Motion Artifact Modeling of Capacitive Electrodes Based on Triboelectric Nanogenerators

Capacitive electrodes are a fundamental component to enable contactless biosignal sensing. However, they are prone to motion artifacts (MAs). In the last decade, MAs have been modeled based on statistics or unverified equations, often being described as unpredictable. In this work, MAs are described with a new physically accurate model based on triboelectric nanogenerator models. The model verification employed electrical measurements, confirming its accuracy while also providing a physical explanation for MAs. The new model represents a critical tool to better understand this phenomenon and develop solutions for these artifacts. It accounts for transverse and longitudinal motions. For the latter, both sliding over the skin or into an air gap are considered. Numerical simulations and measurements were performed both in a test jig and in vivo . In a controlled setup, an agreement between predicted and measured behaviors for different speeds and lengths of motion was observed. The in vivo measurements showed that by estimating and scaling the energy spent to start the motion, the MA amplitude was matched. The measured and predicted spike duration of a transverse motion were 63 ms and 40 ms, while the measured and predicted settling time of a longitudinal motion were 2.6 s and 3.0 s, respectively.