Innovation in Pneumatic Artificial Muscle Control: Breakthrough in Adaptive Fuzzy Sliding Mode Contr

2023-12-19

Pneumatic artificial muscles (PAM) have been widely used in fields such as robotics, rehabilitation, and prosthetics as potential actuators for simulating human motion in recent years. However, due to its nonlinear characteristics, the motion trajectory control of PAM systems has always faced challenges.

Recently, a group of researchers proposed an innovative method of adaptive fuzzy sliding mode controller, which estimates the control parameters of PAM systems through fuzzy logic, significantly improving their motion accuracy and adaptability.

PAM is usually made of rubber and covered with woven yarn on the surface, which can mimic the characteristics of human muscles. When inflated, PAM becomes hard and contracts; When deflated, it will become soft and stretch. However, PAM is a nonlinear system with delays, so an effective control system is needed to adjust its performance.

Traditional control methods have certain limitations in dealing with the nonlinearity and hysteresis phenomena of PAM systems, thus requiring new solutions. The research team is led by Ngoc Tam BUI, Associate Professor at the School of Engineering at Shibuya Institute of Technology in Japan, and Dr. Quy Thinh Dao from Hanoi University. They proposed the AFSMC method, which uses Takagi Sugeno fuzzy algorithm to estimate disturbances and automatically update output values, thereby achieving better motion tracking and adaptability.

In the research, the team first designed a sliding mode controller with control signals, which introduced special variables to estimate disturbances and improve control performance. Then, through adaptive rules, an adaptive fuzzy algorithm was created to automatically update the component rule parameter vector to calculate disturbance variables.

By analyzing the Lyapunov stability conditions of the developed AFSMC algorithm, researchers verified its stability. In addition, a series of experiments have shown that the AFSMC method performs better in motion tracking accuracy compared to traditional sliding mode control methods. When the load frequency is 0.5 Hz, the root mean square error of the AFSMC method is only 2.68 °, while the traditional sliding mode controller is 4.21 °. In addition, the AFSMC method also demonstrates excellent adaptability and can effectively cope with sudden external interference.

Researchers believe that this innovative method has the potential to be applied to robot rehabilitation equipment, assistive devices, and personalized treatment, bringing precise treatment effects to patients with rehabilitation needs. In addition, this method also contributes to the design and development of advanced prosthetics, enhancing their functionality and rehabilitation effects.

Although this study lays the foundation for the motion trajectory control of PAM systems, the research team hopes that this achievement can inspire further exploration and development in the field of rehabilitation technology. According to Associate Professor Ngoc Tam BUI, according to research findings, there may be a commercial rehabilitation system based on PAM in the next 5 to 10 years, which will bring significant benefits to patients with spinal cord injury, stroke, and other rehabilitation needs. The long-term impact of this study is expected to drive continuous innovation and progress in rehabilitation technology.

MIT develops a new framework to enable robots to quickly learn in new environments