Differential impacts of autonomous and connected-autonomous vehicles on household residential location

Abstract

High market penetration of autonomous vehicles (AVs) and connected-autonomous vehicles (CAVs) is expected to impact transportation network performance, which is an important determinant of residential location decisions, especially for households who commute to work by personal vehicle. This study examines and compares the impacts of privately owned AVs and CAVs on the location and commute characteristics as well as the spatial distribution of households within the Triangle Region in North Carolina. A Mixed Multinomial Logit model is developed using recent household survey data to capture household preferences. In addition, the region’s travel demand model, the Triangle Regional Model, is used to predict the network-level impacts of AV and CAV adoption, and cluster analysis is conducted to explore how network performance changes vary with transportation demand and supply zone characteristics at a local and regional level. Residential location patterns are predicted for a number of AV and CAV scenarios for the year 2045 using the outputs of the econometric analysis and the Triangle Regional Model. We find that extensive adoption of private CAVs improves network conditions and encourages households to live farther from work, leading up to a 5.6% increase in suburban and rural households that commute to work by personal vehicles. A high market share of AVs is associated with deteriorated transportation network performance and up to a 2.8% increase in urban households. Results vary by market penetration rate of each technology, mix of AVs, CAVs, and human driven vehicles in the traffic stream, and fuel type (conventional-fuel versus electric vehicles).