This thesis presents a novel robot hand developed to have closer dexterity of human's precision grip, while not requiring the complex control algorithm. The robot hand has total 17 DOF (Degree Of Freedom) and includes 14 back-drivable actuation system...
This thesis presents a novel robot hand developed to have closer dexterity of human's precision grip, while not requiring the complex control algorithm. The robot hand has total 17 DOF (Degree Of Freedom) and includes 14 back-drivable actuation systems. A new performance index for robot hands with the idea of mimicking the human hand's precision grip, the procedure of kinematic optimization of robot hand using this performance index, and the procedure of experimental evaluation to see the grasping capability of the robot hand are explained.
First, to develop a robot hand which can grasp various objects stably and carry out fine manipulation on those objects like a human hand, an investigation and analysis are performed on the human hand's skeletal structure and tendon-ligament system in Chapter 2. In this process, important differences between the human hand and existing robot hands are discovered, which are the nature of the DIP and PIP joint coupling, a different number of DOF in the thumb, and the existence of the IMC joints in the palm. These features are considered during designing the robot hand.
In the second, to overcome the situation that researchers are developing robot hands depending on their inspiration and experience because of the lack of method that can evaluate the performance of robot hands, a new performance index named `Interactivity of Fingers (IF)' is proposed and applied to four existing robot hands to show the usefulness of the IF. Moreover, two mathematical models of a robot hand (one with IMC joints and the other with none) are built and compared to each other to confirm the advantage of the IMC joints in the palm. Then, design optimization of the robot hand is carried out using the IF. The optimization is performed by using a Genetic algorithm (GA). In this process, it is possible to define the position and orientation of the CMC (or Saddle) joint of the thumb systematically.
Next, in Chapter 5, the design process of the robot hand including the aforementioned features in Chapter 2 is described. The robot hand is composed of the finger module, the thumb module, and the IMC joint module so that it is convenient to repair malfunctioning module by simply replacing the damaged module with a new one. Since the robot hand was designed to be manufactured using a 3D-printer, the size of the robot hand is designed to be 1.5 times bigger than that of the human hand because of the weak solidity of frame's material. The kinematic analysis including the forward and inverse kinematics of the robot hand is also addressed.
Finally, the weight, the payload, and the control system, and grasping strategy of the robot hand are provided in Chapter 6. To show the high grasping capability, experimental evaluation on the robot hand using the coordinate center of the workspace intersection volume is carried out. The robot hand's index finger, middle finger, and thumb are directed to the coordinate center of the workspace intersection volume, then the center of gravity of various objects, such as a rectangular sponge (25mm thickness), a roll of insulation tape (30mm radius), a coin of KRW 100 (12mm radius) and KRW 10 (9mm radius) are located at this position. This shows very promising results, that even the tiny object, a coin of KRW 10, is easily grasped.