VBAMS: Vision-Based Sensing for Estimating and Controlling Shrinkage Ratios in McKibben Artificial Muscles

This paper proposes a self-sensing method for the Mckibben Artificial Muscle with a single built-in camera. Mckibben artificial muscles have been the focus of recent robotic systems due to their flexibility and lightweight. Position control of artificial muscles involves sensing the length of the artificial muscle, but most of those sensing methods face challenges in adapting to various geometries and environmental changes. To solve this problem, this study developed a method to estimate the shrinkage ratio of pneumatic rubber artificial muscles by using internal image processing of the VBAMS system (Vision-Based Artificial Muscle Sensing) to obtain the inner diameter of the artificial muscle’s rubber tube virtual artificial muscle. A dome-shaped marker was placed on the flexible rubber tube, one of the components of the artificial muscle. Since the marker moves during the expansion and contraction phase of the muscle, we use a single camera to measure the displacement of the marker and use it to measure the virtual inner diameter. Experiments were conducted with equipment using this system to verify its applicability to feedback control of the actual system.