MODELING SIMULATION AND ANALYSIS OF ROBOT GRASPING POSE FOR NOVEL OBJECTS

Abstract

Modern day robotics aims at bringing precision and motor dexterity of humans to machines. Many scientists are working on mimicking the attributes of a human in order to grasp different object. Initially, the robots were preprogrammed to hold different object but this method is not as effective against a novel object because the image of a novel object does not match with the pre-stored program data. Hence end effector will not be able to pick up the object or it may get damaged. In this paper, a new method is discussed on the dimension and the center of mass. The robot automatically calculates the grasping point and finds the suitable pose of the gripper to grasp the desired object. This proposed work of robot simulation is achieved using MATLAB 8.2 and Creo parametric 2.0. The robot model designed using Creo is fed as input into the MATLAB to generate control signals for the robot to grasp novel objects. By this the robot will automatically grasp a new unknown object by comparing the database already created and be able to handle the object dynamically.