Hysteresis Modeling and Compensation in Antagonistic Tendon-Sheath Mechanisms via Tension Detection Under Various Conditions

This paper proposes a model for estimating hysteresis and compensating for it in an antagonistic tendon-sheath mechanism (TSM) based on the strong correlation between tension and the magnitude of hysteresis. The TSM is an essential mechanical component for delivering accurate displacement in flexible tubes with arbitrary paths, but the hysteresis changes with the shape of the path is a challenge. This study focused on modeling hysteresis using only tension data without relying on path data. Experiments showed that the strong correlation between tension and hysteresis remained consistent despite the variations in velocity, path, and frictional forces acting on the end-effector. A backlash model-based algorithm was proposed to estimate the hysteresis by relying solely on the tension value, which is easier to obtain than path data or other information, using the derived tension-hysteresis model. The variation in tension with the velocity of tendon displacement was modeled using the initialization process. A strategy was proposed to predict and feedforward compensate for the hysteresis through the proposed estimation algorithm. For validation, 36 experimental trials were conducted, varying conditions along arbitrary paths, showing that the estimation model achieved a mean absolute error of 0.19°, and the compensation model achieved a mean hysteresis reduction of 87.7%. These results showed the feasibility of relying solely on tension data for hysteresis estimation and compensation.