Design and image processing supported control of a 4-axis industrial robot arm

Industrial robot arms are widely used in areas such as automotive production, electronic assembly, food packaging, pharmaceutical industry, and logistics. With the advancing software and hardware technologies, numerous studies are being conducted to enhance the control and programming ease of industrial robot arms. These studies utilize various sensors, algorithms, and even contemporary methods like artificial intelligence for the control and programming of robots. The aim of these efforts is to make industrial robot arms work more precisely and efficiently, thus improving production processes. In this thesis, the image processing-based control and programming of a four-axis industrial robot arm developed within the scope of the thesis have been carried out to contribute to the stated goals. The data obtained from the images through the pose estimation method have been processed for use in determining the joints of the human body and controlling the axes of the robot arm. The MediaPipe library, developed for pose estimation, has been used to predict the positions of human body joints from the images. The control of the industrial robot arm has been achieved through a microprocessor card developed within the scope of the thesis, using pose estimation data. The working accuracy of the system has been analyzed with tests conducted at different distances between the camera and the human.