Application of artificial muscle e-rubber for healthcare sensing: verification of measurement properties as a smart insole.

Electroactive polymer (EAP) artificial muscles are gaining attention in robotic control technologies. Among them, the development of self-sensing actuators that integrate sensing mechanisms within artificial muscles is highly anticipated. This study aimed to evaluate the accuracy and precision of the sensing capabilities of the e-Rubber (eR), an artificial muscle developed by Toyoda Gosei Co., Ltd., and to investigate its potential for healthcare sensing applications such as smart insoles. The objective was to transform the eR into a thin capacitor and estimate the applied load by sensing minute changes in the capacitance. The changes in the EAP dielectric constant, electrode area, and inter-electrode distance, all of which define the capacitance, are non-linear functions. The relationship with the external force also exhibits nonlinearity. To address this issue, we experimentally plotted the load and capacitance changes and derived a regression equation. We evaluated the sensing characteristics of both a stand-alone sensor and a sensor embedded in a smart insole, followed by a precision verification of the load estimation using the derived regression equation. Load-capacitance changes were measured up to 400 N at three conditions: 23 °C and 50% humidity, 40 °C and 50% humidity, and 40 °C and 80% humidity. For the standalone sensor, the coefficient of variation was less than 1.25% and the confidence interval was 0.25%, indicating high precision. However, for the sensor embedded within the insole housing, the coefficient of variation increased to less than 8%, and the confidence interval was 1.5%, likely owing to the influence of gaps within the insole structure. Regarding the load estimation equation, a 5th-order polynomial approximation (R >0.999) demonstrated the best fit, indicating that it is sufficiently accurate for healthcare sensing applications. Although capacitance-based sensors are increasingly being used in biomedical monitoring for pressure and load measurements owing to their durability and high sensitivity, their primary challenge lies in the nonlinearity of the sensing results. Although this challenge also exists for capacitance sensors utilizing artificial muscles, our study shows that developing a regression equation based on the experimental relationship between the load and capacitance changes can yield sufficient precision for practical healthcare applications.