# Design and Evaluation of a Conductive-Knit Sensor System for Measuring Forearm Pronation and Supination

## TL;DR

A wearable textile sensor using carbon nanotubes was developed to measure forearm rotation accurately and comfortably without lab equipment.

## Contribution

A novel sensor configuration using two CNT-based knitted sensors at 45° angles minimizes hysteresis and enables reliable rotational angle estimation.

## Key findings

- The sensor system achieved ~2.5° RMS error in offline analysis and ~5° in real-time compared to a gyroscope.
- Nonlinear (cubic) regression models outperformed linear models in estimating rotational angles.
- The system offers breathability, flexibility, and comfort without direct skin attachment.

## Abstract

Wearable sensors enable continuous monitoring of human motion in daily life, but assessing forearm pronation and supination usually requires specialized laboratory equipment. This study aimed to design and evaluate a knitted textile sensor system based on carbon nanotube (CNT) conductive yarns for unobtrusive measurement of rotational forearm movements. To overcome the hysteresis inherent in knitted strain sensors, we implemented a mechanism using two sensors positioned at 45° on the forearm, arranged so that only one sensor undergoes elongation during each motion. This configuration leverages the sensor’s stable elongation response while minimizing the influence of unloading hysteresis, thereby enabling reliable estimation of rotational angles. Electrical characterization confirmed consistent strain resistance during elongation, and regression analysis showed that nonlinear (cubic) models provided superior accuracy, compared to linear models. Validation experiments demonstrated that the system achieved errors of about 2.5° root-mean-square in offline analysis and approximately 5° in real-time estimation, compared with a gyroscope. Importantly, the system did not require direct skin attachment and offered the advantages of breathability, flexibility, and comfort inherent to textile-based designs. These results highlight the feasibility of CNT-knitted sensors as a lightweight, cost-effective, and user-friendly platform for capturing complex upper limb movements, with promising applications in rehabilitation monitoring, sports performance assessment, and everyday healthcare technologies.