Design and Instrumentation of Environment-Powered Systems
OPEN ACCESS
Author / Producer
Date
2020
Publication Type
Doctoral Thesis
ETH Bibliography
yes
Citations
Altmetric
OPEN ACCESS
Data
Rights / License
Abstract
Energy Harvesting presents a key technology to sustainably supply the billions of devices in the emerging Internet of Things (IoT). Converting physical signals such as radiation, temperature, vibration, etc. into electrical energy promises virtually unlimited energy to supply cyber-physical systems (CPSs) in a long-term and scalable manner. However, with an energy supply depending on a spatially and temporally variable environment significant non-determinism is introduced into the system. In this thesis we explore the potential and limitations of supplying cyber-physical systems (CPSs) from environmental energy using only minimal energy buffering. We introduce novel design methodologies to supply applications reliably and efficiently, explore the energy yield of thermoelectric harvesting, and optimize the utility of data transmissions in infrastructure-less monitoring. Furthermore, we introduce a testbed and measurement support to assist designers in design aspects arising in energy harvesting systems. Specifically, we make the following contributions: - We introduce a novel measurement tool that combines high accuracy and portability. Enabling joint in-situ observations of the ambient, multiple energy flows, and application states, it provides critical insights during the design and verification of energy harvesting systems. - We present a testbed for the emulation of radiation and temperature environments. In combination with a programmable, time- and event-triggered current sink, it enables fast and repeatable exploration, dimensioning and validation of energy harvesting system design aspects. - We introduce the first model for thermoelectric energy harvesting at the ground-to-air boundary that incorporates all components from the physical signal to the application. In combination with a newly proposed rectifier circuit, an optimized harvesting system is implemented. Extensive real-world evaluation attests the accuracy of the model and demonstrates unprecedented output power in the given harvesting scenario. - We propose a novel energy management principle that decouples the energy harvesting and electrical load using a minimal energy buffer to allow each end to operate at is optimal operating point. An energy management unit (EMU) implementing this principle is designed and extensively evaluated. Efficient and reliable operation is demonstrated, even when the input power is significantly lower than the application requirements and exhibiting high variability. - We study the utility of data transmitted in an infrastructure-less communication scenario supplied by energy harvesting. Using a model-based optimization approach, we derive a new data transmission scheme for long-term batteryless monitoring applications. Evaluation using a batteryless sensor nodes demonstrates accurate abstraction of the scenario using our model and significant gain in utility at minimal run-time overhead. The methods and solutions presented are implemented and extensively evaluated under lab and real-world conditions. From these, we conclude that the methods and design tools presented enable efficient design and thorough evaluation of energy harvesting systems.
Permanent link
Publication status
published
External links
Editor
Contributors
Book title
Journal / series
Volume
Pages / Article No.
Publisher
ETH Zurich
Event
Edition / version
Methods
Software
Geographic location
Date collected
Date created
Subject
Energy harvesting; Transient computing
Organisational unit
03429 - Thiele, Lothar (emeritus) / Thiele, Lothar (emeritus)
Notes
Funding
157048 - Transient Computing Systems (SNF)