Holistic Energy Management with uProcessor Co-Optimization in Fully Integrated Battery-Less IoTs


ully integrating on-chip power management modules and microprocessor into a system-on-a-chip (SoC) provides many benefits including lower cost and higher levels of optimization, especially for battery-less energy harvesting operation. However, previous studies have only focused on optimization of individual module, e.g. power converter. There is a lack of systematic optimization of energy efficiency considering microprocessor, regulator, and harvester. This paper performs a holistic study on power efficiency of the whole energy harvesting system including solar cells, on-chip voltage regulators and microprocessors. An optimal scheduling and operation strategy is proposed for achieving the best efficiency and system level performance while avoiding local minimum as in conventional approach. We show that the minimum energy point is different from conventional approaches without a holistic view of the system. We demonstrated the study and proposed scheme using a battery-less solar energy harvesting system and a 65nm fully integrated test chip with 20% additional energy savings.

2018 31st IEEE International System-on-Chip Conference (SOCC)