Enhanced Infrastructure-Enabled Cooperative Perception for Autonomous Driving

An enhanced cooperative perception system based on Roadside Units (RSUs) and Vehicle-to-Infrastructure (V2I) communication is proposed. The objective is to improve the perception capabilities of Connected Autonomous Vehicles (CAVs) by leveraging perception information obtained from RSUs. CAV’s perception performance determines the reliability of autonomous driving. In the proposed system, the infrastructure-enabled perception algorithm realizes wide-range and high-precision perception, while V2I communication transmits data collected by the RSUs to CAVs. In this way, the enhanced cooperative perception system can furnish CAVs with precise and prompt awareness of their Region of Interest (RoI), even when the RoI extends beyond the CAV’s perceivable range. The proposed system has been evaluated using a software-in-the-loop approach. Experiments have confirmed that the proposed system shows superiority in enhancing safety and efficiency for CAVs.