A triboelectric sensor with an inverted pendulum design for machine and human movement monitoring in freestanding mode

Abstract The vast availability of ambient mechanical vibrations in the natural environments and our daily activities has spurred the advancement of triboelectric sensors for vibration sensing. However, the operation of the traditional triboelectric vibration sensors is usually constrained to contact-separation mode, limiting their functionality to transverse vibrations and making them unsuitable for longitudinal and rotary movements. The inherent wear and tear associated with the contact-separation mode further restricts their practical application. To address these limitations, this study presents a new design for a vibration sensor that employs a spring-assisted inverted pendulum structure. The vibration sensor exhibits a 10 Hz resonant frequency and produces maximum triboelectric output at the vibration amplitude of 5 mm. As a proof of concept, the vibration sensor successfully detects balanced and unbalanced fan blades by analyzing the obtained data via Fast Fourier Transform (FFT) analysis. Besides machine monitoring, the proposed vibration sensor can be integrated into a wearable device for tracking human wrist movements. Combined with a machine learning algorithm, the sensor has achieved an accuracy level of greater than 95% in recognizing four distinct wrist motions: bending, rotating, waving, and handshaking. These findings have proven that the proposed triboelectric sensor design can be effectively integrated into wearable technologies, smart factories, and virtual/augmented reality systems, significantly broadening the triboelectric sensor applications.