Developed and delivered a radar-based real-time object detection and risk assessment feature for the OpenHUTB/nn repository, focusing on enhancing driverless car safety metrics within simulation environments. The work involved integrating radar measurements with the existing object tracking pipeline, enabling accurate distance estimation and hazard scoring for detected objects. Leveraging Python and applying techniques such as Kalman filtering and computer vision, the feature supports real-time risk assessment and decision-making. The implementation provided an end-to-end solution for evaluating safety metrics, combining machine learning and data visualization to improve the simulation’s ability to assess and respond to potential hazards in real time.