Smart Rideshare Matching – Feasibility of Utilizing Personalized Preferences

Project Abstract

Current transport-share systems or carpooling typically rely on users to actively request or offer a ride and to coordinate the time and pickup location. Services such as Lyft and Uber have addressed this problem by using location to provide ride services that are convenient and on-demand. The on-demand and convenience aspects of transportation might also be the main reason behind using personal cars as they allow to combine commutes with other activities (e.g., picking up kids to and from school, running errands, going to off-campus meetings, etc.). This convenience, however, comes at a great personal and societal cost including traffic congestion, parking demand, stress, and health problems. Despite various agencies' incentives and discounts for ridesharing, this kind of service has not been widely used for obvious reasons mentioned above as well as hassled coordination, scheduling requirements, commitment, and having to actively request or offer rides. In this project, the research team proposes to conduct a case study using a university community to increase engagement in ridesharing in the UVA community by building a proactive context-aware matching and recommendation system that matches the community members based on predicted ride events inferred from their calendars and routines (e.g., shared time and location of events in Outlook calendar).

Universities Involved

University of Virginia 

Principle Investigators

Dr. Brian Park

Dr. T. Donna Chen

Dr. Afsaneh Doryab

Dr. Andrew Mondschein 

Funding Sources and Amounts

USDOT: $100,000

Start Date

September 1, 2023

Completion Date

September 1, 2024

Expected Research Outcomes & Impacts

The authors anticipate advances in modeling and prediction of personal activities from the sporadic calendar and GPS data. New algorithms for efficient context-aware peer-to-peer matching that will also promote the development of Ubiquitous Computing as well as Social Computing. Finally, the data collected from the application usage can inform more efficient and accurate proactive route guidance algorithms once implemented at scale.

Subject Areas

Connected and Automated Vehicles, Ridesharing,