Design and Control of a 2-DOF Forearm Robotic System Driven by Pneumatic Artificial Muscle Actuators

• Main Authors: Wei-Chan Hsu, 徐維謙
• Other Authors: Mao-Hsiung Chiang
• Format: Others
• Language: en_US
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/36748708216125412697

碩士 === 國立臺灣大學 === 工程科學及海洋工程學研究所 === 102 === During the second half of the 20th century, assistive robotics and exoskeletons appeared as a new type of robots that can either provide assistive support for patients with impaired limbs or augment the strength of human beings. Although electromagnetic motors may be the most widely used actuators in the robotic area, pneumatic actuators have their own advantages, such as lightness, cleanness and easy maintenance. Of all the pneumatic actuators, pneumatic artificial muscles (PAMs) may be the most promising one for the design of assistive or rehabilitation robots because of their inherent compliance and high power to weight ratio, despite their high nonlinearity. This thesis aims to design and control a 2-DOF forearm robotic system actuated by two pairs of PAMs, following the research of [1], as a step toward our future objective of the development of an upper-limb rehabilitation robot. For the purpose of the study, basic biomechanics of the human body is briefly discussed, and then the design as well as the test rig of the robotic system is demonstrated. To understand the properties of the system, the kinematic and dynamic models of the mechanism are derived, followed by the mathematical modeling of the pneumatic components used in the system. These modeling approaches lend themselves to the controller design, in that the pressures of the PAMs that are required to produce a desired force under a certain contraction ratio can be easily calculated, such that a model-based controller can be designed. For the controller design, a torque controller with a model-based feedforward controller and a feedback PID controller is first developed, and then a sliding mode controller is combined with the torque controller to form a joint angle controller. With the help of a velocity observer, a decent velocity signal can be employed for the control of the system. In addition, the estimated disturbance by the disturbance observer is used as a cancellation signal for better angle tracking performance. Simulations are carried out on the elbow joint to test the model and the controller, and experiments are conducted to verify the efficacy of the design and control of the system. The results show that the smooth angle tracking performance can be achieved.