior, these trips comprise a ready data set to analyze the efficacy of novel transportation systems. After determining where each individual wants to go, from where, and when, one can begin to engineer systems to serve this demand using autonomously driven vehicles. Specifically, this data set is designed to gain insight into a national system of autonomous taxis that could (a) match the comfort and convenience of personal cars, (b) exceed the accessibility of mass transit, and (c) deliver wide-ranging benefits such as alleviated congestion, reduced pollution, and increased vehicle safety.