A Battery-free Human Activity Recognition System Based on Kinetic Energy Harvesting

Abstract

Nowadays, there are more and more wearable monitoring devices for multiple applications. Most of these existing devices are powered by chemical batteries. As the Internet of Things (IoT) nodes increase, replacing and disposing of batteries will become prohibitively labor-intensive and environmental-unfriendly. Energy harvesting (EH) is a promising technology that scavenges wasted ambient energy, which is from solar radiation, vibration, and wind. In this paper, we introduce a real-time, low-cost human activity recognition (HAR) system based on kinetic energy harvesting (KEH), where the harvester works as both energy source and sensor. The adopted well-rounded energy-aware circuit allows the whole system to be low-power as well as low-cost. Utilizing the unique relationship between human arm swing and harvested energy, the information can flow along with energy inside the system. By knowing the interval between transmitted packets, the proposed system can realize HAR in real-time. In addition, an all-in-one prototype has been fabricated for validating the performance of the proposed system. Lab tests and field tests demonstrate that the proposed system can reliably recognize different human activities, such as standing, walking, jogging, and running. As a cyber-electro-mechanical co-design, this study brings a promising solution for pervasive self-powered HAR and battery-free IoT systems.