A general framework for modeling shared autonomous vehicles with dynamic network-loading and dynamic ride-sharing application. (July 2017)

• Record Type: Journal Article
• Authors: Levin, Michael W.; Kockelman, Kara M.; Boyles, Stephen D.; Li, Tianxin
• Is Part Of: Computers, environment and urban systems. Volume 64(2017)
• Page Start: 373
• Page End: 383
• Publication Date: 2017-07
• Subjects: City planning -- Data processing -- Periodicals; Regional planning -- Data processing -- Periodicals; 303.4834
• DOI: 10.1016/j.compenvurbsys.2017.04.006
• Languages: English
• ISSNs: 0198-9715

Abstract

Abstract: Shared autonomous vehicles (SAVs) could provide low-cost service to travelers and possibly replace the need for personal vehicles. Previous studies found that each SAV could service multiple travelers, but many used unrealistic congestion models, networks, and/or travel demands. The purpose of this paper is to provide a method for future research to use realistic flow models to obtain more accurate predictions about SAV benefits. This paper presents an event-based framework for implementing SAV behavior in existing traffic simulation models. We demonstrate this framework in a cell transmission model-based dynamic network loading simulator. We also study a heuristic approach for dynamic ride-sharing. We compared personal vehicles and SAV scenarios on the downtown Austin city network. Without dynamic ride-sharing, the additional empty repositioning trips made by SAVs increased congestion and travel times. However, dynamic ride-sharing resulted in travel times comparable to those of personal vehicles because ride-sharing reduced vehicular demand. Overall, the results show that using realistic traffic flow models greatly affects the predictions of how SAVs will affect traffic congestion and travel patterns. Future work should use a framework such as the one in this paper to integrate SAVs with established traffic flow simulators. Highlights: We present a framework for modeling shared autonomous vehicles (SAVs) compatible with a general class of traffic simulation. We implement our framework using a cell transmission model simulator on a city network. SAVs can greatly increase congestion, and therefore road congestion should be included in all SAV models. We compare polynomial-time heuristics for dynamic ride-sharing and preemptive relocation for SAVs.