A research team from Bristol, United Kingdom (UK) utilized the Simulation of Urban Mobility (SUMO) traffic generator to microscopically model and control each vehicle and traffic light. The SUMO model was accompanied by a MATLAB framework that automated the simulation steps, enabling easier generation and validation of different scenarios. The MATLAB framework worked in two phases. At first, the team generated the required structures and files for SUMO, based on several user-defined parameters. Later, using TraCI4Matlab framework, different scenarios were executed in a paralleled fashion. The research team also used the Intelligent Driver Model (IDM) as their car-following model within the simulation for all vehicle types.



The effect of the road layout on traffic flows was assessed for five cities: Manhattan, Paris, Berlin, Rome and London. Uniform scenarios for all cities were handled by working with demographic information about with the multi-modal public transportation systems in each city, parking availability, and the design of pedestrianized streets or cycling paths, but this data were not complete enough so they used different vehicle densities in each city to support the simulation of different traffic patterns. Key performance metrics evaluated included traffic congestion, average speed and average trip time.



FINDINGS



CAVs have the potential to decrease average trip times by up to three times under heavy traffic conditions. The numerical studies show that the traffic congestion can be reduced by up to a factor of four, while the average flow speeds of CAV group remains closer to the speed limits and can be up to 300 percent greater than the human-driven vehicles



Traffic situations indicated that even a small market penetration of CAVs can have a substantial net positive effect on the traffic flows.