Dynamic Pose Estimation Based On 3D Point Cloud

LiDAR(Light Detection And Ranging)is is a kind of sensor with small size,light weight,high resolution and strong anti-interference ability.With the increasing performance,its application value is increasing.Currently the world space powers are in the vigorous development of the rendezvous lidar technology.In this paper,the application of Li DAR in the final approach stage and docking stage is studied.Based on the 3D point cloud data,the position and pose of the space station are estimated automatically.The main research contents are as follows:1.A simulation system based on Gazebo 3D simulator is built for the attitude estimation.The system can simulate the motion mode of the space station and realize the collection,transmission,reception and processing of 3D point cloud data of the space station.2.By using ICP algorithm and RANSAC algorithm,the pose of the space station is estimated.The accuracy and running time of the method are analyzed,and the effectiveness of the algorithm is verified.3.Aiming at the problem of slow speed of ICP algorithm,a coarse registration method based on RANSAC algorithm is proposed.Experimental results show that the proposed method can effectively improve the processing speed and precision of ICP algorithm,and has a certain application value in engineering application.4.The accuracy of ICP algorithm will decline because of the decrease of the number of points in the point set in the second half of docking stage of space rendezvous and docking,this paper uses RANSAC algorithm to estimate the position of the space station in the second half of the rendezvous and docking stage.Experimental results show that the RANSAC algorithm is more accurate than the ICP algorithm,and the operation speed is faster than the ICP algorithm.