Research Of Human-Computer Interaction System Based On Gesture Recognition And Robots Formation

With the continuous development of human-machine interaction,communication between people and the robot is also becoming more common.To replace complex cumbersome procedures,a more intuitive way of human-computer interaction has become the urgent needs of the people.Therefore,based on the visual identification of human-computer interaction has been gained comprehensive attention.The visual identification for the object of human recognition can be divided for face recognition and gesture recognition.For robot control,gesture control compared to the expression control has more practical meaning.Therefore,this article studies the robot formation control system based on gesture recognition.The research content is divided into two aspects,on the one hand,the human-computer interaction platform which is based on gesture recognition technology,on the other hand group of robot formation control method is proposed.For gesture recognition part,this article has chosen the most popular device camera named Kinect as image acquisition device.Compared with the ordinary camera,this camera device is able to provide users with the data of depth and the information of bone.And Microsoft also provides special development kit which is SDK.according to the characteristics of the device camera,this paper puts forward a kind of gesture recognition method based on depth information.Static hand gestures recognition can be divided into four steps,in turn,is a human body skeleton model,static gesture segmentation,gesture contour extraction and fingertip detection.Firstly we need to get the human body skeleton model,and then introduce the gesture segmentation process in detail,proposed a gesture segmentation method based on skin color detection combined with the depth of the image information.Secondly describe the process of extracting the contour with the method of Moore-Neighbor algorithm in detail.Then the convex hull is put forward to detect curvature,in combination with the hierarchical clustering,K-curvature algorithm and fingertip detection algorithm.The accuracy of the final recognition algorithm is verified by experiment.For multi-robot formation,this article selects the more widely used robot named TurtleBot as the research object.The characteristics of the multi-robot system are analyzed first,and then analyzes the possible problems in formation,and then analyzed the location problem of robot,and puts forward a visual locating method based on USB camera,which is accurate and reliable.Finally,we analyze the method of robot formation,and design several behaviors of the robot and obstacle avoidance method.In the last part of the article,we set up the human-computer interaction system based on gesture recognition and robots formation.Users can switch different gestures to control the robots' motion state and formation.several experiments have proved the accuracy and practicability of our human-computer interaction system.The human-computer interaction system we design in this paper has stronger practicability,can solve the problems in the process of human-computer interaction well.The proposed gesture recognition method is accurate,effective and has a good application prospect.Add gestures intervention in formation,can better improve the work efficiency of the robot.In this paper,the human-computer interaction system has a good application value in the field of service robots.