A dual-modal cooperative energy conversion biomechanical energy harvester for wearable devices and AI-assisted motion monitoring

Abstract The rapid advancement of Artificial Intelligence has continuously enhanced the self-sensing and autonomous decision-making capabilities of wearable devices, thereby further accelerating the fulfillment of growing demands for health monitoring. To address the challenges related to power supply and motion monitoring in conventional wearable technologies, a dual-modal cooperative conversion biomechanical energy harvester is proposed. It is designed to ensure wearing comfort while enabling efficient energy harvesting and accurate motion monitoring. It integrates two synergistic energy conversion mechanisms: one enables energy harvesting during the negative work phase of the knee joint, and the other achieves efficient electrical conversion through the release of stored energy from a coil spring. This dual-modal cooperative design significantly extends energy harvesting duration and improves power generation efficiency. Experimental results show that at 4 km/h treadmill speed, the harvester achieves an average output power of 263.05 mW and a peak power density of 0.26 W/kg. Moreover, the electromagnetic signals generated by the harvester can accurately identify different motions when processed using the Gated Recurrent Unit model. Consequently, this study offers a novel strategy for the development of wearable systems and is expected to advance smart health management toward safer, greener, and more sustainable solutions.