Kinect-based Service Robot For Desktop Cleaning

Along with the progress of the society and the development of science and technology, the types of robot increases, the technical level develops as well, more and more service robot appear in people’s daily life. The content of the thesis is developed based on "The National Natural Science Foundation of China" project from Tsinghua University. The thesis is talking about a system of service robot used for desk cleaning based on visual sensor. The aspects of the thesis are as follows:First of all, this thesis introduces the background and significance of the appearance of the service robots, and give the definition of the service robot both in western countries and our country. Then it describe different types of service robot especially the cleaning service robot.Secondly, this thesis made comparatively detailed introduction of all the parts of the facilities in the system. Kinect is the visual sensor of the system, and it was calibrated in order to obtain the coordinate of points in the manipulator coordinate system.Then background subtraction algorithm is used to obtain the contour of the object appeared in the scene.And in order to eliminate the influence of object motion process and the hand interference when putting the object, frame differential method is used.Thirdly, trajectory planning of manipulator is conducted after the target grasp point is determined. RRT(Rapidly-exploring Random Tree) algorithm is used as the path planning algorithm, and a kind of improved algorithm of the gradient projection method--the most like gradient method is used for the inverse computation.Finally, in order to prove this desk cleaning system can work effectively, lots of experiments were made, and the results indicated that this system can independently complete desktop cleaning tasks successfully.