Geometric Constraints and Rough-Fine Registration-Based Localization Method for Social Intelligent Transportation Systems

Localization and pose estimation algorithms play an important role in intelligent transportation systems (ITSs), as ITS need to accurately sense and understand the traffic environment to support autonomous navigation, traffic flow management, and autonomous material handling. This article proposes a pose estimation method in the front end of lidar odometry with geometric constraints. The proposed method can accurately capture the geometric information in the environment and ensure the effectiveness of the point cloud participating in the registration to improve the accuracy of registration. In the back end, an enhanced pose estimation strategy combining rough registration and fine registration is adopted to further improve localization accuracy. Comprehensive experimental results show that the proposed method achieves higher localization accuracy against other baselines, which also demonstrates that the proposed method can cope with challenging scenes such as complex road conditions and dynamic objects.