FlexSense: Flexible Infrastructure Sensors for Traffic Perception

The future of transportation is connected, automated and roadside supported. Despite a booming number of test facilities being built around the world to conduct fundamental research on Intelligent Transportation Systems (ITS), there is a lack of variability in the observed scene due to current urban planning. To enrich the scenery for further technology development and to promote the application of initial research results, we develop FlexSense as a flexible test sensor system. The sensor system can be easily deployed in a wide variety of urban environments. It also meets the challenges of different weather and lighting conditions for robustness in application trials. Our mobile sensor setup leverages the development of infrared spectrum sensing and radar detection, providing both 2D image and 3D point cloud data as raw data outputs. In addition, we develop a software framework for environmental perception based on this sensor setup, including 3D object detection, multimodal sensor data fusion and multi-target tracking. We also provide a dataset collected with this sensor setup and the corresponding 3D bounding box labels to the research community. The dataset is publicly available and can be accessed via this website: https://fraunhoferivi.github.io/FlexSense/.