Abstract
Human activity monitoring plays a critical role in clinical diagnostics, fitness assessment, remote patient care, and related domains. In this study, we design and develop a biodegradable, environmentally friendly triboelectric nanogenerator (TENG) sensor using corn silk and integrate it with machine learning techniques for intelligent human motion detection. The work emphasizes the use of sustainable and biodegradable materials to advance green electronic systems. Corn silk, an abundant agricultural byproduct, serves as the primary triboelectric material in the proposed sensor. The corn silk-based TENG is capable of charging commercial capacitors and generates an output voltage of 101 V. In addition, the autonomous eco-friendly sensor is evaluated as a wearable device for capturing and distinguishing user activities such as jumping, running, and walking. To enhance activity recognition accuracy, an ensemble machine learning model based on a Histogram gradient boosting classifier is integrated with the TENG output signals. The model achieves a classification accuracy of 98.7% across the three targeted activities. Furthermore, a graphical user interface is developed to seamlessly combine the optimized machine learning model with real-time TENG sensor inputs for efficient human activity recognition. Overall, this study presents a customized TENG-based sensing methodology and highlights its potential for developing intelligent, self-powered, and human-motion monitoring systems.