Parameter Estimation of Non-Paved Roads for ICVs Using 3D Point Clouds

Road parameter estimation is important for intelligent and connected vehicles (ICVs) operating on non-paved roads as it may influence their path planning and motion control. This paper presents a method for the estimation of longitudinal slopes, lateral slopes, and roughness of non-paved roads using 3D point clouds. Firstly, the regions of interest (ROIs) of ground are extracted by rasterizing the point clouds with grids, and divided into blocks according to the densities of point clouds. Next, longitudinal and lateral slopes are estimated by calculating the angles between two preference planes fitted using Random Sample Consensus (RANSAC) and Least Squares. Finally, an index of roughness, which is similar to International Roughness Index (IRI), is proposed for road roughness estimation in different grids. Experimental tests on non-paved roads demonstrate that the proposed algorithm has satisfactory performance in terms of the estimation accuracy of road slopes and roughness.