Robust Vehicle Localization Using Entropy-Weighted Particle Filter-based Data Fusion of Vertical and Road Intensity Information for a Large Scale Urban Area

Abstract

This letter proposes a robust vehicle localization method based on a prior point cloud in urban area. The high resolution point cloud collected six months ago is provided from Singapore Land Authority around One-north area in Singapore, because the data are outdated there are many changed aspects of the environment such as redrawn road markings, construction areas, and changing tree shapes. In response, this paper proposes a novel fusion algorithm based on a particle filter using vertical and road intensity information for robust localization. Whereas the state-of-the-art fusion algorithm focus on optimization of the vehicle pose based on multiple measurements, the proposed method estimates a robust vehicle pose by considering the reliability of each feature from the prior map. We also propose an efficient management strategy of a grid map that includes multilayer vertical and road intensity information for real-time operation. A sensor system equipped on a vehicle consists of 32 channels of 3D Light Detection And Ranging, IMU, wheel odometry, which are used for the proposed algorithm. RTK-GPS, wheel odometry, and iterative closest point algorithm are utilized by using a graph-structure optimization method in off-line for estimating ground truth. The total data set for the demonstration is collected in a 19.9 km way in an urban area. The proposed approach has successfully performed an autonomous vehicle driving in an urban area.