Object Perception Framework for Connected and Automated Vehicles: A Case Study

Connected and Automated Vehicle (CAV) technology plays an increasingly crucial role in advancing the development of transportation systems. However, due to the natural limitations of the occlusion and receptive range, the perception capability of existing CAVs can be further enhanced by involving emerging technologies such as roadside perception and multi-sensor fusion. To improve the perception capability of CAVs, we proposed a perception framework that consists of three perception subsystems: 1) a stock subsystem that represents the built-in perception system of the CAV, 2) an external subsystem that enhances the onboard perception system, and 3) a roadside subsystem that enhances the CAV perception performance based on cooperative perception. Different fusion schemes have been developed to efficiently fuse the information from these different subsystems. A case study has been conducted and the experimental results demonstrate that the proposed perception system can improve object detection performance by approximately 5% under
$100m$
, and approximately 10% at
$200m$
perception range.