A Cloud-Based Cooperative Perception System for Object Detection Using Roadside Cameras

Improving a vehicle's situational awareness is key to making fully autonomous driving possible. Cooperative perception systems are becoming increasingly essential in achieving that goal. This paper presents a novel approach that uses a single roadside camera, a 5 G communication link, and a cloud computing platform to enhance autonomous vehicle perception and decision-making. The paper focuses on evaluating the feasibility of detecting objects in the cloud using machine learning algorithms and then sharing the detection information with connected autonomous vehicles. While the initial tests were performed in a bench setup, the results showed high detection accuracy and low latency. These results highlight the potential of cloud-based roadside perception systems to effectively complement vehicle onboard perception systems.