Position and attitude estimation and motion tracking using Kalman filter for autonomous docking system

摘要

A visual sensing system for an autonomous docking robot is developed in this study. The system recognizes a target object moving with no external forces and measures the position and attitude by a stereo vision. In order to reduce the measurement errors, Kalman filter is adopted. The position and attitude in the future is calculated from the estimated value by Kalman filter. Then, the motion of the target is tracked based on this estimation using binary images. The efficiency of the proposed method is demonstrated by an experiment, where the object under uniform motion and equiangular velocity motion is captured by the stereo vision.