Abstract
This research investigates the perception capabilities of vehicle-road collaboration under cloud control in intelligent transportation systems. It proposes a cloud-controlled sensing system that uses Dempster-Shafer (DS)-based evidence theory to construct an Evidence Occupancy Grid Map (EOGM). The system integrates local LiDAR scans into a global EOGM, which is then shared among vehicles via the cloud. Experimental results show that the cloud-based system improves detection performance compared to traditional vehicle or road-side detection, especially under varying penetration rates of self-driving vehicles.