Intersection Mobility for Polaris Car

Autonomous Vehicles society is blooming and evolving as driving safety is expected to improve by autonomous vehicles (AVs) compared with human driving. Intersection mobility is one of the challenging tasks for AVs; thus, this paper proposes two algorithms. The first proposed algorithm uses a driving space to generate a motion trajectory based on LiDAR data and road geometry. The other algorithm provides C
²
smooth path Bézier curve by joining two cubic Bézier curve segments. Information pieces from Polaris car's position during turning are utilized to compute the intersection road's corner radius. The proposed algorithms have been tested in multiple simulation environments with different traffic conditions in Gazebo using Polaris car. Simulation results show that Polaris car can turn continuously and smoothly through the intersection square, demonstrating the effectiveness of the proposed algorithms.