A Framework for Integrating Automated Vehicles with Public Transit: Using Differences in Rider Demand and Social Networks

Public transit agencies operate under a fundamentally different mandate than private mobility providers. Where private operators optimize for demand, transit agencies exist to guarantee access for those who cannot drive, cannot afford alternatives, and live in places where demand alone would never justify service. This distinction matters because the rapid expansion of autonomous vehicles in cities like San Francisco, driven by private Transportation Network Companies, is optimizing mobility for those who already have options, while leaving behind those who depend on the public network. Yet the research for public institutions can harness new vehicle technologies in service of equity rather than profit. This proposal builds on that premise. The research team introduces a mobility science framework that identifies targeted trip-chaining between AVs and fixed-route transit by leveraging two newly available data sources: mobile phone-derived travel records that provide unprecedented resolution into individual trip characteristics, and inter-zipcode social network signals. This offers an additional behavioral lever for ridesharing eligibility. Building on an existing matching and pooling algorithm, the proposed framework integrates a logit-based mode-choice module and a congestion-aware dynamic traffic assignment module to enable public transit agencies to guide high-level system design while encouraging equitable deployment of autonomous mobility on demand (AMoD).