A Reconfigurable Energy Storage Architecture for Energy-harvesting Devices

The Computing Landscape of the 21st Century

“…Emerging wireless backscatter networking enables communication at extremely low power [15]. Intermittent computing allows sensing and complex processing on scarce energy [5,17]. Such capabilities enable a new breed of sensors and actuators deployed in the human body to monitor health signals, in civil infrastructure, and in adversarial environments like outer space [5].…”

“…Intermittent computing allows sensing and complex processing on scarce energy [5,17]. Such capabilities enable a new breed of sensors and actuators deployed in the human body to monitor health signals, in civil infrastructure, and in adversarial environments like outer space [5]. RF beamforming extends the capability of batteryless, networked, in-vivo devices [19].…”

Reconsidering batteries in energy harvesting sensing

“…Many intermittent systems papers, however, dismiss them as expensive [17][18][19][20], short-lived [9, 17-20, 26, 47]. temperature-sensitive [9,[17][18][19][20]26], less efficient [17][18][19][20], bulky [17][18][19][20]47], and dangerous [17][18][19][20]. In this section, we re-examine each of these arguments, in the context of newly available modern battery chemistries, including Lithium Titanate (LTO) and Lithium Iron Phosphate (LiFePo 4 ), and low power battery management ICs that simplify energy harvesting designs [44], concluding that these new chemistries do not exhibit the same detracting qualities as those available a decade ago.…”

“…Even with an order of magnitude capacity reduction to achieve high cycle lifetimes, batteries have a usable energy density of 5-50x greater than supercapacitors of similar physical size. The batteries shown [18], the Solar Monjolo [6] and the Capybara temperature monitor and alarm [9], which have total capacitances and energy capacities of 119 µF (0.41 µWh at 5 V), 500 µF (1.7 µWh at 5 V) and 8.8 mF (8.3 µWh at 2.6 V), respectively. Capacitors (d) [7] and (f) [30] are large supercapacitors available on the Capybara platform and have the capacitances and energy capacities of 300 mF (300 µWh at 2.7 V) and 220 mF (540 µWh at 4.2 V) respectively.…”

“…Federating energy stores for each predefined component workloads decreases latency between tasks and eases modularity [17]. Reconfigurable energy buffer sizing lowers inter-task latency for arbitrary workloads [9]. Even with these solutions, however, it is still not easy to program and use capacitor-based energy harvesting sensors, and they still have decreased availability and energy utilization compared to a solution with larger energy storage capacity.…”