Examining the effects of Automated Mobility-on-Demand services on public transport systems using an agent-based simulation approach

Several cities worldwide have been attempting to adopt "car-lite" policies to reduce traffic congestion and urban pollution. In addition to measures such as re-designing neighbourhoods and encouraging active modes, much expectation has been placed on the adoption of new and innovative modes, particularly shared AVs, or Automated Mobility-on-Demand (AMOD). Although the realisation of AMOD appears to be imminent, little is known about its potential effects on current transport systems. In this paper, using SimMobility, an agent-based microsimulation platform, we explored the impact of AMOD on public transport (PT). Two AV adoption scenarios were simulated: (1) "Partial Automation" where AMOD is introduced alongside existing modes, and (2) "Full Automation" where the use of private human-driven vehicles is prohibited upon the implementation of AMOD. We found that, compared to the base case (where there is no AMOD), the share of PT usage decreased significantly in the Partial Automation scenario whereas it increased in the Full Automation scenario. While the overall congestion level was reduced in the Full Automation scenario, road in the Partial Automation scenario tended to suffer from high travel demand. The increased demand for PT also prompts for a revision of current service schedules. The temporal and spatial analyses of PT demand between scenarios have brought some useful implications on the implementation of AMOD for urban and transport planners.