FTO-Sim: an open-source simulation framework for evaluating cooperative perception in urban areas

Abstract In urban areas, static and dynamic occlusions frequently obstruct the field of view and impair the reliable detection of vulnerable road users (VRUs), posing a major challenge to the safe deployment of connected and automated vehicles in complex urban traffic. To address this, the following paper presents an open-source simulation framework for evaluating cooperative perception under realistic occlusion conditions, tailored to assessing VRU safety in urban traffic. The framework uses SUMO for microscopic traffic simulation and a Python-based ray-tracing module, enabling explicit modeling, visualization, and evaluation of occlusion effects without requiring complex co-simulation frameworks. In addition to conventional floating car observers, the framework introduces a novel observer type, the floating bike observer, extending the scope of cooperative perception studies. Several evaluation metrics are implemented, including relative visibility, level of visibility, and VRU-specific detection rates, which enable the systematic assessment of spatial perception coverage, detection reliability, and safety in critical road user interactions. The framework is fully open-source and complemented by simulation examples that replicate published studies, ensuring transparent validation of previous results while providing a basis for adapting it to new research questions in cooperative perception.