Virtual Based Antropomorhic Gripper application for Automation Grasping Tasks

The following work presents the followed methodology for development a low cost Virtual Based Anthropomorphic Gripper application for Automation Grasping Tasks of common object geometries in a robotics work-cell environment containing a 6 DOF Industrial Robot and an Anthropomorphic hand based device attached to it. In order to comply the low cost development requirement, a set of parameters are defined so the programming and editing of the virtual environment is within this scope. Next, the study of the human hand, the mechanical hand and its kinematics allows the recreation of virtual objects resembling their behavior and visual looks. By studying human grasp postures it can be seen that each finger has a position from which a grasping table can be created, this information also works for an anthropomorphic gripper as they share similar anatomy. From that study three basic shapes are chosen to test the application, a spherical, cylindrical and planar ones, these are available on the graphic user interface, so it can be executed on the desire virtual device. For testing purposes, a real anthropomorphic device is virtualized and from the virtual grasps performed, an information file is created for been executed on the workcell The developed application allows the virtualization of a six to nineteen degrees of freedom either using commercial or open source 3D modeling software, initial applications are on offline grasp task planning. The program execution is operating system independent, which makes it portable as well of the creation and edition of 3D geometries. Future work can be for training and educational purposes and the research expanded for improving the grasping technique using supervisory systems and shape recognition.