Agent-based simulation of city-wide autonomous ride-pooling and the impact on traffic noise

Abstract

Pooled on-demand services promise to provide a convenient mobility experience and increase efficiency of road transport. We apply an established ride-pooling algorithm within the simulation framework MATSim to an autonomous fleet serving almost 2 million requests in Munich. Two mode choice scenarios are implemented, one substituting all car trips by ride-pooling, another one with free mode choice. For both scenarios we compare a stop-based and a door-to-door service in terms of system efficiency and noise emissions, applying an updated noise prediction model in MATSim. The results contribute to the systematic analysis of ride-pooling and show the effects of the proposed policies and service designs, which are essential for an efficient system with low noise exposure. Replacing all car trips by a stop-based ride-pooling system leads to a drastic noise reduction in residential areas whereas door-to-door systems may even increase noise exposure due to additional pick-up/drop-off rides and detours.