Using convex hulls with octree/voxel representations of point clouds to assess road and roadside geometric design for automated vehicles

Abstract

This paper aims to quantitatively assess the geometric design of road and roadside infrastructure for Autonomous Vehicles (AVs) using point cloud data while addressing the limitations and recommendations of previous studies. This research utilizes an ultra-dense point cloud to represent a digital twin of existing roads and simulates a set of AV sensors and phantom targets on both the road and roadside areas within this environment. A quadratic surface model, convex polyhedra/hulls with an octree/voxel data structure, a semantic segmentation variant of PointNet++, and YCbCr colour spaces were used by us to map road/roadside design readiness measures. Using the developed method, several case studies are presented to identify locations with substandard design conditions for AVs. This work can help infrastructure operators, and AV professionals make data-driven decisions regarding smart physical and digital infrastructure investments.