Automated generation of traffic signals and lanes for MATSim based on OpenStreetMap

Abstract

Abstract OpenStreetMap (OSM) offers information on traffic networks like the position of intersections and streets, road categories, speed limits, number of lanes etc. Based on this information, network input data for transport simulations, e.g. the agent-based transport simulation MATSim can be generated. A MATSim network is modeled as a graph with nodes (intersections) and links (streets). Previously, data about traffic signals and lanes could only be provided to MATSim simulations by manual data preparation or had to be emulated by reducing flow capacities of links that lead towards signalized intersections. To be able to model inner-city traffic more realistically, this study extends the input generation of MATSim. Information about traffic signals and lanes is extracted from OSM and automatically processed to create signal and lane input files for the simulation. Processing data about traffic signals and lanes is required to optimize fixed-time plans, as well as to develop and validate algorithms for traffic-adaptive signals. This study provides the basis for further development of traffic lights in the transport simulation MATSim. An application of a MATSim simulation including traffic signals and lanes in the Berlin metropolitan area verifies the developed method.