A system of shared autonomous vehicles for Chicago: Understanding the effects of geofencing the service

Abstract

With autonomous vehicles (AVs) still in the testing phase, researchers and planners must resort to simulation techniques to explore possible futures regarding shared and automated mobility. An agent-based discrete-event transport simulator, POLARIS, is used in this study to simulate travel in the 20-county Chicago region with a shared AV (SAV) mobility option. Using this framework, the effect of an SAV fleet on system performance when constrained to serve within geofences is studied under four distinct scenarios: service restricted to the city, to the city plus suburban core, to the core plus exurban areas, and to the entire region — along with the choice of dynamic ridesharing (DRS) versus solo travel in an SAV. Results indicate that service areas need a balanced mix of trip generators and attractors, and an SAV fleet’s empty VMT (eVMT) can be noticeably reduced through suitable geofencing and DRS. Geofences can also help lower response times, reduce systemwide VMT across all modes, and ensure uniform access to SAVs. DRS is most useful in lowering VMT and %eVMT that arises from sprawled land development, but with insufficient demand to share rides, savings from the use of geofences is higher. Geofences targeting neighborhoods with high trip density bring about low response times and %eVMT, but fleet sizes in these regions need to be designed for uniformly low response times throughout a large region, as opposed to maximizing vehicle use in a 24-hour day.

Joint Econometric Model Framework for Transportation Network Company Users’ Trip Fare and Destination Choice Analysis

“…From Table 2, it is evident that trip fare is likely to increase as the distance between origin of the trip and the CBD increases. The result represents the supply side challenge (or rerouting costs) for drivers to pick up riders away from the CBD (see Gurumurthy et al for similar findings) ( 56 ). The negative coefficient for interaction of distance variable and PM peak period indicates that, during the PM peak, the impact of distance from the CBD is moderated, possibly because of increase in expected supply for TNC.…”

Curb Allocation and Pick-Up Drop-Off Aggregation for a Shared Autonomous Vehicle Fleet

“…This restricted SAV service to a smaller region of the Austin area, while still allowing all other modes to serve the entire 6-county Austin area. Gurumurthy et al (2021) studied four different SAV geofences in the Chicago area, each serving an increasingly larger area and all also including DRS. By focusing on a small slice of the region and excluding sprawling development while still covering a large enough area to provide a balance between trip generators and attractors, fleet efficiency increased.…”