The emergence of billions of Internet-of-Things (IoT) devices is significantly improving the quality of life and productivity. However, these devices must be fully autonomous and miniaturized to achieve ubiquitous deployment across various applications. Integrated circuits (ICs), as the foundation of modern electronic devices, face a wide range of design challenges to fully enable the IoT era, such as achieving high performance with limited energy due to a small form factor. Overcoming these challenges requires extreme energy efficiency and edge intelligence. Especially in the post-Moore’s Law era, IC design must optimize for both energy processing and edge computing by integrating these components together with cooptimization techniques into systems-on-chip (SoCs). This talk will introduce my work aiming to make the next-generation IoT devices more energy efficient and intelligent by leveraging advanced integrated circuit techniques from three main perspectives: energy processing, edge computing, and integrated systems. From an energy perspective, I will introduce two energy-harvesting designs including a piezoelectric energyharvesting system and a fully autonomous multi-modal energy-harvesting platform. These designs maximize the energy-extraction capability for self-powered systems with ultra-low quiescent power and high efficiency. To power emerging nanowatt devices, my work provides a complete power management solution with the lowest power records for two types of DC-DC converters and a voltage reference. From an edge computing perspective, I will discuss our recent machinelearning accelerator that uses in-memory computing (IMC) to explore the signal-to-noise ratio of different IMC banks in 28nm technology. From a system perspective, I will present an ultra-lowpower IoT SoC that integrates both energy delivery and edge computing features using a proposed triple-mode power management technique to maximize system-level energy efficiency. Finally, I will conclude my talk with future research directions, potential funding sources, and my teaching interests.
Shuo Li is currently a Postdoctoral Research Associate at the University of Illinois at Urbana Champaign. He received the Ph.D. degree in electrical engineering from the University of Virginia in 2021, the M.S. degree from Fudan University in 2016, and the B.Eng. degree from the University of Electronic Science and Technology of China in 2013. His research interests include analog/digital/mixed-signal integrated circuits and systems, energy harvesting and power management units, in-memory computing accelerators for machine-learning applications, and low-power systems-on-chip for emerging IoT devices. Dr. Li was a recipient of the IEEE International Symposium on Circuits and Systems (ISCAS) Best Paper Award in 2020 and a winner of the IEEE SSCS 2019-2020 International Student Circuit Contest. His research has been published in top peer-reviewed circuit design conferences and journals including ISSCC, JSSC, VLSI, and CICC. He also serves as a reviewer for the JSSC, TCAS-I, TVLSI, and TBioCAS.