Electro-Pneumatic Interface Framework for PAM-based Humanoid Robot Motion Control with EMG

In this work, a real-time electro-pneumatic interface framework for emulating human muscular motion on a pneumatic-actuated muscle-based robot is presented. Pneumatic-actuated muscles are flexible actuators that contract and expand using pressure input and are widely used in medical and robotic research to generate motions that mimic human muscles. Electromyographic signals that measure the nerve stimulation of the muscles in an organism, can be used to emulate real-time muscular motion in Pneumatic Artificial Muscle based robots. Here, the proposed electro-pneumatic interface framework takes a real-time electromyographic signal as input and outputs corresponding pneumatic signals for actuating the pneumatic artificial muscles. The electromyographic sensors, attached to the human body, pick up the signals during motion, and the software electro-pneumatic interface computes the required differential pressure for the robot’s motion. In this approach, experimental data is used to model the transfer functions between variables such as the electromyographic signal, robot deflection angle, and differential pressure for the pneumatic artificial muscles. The proposed framework is experimentally validated on a humanoid robot made of pneumatic artificial muscles. The simulation and experiment results show the virtues of the proposed methodology.