Traversability Estimation Based on Occupancy Grid for Autonomous Driving in Extreme Environments

Abstract

In autonomous driving, traversability estimation is a technique for quantifying danger around a vehicle and is used for path planning.Although understanding the environment around the vehicle is essential for traversability estimation, it is impossible to estimate the blind spots of sensors, such as occlusions by obstacles. In this study, we effectively use point clouds measured by LiDAR to restore information on blind spots and realize traversability estimation.