Visual stabilization control on the wheeled mobile robot in a dynamic scene

Abstract

The invention discloses a design of a visual stabilization system of a wheeled mobile robot in a dynamic scene. The invention provides a novel method based on a monocular wheeled mobile robot, and therobot can perform visual servo under the condition that a visual target moves. Compared with an existing method, when the feature points are fixed, the mobile robot can correctly reach the expected pose, and it is difficult for the mobile robot to reach the expected pose after the feature points move. In order to monitor the movement of feature points, a monitoring camera is added in an experiment scene, and after the feature points move, the mobile robot can still reach an expected pose. And calculating the relationship between the moved feature points and the monitoring camera by utilizinga POSIT algorithm through the image of the monitoring camera, and further calculating the relationship between the corresponding coordinate system and the feature points, and calculating a rotation matrix and a translation vector between the current pose and the expected pose according to the real-time image feedback information by utilizing a coordinate system transformation rule. And a controller based on a polar coordinate representation method is used for driving the mobile robot to reach the expected pose. Simulation and experiment results prove that the method is effective and reliable.

Description

technical field [0001] The invention belongs to the technical field of computer vision and mobile robots, in particular to a vision stabilization control method for a wheeled mobile robot in a dynamic scene. Background technique [0002] In daily life, human beings receive most of the external information through the eyes, which shows the importance of visual sensors. Installing vision sensors on mobile robots can make them more intelligent and have a wider range of applications, such as intelligent transportation, deep sea exploration and home services. With the advancement of control theory and visual perception, more and more people are attracted to study robot visual stabilization control. For monocular vision systems, the lack of 3D models of visual features is an inherent problem, coupled with the non-integrity constraints of wheeled mobile robots, it brings greater challenges to robot vision stabilization. Some existing methods are to establish fixed feature points ...

AI Technical Summary

Problems solved by technology

However, the image feature points must remain unchanged throughout the visual servoing process. In practical applications, the feature points are sometimes moved due to work space factors and many other reasons.

Images