The Internet-of-Things (IoT) paradigm is under constant development and is being enabled by the latest research work from both industrial and academic communities. Among the many contributions in such diverse areas as sensor manufacturing, network protocols, and wireless communications, energy harvesting techniques stand out as a key enabling technology for the realization of batteryless IoT end-node systems. In this paper, we give an overview of the recent developments in circuit design for ultra-low power management units (PMUs), focusing mainly in the architectures and techniques required for energy harvesting from multiple heterogeneous sources. The paper starts by discussing a general structure for IoT end-nodes and the main characteristics of PMUs for energy harvesting. Then, an overview is given of different published works for multisource power harvesting, observing their main advantages and disadvantages and comparing their performance. Finally, some open areas of research in multisource harvesting are observed and relevant conclusions are given.
3 April 2019 in IEEE Transactions on Industrial Electronics
Modern Internet-of-Things (IoT) devices rely on energy harvesting to extend their battery life and create maintenance-free applications. However, energy from ambient sources is intermittent and unpredictable, which compromises the reliability of any system. Harvesting from multiple sources allows the collection of higher amounts of power, producing robust operation even in the presence of ambient variations. This paper describes the design of a fully integrated system that combines the energy delivered by four inputs. The proposed architecture applies a smart harvesting strategy based on the available power from the transducers. The two inputs with the most power are automatically selected and combined, while maximum power tracking is performed. If the other two sources still show some usable power, a two-way energy storage scheme allows their use to sustain the circuit's operation. The system was fabricated in a 0.13-μm CMOS technology. Experimental results demonstrate a peak tracking efficiency of 96%. The power conversion efficiency has a maximum of 87.2% at 198 μW of output power and 72% at 981 μW.