Pressure-based dispatch for shared autonomous vehicles

Principal Investigator:

Michael Levin, Assistant Professor, Civil, Environmental and Geo-Engineering

Project Summary:

Shared autonomous vehicles (SAVs), an autonomous taxi service, have been proposed as a possible replacement for personally-owned vehicles (Fagnant and Kockelman, 2014, 2016). Because the majority of taxi (or ride-sharing) operating costs are used for driver wages, companies such as Uber and Lyft are actively researching replacing their vehicle fleets with SAVs. By removing the driver, costs could become similar to that of owning a personal vehicle. If SAVs are used for a large proportion of daily travel, significant problems arise. First, how should SAVs be matched with passengers? Finding the optimal matching can be formulated as an NP-hard dial-a-ride problem, but it cannot be solved in real-time for large (e.g. 10,000) numbers of vehicles and passengers. Furthermore, large numbers of SAVs will contribute to traffic congestion, unlike current taxi services. Another question is how to route SAVs to minimize operating costs and passenger travel times.


Project Details: