Pose Estimation of Moving Vehicles Based on Heuristic Rules for Autonomous Driving

Estimating the pose information on moving vehi-cles is one of the most fundamental functions of autonomous driving for detecting and tracking moving objects. The current methods are often based on the prior information of the point cloud distribution, however, the distribution in the actual scene changes dynamically, and is easy not to meet a priori. This paper proposes a 3D pose estimation method based on heuristic rules. In a single frame point cloud, the edge length and edge visibility are used to heuristically screen the candidate orientations established by the convex hull sampling method. On consecutive frame point clouds, tracking heuristics are employed to obtain smoother orientation estimation results. The experimental results on the KITTI dataset and the data from the actual scene of vehicles show that the proposed method reduces the average object-orientation error and improves the accuracy of estimation effectively.

[1]  Linglong Lin,et al.  Optimal Vehicle Pose Estimation Network Based on Time Series and Spatial Tightness with 3D LiDARs , 2021, Remote. Sens..

[2]  John M. Dolan,et al.  L-Shape Fitting-Based Vehicle Pose Estimation and Tracking Using 3D-LiDAR , 2021, IEEE Transactions on Intelligent Vehicles.

[3]  IEEE International Conference on Robotics and Automation, ICRA 2021, Xi'an, China, May 30 - June 5, 2021 , 2021, ICRA.

[4]  Yang Liu,et al.  Estimation of 2D Bounding Box Orientation with Convex-Hull Points - A Quantitative Evaluation on Accuracy and Efficiency , 2020, 2020 IEEE Intelligent Vehicles Symposium (IV).

[5]  Ming Yang,et al.  Robust Point Set Registration Using Signature Quadratic Form Distance , 2020, IEEE Transactions on Cybernetics.

[6]  Jie Li,et al.  Probabilistic 3D Multi-Object Tracking for Autonomous Driving , 2020, ArXiv.

[7]  2020 IEEE Intelligent Vehicles Symposium (IV) , 2020 .

[8]  Sendai,et al.  2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , 2020 .

[9]  Bin Yang,et al.  Vehicle Pose Estimation Based on Edge Distance Using Lidar Point Clouds (Poster) , 2019, 2019 22th International Conference on Information Fusion (FUSION).

[10]  Kaiqi Liu,et al.  Fast Dynamic Vehicle Detection in Road Scenarios Based on Pose Estimation with Convex-Hull Model , 2019, Sensors.

[11]  Yang Liu,et al.  BoxNet: A Deep Learning Method for 2D Bounding Box Estimation from Bird's-Eye View Point Cloud , 2019, 2019 IEEE Intelligent Vehicles Symposium (IV).

[12]  2018 IEEE International Conference on Robotics and Biomimetics (ROBIO) , 2018 .

[13]  Fa Wang,et al.  An Efficient L-Shape Fitting Method for Vehicle Pose Detection with 2D LiDAR , 2018, 2018 IEEE International Conference on Robotics and Biomimetics (ROBIO).

[14]  Sergiu Nedevschi,et al.  A Fast Ransac Based Approach for Computing the Orientation of Obstacles in Traffic Scenes , 2018, 2018 IEEE 14th International Conference on Intelligent Computer Communication and Processing (ICCP).

[15]  Fei-Yue Wang,et al.  MFR-CNN: Incorporating Multi-Scale Features and Global Information for Traffic Object Detection , 2018, IEEE Transactions on Vehicular Technology.

[16]  Hans-Joachim Wünsche,et al.  An Orientation Corrected Bounding Box Fit Based on the Convex Hull under Real Time Constraints , 2018, 2018 IEEE Intelligent Vehicles Symposium (IV).

[17]  Myoungho Sunwoo,et al.  L-Shape Model Switching-Based Precise Motion Tracking of Moving Vehicles Using Laser Scanners , 2018, IEEE Transactions on Intelligent Transportation Systems.

[18]  Xiao Zhang,et al.  Efficient L-shape fitting for vehicle detection using laser scanners , 2017, 2017 IEEE Intelligent Vehicles Symposium (IV).

[19]  Andreas Geiger,et al.  Vision meets robotics: The KITTI dataset , 2013, Int. J. Robotics Res..

[20]  Michael Himmelsbach,et al.  Tracking and classification of arbitrary objects with bottom-up/top-down detection , 2012, 2012 IEEE Intelligent Vehicles Symposium.

[21]  脇元 修一,et al.  IEEE International Conference on Robotics and Automation (ICRA) におけるフルードパワー技術の研究動向 , 2011 .

[22]  Quanshi Zhang,et al.  Moving object classification using horizontal laser scan data , 2009, 2009 IEEE International Conference on Robotics and Automation.

[23]  Yassine Ruichek,et al.  Representing and Tracking of Dynamics Objects using Oriented Bounding Box and Extended Kalman Filter , 2008, 2008 11th International IEEE Conference on Intelligent Transportation Systems.

[24]  Robert A. MacLachlan,et al.  Tracking of Moving Objects from a Moving Vehicle Using a Scanning Laser Rangefinder , 2006, 2006 IEEE Intelligent Transportation Systems Conference.

[25]  Mohan M. Trivedi,et al.  Moving shadow and object detection in traffic scenes , 2000, Proceedings 15th International Conference on Pattern Recognition. ICPR-2000.