Integration of C-V2X with an Obstacle Detection System for ADAS Applications

Cellular Vehicle-to-Everything (C-V2X) technology is essential for enhancing road safety by connecting vehicles, roadside infrastructure, and networks. It improves safety by transmitting vital warnings and traffic updates that vehicle sensors might miss. This paper presents a novel C-V2X-based object detection system designed for intersections, leveraging cutting-edge 5G sidelink technology and advanced Deep Neural Networks (DNNs) deployed on edge computing units. The system utilizes Basic Safety Messages (BSMs), which convey timestamps and location data of detected objects, sent from a Roadside Unit (RSU) to On-Board Units (OBUs) via 5G sidelink. The proposed system was tested using an outdoor testbed at the GMMRC testing track at Kettering University. The system continuously monitors the intersection, with the RSU broadcasting detection results to nearby vehicles equipped with OBUs. Utilizing the depth perception capabilities of the stereo camera, the system accurately determines the distance of objects from the camera. Performance evaluations conducted through real-time experiments on the test track demonstrated the system's efficiency, achieving an average latency of just 1 millisecond in V2X communication across various scenarios. The 5G-enabled communication offers minimal latency, high data rates, and reliable connections, making it exceptionally well-suited for C-V2X applications. These results highlight the considerable potential of this technology for advancing vehicular communication and safety across diverse applications.