Enhancing grasping capabilities through a three-finger gripper with full degrees of freedom

The paper presents the development and successful implementation of a three-finger robotic gripper with six degrees of freedom, aiming to enhance robotic dexterity and manipulation capabilities by replicating human finger movements and enabling advanced grasping strategies. The robotic fingers 1 and 2 design emulated human finger flexion/extension and adduction/abduction motions using tendon-driven and four-link mechanisms, respectively. The primary focus was on robotic finger 3, which mimicked the human thumb's opposition motions and featured a remarkable 145-degree rotation capability along the gripper base. This rotational ability allowed the gripper to grasp objects in multiple orientations. The system uses vision-based control and image processing techniques to empower the gripper with object recognition and precise positioning of the robotic fingers. Image processing algorithms enabled real-time decision-making, adapting the gripper's strategy based on the object's size, shape, and orientation. For the experiment, nine objects of different shapes and sizes were selected. The objects were analyzed through the image processing algorithm, resultingin grasping information with approximately 90% precision. The experimental phase demonstrated satisfactory performance in handling objects with diverse geometries.