Research On Xylophone Playing Robot Based On Desktop Robotic Arm

With the continuous development of robotics technology and the strengthening of national strategic forces in science and technology warfare,the robotics industry has been flourishing in China,with new robots constantly being developed to serve all aspects of our work and life.Miniaturized robots have become a new favorite in the robotics application market because of their small size,light weight,and diverse functions.The expansion of the functions of these robots not only makes us feel the charm of industrial technology,but also gives us a different experience in many aspects of its application.Among them,the performance robot,as a kind of performance special robot,combining music performance and robotics,has attracted much attention.At this stage,there are few studies on this type of robot,and in view of this,this paper designs a robot with xylophone performance function based on a desktop-level six-axis robot arm.The specific research contents are as follows.First,the spatial trajectory of the six-axis robot arm is solved by using kinematics to establish the kinematic model of the robot arm,and the mathematical relationships between the positional matrix and the six joint angles are solved by solving the forward and inverse kinematics of the robot arm.The results of the forward and inverse kinematics derivation were checked by using simulation experiments,and the Monte Carlo method was used to solve the working environment of the robotic arm to obtain the range of motion of the robotic arm and complete the motion analysis of the robotic arm.Secondly,in order to complete the positioning of the xylophone by capturing images,camera calibration of the camera is required and the internal and external parameters of the camera are obtained.The camera arrangement of eye outside the hand is used to calibrate the position relationship between the robotic arm and the camera,and the hand-eye conversion matrix is obtained,which enables the vision system to provide the target point location for the robotic arm.For the recognition of xylophones in different environments,a key recognition method with HSV+ geometric features is proposed based on the color,shape and environmental characteristics of the keys,and the keys in the image are successfully recognized and the location coordinates of the center of the key surface are obtained.Then,the segmentation of the simple score and the extraction of notes were studied.According to the structural characteristics of the sketch,the sketch image is segmented and extracted by line using the horizontal projection method,and the Hoff transform is used to find the bar lines in the line extraction map,and the line sketch image is segmented according to the bar lines and divided into a number of bar images,and then the contrapuntal lines in the image are removed,and all the notes in the bar image are segmented into single notes by the de-edge operation,which completes the segmentation of single notes from the sketch image.In order to extract the note information from a single note image,this paper proposes a single-input,three-output note recognition method based on convolutional neural network,in which three numeric,pitch,and time-value sub-networks are constructed in the same network model,and the trained model is used to complete the extraction of note information.Finally,the xylophone playing platform of the robot arm is built,and the trajectory planning of the playing action of the robot arm is carried out,and the playing performance of the robot arm is tested and analyzed through experiments.