作者: Ke-Hu Yang , Wen-Sheng Yu , Xiao-Qiang Ji
DOI: 10.1007/S11633-012-0647-Z
关键词:
摘要: The rotation matrix estimation problem is a keypoint for mobile robot localization, navigation, and control. Based on the quaternion theory epipolar geometry, an extended Kalman filter (EKF) algorithm proposed to estimate by using single-axis gyroscope image points correspondence from monocular camera. experimental results show that precision of robot's yaw angle estimated EKF much better than given image-only gyroscope-only method, which demonstrates our method preferable way autonomous applications.