Description
TitleData-driven cyber-physical systems for socially informed mobility
Date Created2021
Other Date2021-10 (degree)
Extent1 online resource (xvii, 156 pages) : illustrations
DescriptionWith the ever-increasing concerns for air pollution and energy security, in recent years, we are witnessing a rapid emergence of innovative mobility systems (e.g., electric vehicles and shared mobility like carsharing, ridesharing, and bikesharing). These mobility systems are believed to be able to make a clearer environment for us. In addition, the fast development of advanced sensing technologies and communication devices provides us a good opportunity to collect and accumulate long-term massive data from these mobility systems, which are essential for us to extract domain knowledge and understand their mobility patterns, charging issues, user behaviors, as well as the potential roadblocks from their promotion process. These understandings and knowledge based on real-world data can benefit different parties including governments, providers, drivers, and passengers/users, which cannot be achieved with only short-term small-scale data. For example, we can further enhance the mobility systems and provide better decision-making services after identifying some drawbacks of current mobility systems during their evolving process.
Hence, in this dissertation, we aim to understand innovative mobility systems and physical phenomena by data-driven long-term measurement and prediction based on large-scale data, and then manage and enhance these mobility systems by socially informed decision making. We propose a data-driven framework that technically integrates measurement, prediction, and decision making to achieve the above two objectives. In particular, we present three works for socially informed mobility: one work called FairMove for fairness-aware electric taxi fleet management, one project named Record for satisfaction-aware electric carsharing fleet management, and the last work called sharedCharging for collaborative heterogeneous fleets management. FairMove is designed under a set of findings obtained from a data-driven investigation and field studies, which aims to improve the overall profit efficiency and profit fairness of electric taxi fleets by considering both the passenger travel demand and taxi charging demand. In FairMove, We formulate the electric taxi (e-taxi) displacement problem as a Markov decision process, and then we propose a fairness-aware multi-agent actor-critic approach to tackle this problem. We implement and evaluate FairMove with real-world streaming data from more than 20,100 electric taxis, coupled with the data of 123 charging stations, which constitute, to our knowledge, the largest all-electric taxi network in the world. The extensive experimental results show that our fairness-aware FairMove effectively improves the profit efficiency and profit fairness of the Shenzhen electric taxi fleet by 25.2% and 54.7%, respectively. The Record system considers not only the highly dynamic user demand for vehicle repositioning (i.e., where to relocate) but also the time-varying charging pricing for charging scheduling (i.e., where to charge). In Record, we design a dynamic deadline-based deep reinforcement learning algorithm, which generates dynamic deadlines via usage prediction combined with an error compensation mechanism to adaptively learn the optimal locations for satisfying highly dynamic and unbalanced user demand in real time. We implement and evaluate the Record system with 10-month real-world electric carsharing data, and the experimental results show that our Record effectively reduces 30.2% of vehicle movements. sharedCharging is the first data-driven framework to coordinate the charging events of large-scale heterogeneous fleets with different operating patterns and social purposes, which aims to improve the overall charging efficiency by sharing charging resources. All the three works pave the way for future shared autonomous mobility systems.
NotePh.D.
NoteIncludes bibliographical references
Genretheses
LanguageEnglish
CollectionSchool of Graduate Studies Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.
Statistical Profile