Parallel Processing of Large-Scale Point Cloud Data for Connected and Automated Mobility – A Survey

In the rapidly advancing fields of autonomous driving, advanced driver assistance systems (ADAS), and intelligent transportation systems (ITS), the use of light detection and ranging (LiDAR) sensors is essential for accurately detecting distances to surrounding objects. However, single LiDAR sensors face limitations such as occlusions caused by surrounding objects. To address these issues, cooperative perception technologies that utilize point cloud data from multiple vehicles have been explored. Despite these advancements, efficiently processing the large-scale, complex point cloud data generated by multiple vehicles remains a challenge. This paper surveys recent developments in parallel processing techniques for managing and processing large-scale point cloud data. These advancements are crucial for enhancing the performance of autonomous driving and ITS, enabling more efficient and scalable processing of point cloud data on the road.