DYNAMIC AND WEB-BASED 4D VISUALIZATION OF STREETSPACE ACTIVITIES DERIVED FROM TRAFFIC SIMULATIONS AND SEMANTIC 3D CITY MODELS

Abstract. Semantic 3D city models can serve as anchor points for different components of urban digital twins. In addition to static 3D models such as buildings, transportation infrastructure, vegetation, or city furniture, this can also include dynamic processes such as traffic movement or changing traffic signals. Integrating these aspects into a dynamic, realistic, and accessible 4D visualization presents a number of requirements and challenges, which are discussed. While the City Geography Markup Language (CityGML) is a well established OGC standard for modeling and exchanging semantic 3D city models, the Cesium Language (CZML) provides capabilities for visualizing time-dependant properties that can be displayed in the Cesium virtual globe. Results of the open-source microscopic traffic simulation tool SUMO include information on locations and orientations of vehicles, bicycles, pedestrians and other traffic members as well as traffic signal information and can be exported as a CSV table. In order to allow a 4D visualization of these simulation results, a scheme for deriving 3D + t CZML documents from data given as 2D + t CSV files is presented. Additionally, dynamically changing traffic lights are integrated. Based on data available for the city of Munich, a CityGMLcompliant streetspace model is generated. This enables the combination of traffic simulation results and semantic 3D city models within a common Cesium based web-visualization and allows a direct and platform independent access to realistic visualizations of streetspace activities. Since this often results in a huge amount of data, a spatio-temporal tiling strategy enabling the visualization of large CZML data is presented.