T. Donna Chen

Management of a Shared Autonomous Electric Vehicle Fleet: Implications of Pricing Schemes

The market potential of a fleet of shared autonomous electric vehicles (SAEVs) is explored by using a multinomial logit mode choice model in an agent-based framework and different fare settings. The mode share of SAEVs in the simulated midsize city (modeled roughly after Austin, Texas) is predicted to lie between 14% and 39% when the SAEVs compete with privately owned, manually driven vehicles and city bus service. The underlying assumptions are that SAEVs are priced between

Roles of Vehicle Footprint, Height, and Weight in Crash Outcomes: Application of a Heteroscedastic Ordered Probit Model

0.75/mi and

The effect of crash characteristics on cyclist injuries: An analysis of Virginia automobile-bicycle crash data

1.00/mi, which delivers significant net revenues to the fleet owner–operator under all modeled scenarios; that they have an 80-mi range and that Level 2 charging infrastructure is available; and that automation costs are up to

THE ELECTRIC VEHICLE CHARGING STATION LOCATION PROBLEM: A PARKING-BASED ASSIGNMENT METHOD FOR SEATTLE

25,000 per vehicle. Various dynamic pricing schemes for SAEV fares indicate that specific fleet metrics can be improved with targeted strategies. For example, pricing strategies that attempt to balance available SAEV supply with anticipated trip demand can decrease average wait times by 19% to 23%. However, trade-offs exist within this price setting: fare structures that favor higher revenue-to-cost ratios—by targeting travelers with a high value of travel time (VOTT)—reduce SAEV mode shares, while those that favor larger mode shares—by appealing to a wider VOTT range—produce lower payback.

Operations of a shared, autonomous, electric vehicle fleet: Implications of vehicle & charging infrastructure decisions

This study used a heteroscedastic ordered probit model to distinguish the effects of vehicle weight, footprint, and height on the severity of injuries sustained by vehicle occupants while controlling for many occupant, roadway, and other characteristics. Model results suggest that the impact of physical vehicle attributes on crash outcomes depends on the number of vehicles involved and is typically more significant in one-car crashes than in two-car crashes. Although vehicles with larger footprints and shorter vehicles were estimated to reduce the risk of serious injury for their occupants in single-vehicle crashes, they appeared to be less crashworthy in two-vehicle collisions. Heavier vehicles were anticipated to be more crashworthy regardless of crash type. Under evolving U.S. fuel economy standards, moderate changes in the weights, footprints, and heights of light-duty vehicles are estimated to have a relatively small impact on crash severities, whereas other factors, such as seat belt use, driver intoxication, and the presence of roadway curvature and grade, are estimated to influence crash outcomes much more noticeably.

Carsharing’s life-cycle impacts on energy use and greenhouse gas emissions

This paper examines bicyclist, automobile driver, vehicle, environmental, and roadway characteristics that influence cyclist injury severity in order to determine which factors should be addressed to mitigate the worst bicyclist injuries. An ordered probit model is used to examine single bicycle-single vehicle crashes from Virginia police crash report data from 2010 to 2014. Five injury severity levels are considered: fatalities, severe injuries, minor or possible injuries, no apparent injuries, and no injury. The results of this study most notably found automobile driver intoxication to increase the probability of a cyclist fatality six fold and double the risk of a severe injury, while bicyclist intoxication increases the probability of a fatality by 36.7% and doubles the probability of severe injury. Additionally, bicycle and automobile speeds, obscured automobile driver vision, specific vehicle body types (SUV, truck, and van), vertical roadway grades and horizontal curves elevate the probability of more severe bicyclist injuries. Model results encourage consideration of methods to reduce the impact of biking and driving while intoxicated such as analysis of bicycling under the influence laws, education of drunk driving impacts on bicyclists, and separation of vehicles and bicycles on the road. Additionally, the results encourage consideration of methods to improve visibility of bicyclists and expectation of their presence on the road.