Wearable electronic devices collect user data to provide personalized experiences and real time interactions for the metaverse, thereby driving its development. This paper introduces an innovative self-powered sensing smart monitoring system (SSSMS) designed to harvest human low-frequency inertial energy and achieve gait recognition and fall monitoring (GR-FM) capabilities. The SSSMS enhances energy harvesting efficiency through Halbach electromagnetic enhancement mechanisms, achieving self-sustainability. Additionally, the conical roller triboelectric nanosensor (CR-TENS) accurately captures various human gaits, transmitting these states in the form of electrical signals to a computer via OpenBCI in real time. These signals are integrated with a deep learning-driven diagnostic module to achieve GR-FM. The experimental results demonstrate that under optimal external excitation, the output power is 2.27 mW, with a power density of 32 W/m3. After 100,000 cycles, both the EMG and CR-TENS modules still maintain good performance, sufficiently providing stable and continuous charging for the battery and offering a stable signal for signal processing. Both the EMG and CR-TENS modules exhibit robust feature capture capabilities, effectively sensing different excitation frequencies, amplitudes, and gait patterns. The SSSMS based on the GRU deep learning model achieves recognition accuracies of 96.13 %, 96.60 %, and 92.22 % for frequency, amplitude, and gait, respectively. Deployment experiments demonstrate high accuracy and sensitivity in frequency, amplitude, gait recognition, and fall monitoring, highlighting the advantageous design of EMG and CR-TENS for perceiving subtle motion state changes. The SSSMS demonstrates outstanding capabilities in self-sustainability and real time user embodiment information capture, showcasing its potential as a next-generation interaction device for the metaverse.
Read full abstract