L-Shape Fitting-based Vehicle Pose Estimation and Tracking Using 3D-LiDAR

Abstract

Detecting and tracking moving vehicles is one of the most fundamental functions of autonomous vehicles driving in complex scenarios, as it forms the foundation of decision making and path planning. In order to estimate the pose information of moving vehicles accurately, 3D-LiDAR is widely used for accurate distance data. This paper proposed a real-time tracking algorithm based on L-Shape fitting. The algorithm detects the corners of moving vehicles and uses RANSAC to take a limited amount of noisy data. In addition, a vehicle tracking system with multi-weight Rao-Blackwellized Particle Filtering (RBPF) is built upon the orientation estimation given by L-Shape fitting. The proposed algorithm is validated on the KITTI dataset and a manually labeled dataset acquired from an autonomous vehicle at Carnegie Mellon University. Furthermore, the proposed solution is implemented in an autonomous vehicle at Tongji University. The experiments illustrate that the proposed algorithm achieves real-time performance, mitigates the effect of noisy data and improves estimation accuracy.