Development of Cloud Action for Seamless Robot Using Backpropagation Neural Network

Abstract

This paper presents the cloud action model for a five DOF Seamless robotic arm using the inverse kinematics solution based on artificial neural network (ANN). Levenberg-Marquardt method is used in training algorithm. The desired position and orientation of the end effector is defined as the input pattern of neural network. In addition, we propose the cloud action which is the movement patterns of the robotic arm. The cloud action platform is created in order to perform the basic behaviors of the Seamless robot such as "Catch", "Approach", "Interest", "Look around", "Alert" and "Avoid" actions. Experimental results show the suitable structure of artificial neural network used for solving the inverse kinematics equation, and the testing points in the robot's workspace were verified with the robotic arm.