[SIST Student Seminar] Low-power Endpoint Intelligent Systems Based on Energy Harvesting and Tiny Machine Learning

Topic: Low-power Endpoint Intelligent Systems Based on Energy Harvesting and Tiny Machine Learning

Speaker: CHEN Zijie

Date and time: 14:00–15:00, October 13

Venue: SIST-1C 101

Host: LIANG Junrui

Abstract: 

While large models continue to captivate the world, the development of lightweight and low-power technologies has always held significant importance. With the advance of the Internet of Things (IoT) and artificial intelligence (AI) technologies, artificial intelligence of things (AIoT) systems are becoming increasingly prevalent for pervasive smart sensing. The topic of this presentation is “Low-power Endpoint Intelligent Systems Based on Energy Harvesting and Tiny Machine Learning,” encompassing three main areas of focus: battery-free IoT, self-powered sensing, and edge AI. The speaker will introduce his related works in detail, including the motion-powered human activity recognition (HAR) system, the energy-efficient on-device predictive maintenance (PdM) system, and the first self-powered PdM system utilizing the energy harvester simultaneously as both a power source and a self-powered sensor. These works were published in IEEE IOTJ, IEEE TIM, IEEE/ACM ISLPED and contribute to the promising future of ubiquitous AIoT and pervasive sensing.

Biography:

Mr. CHEN Zijie received his B.E. degree in electronic information engineering from Hangzhou Dianzi University in 2021. He is now studying for a Master’s degree at ShanghaiTech University. His research interests cover edge intelligence (TinyML), smart sensing, and battery-free IoT.