Battery-free Internet of Thingsmahbub/PDF_Publications/IOT_DSP_KL_2016.pdf · 1. [IC2015]...
Transcript of Battery-free Internet of Thingsmahbub/PDF_Publications/IOT_DSP_KL_2016.pdf · 1. [IC2015]...
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Never Stand Still Faculty of Engineering Computer Science and Engineering
Click to edit Present’s Name
Never Stand Still Faculty of Engineering Computer Science and Engineering
Battery-free Internet of Things
Making the most of energy harvesting
Mahbub Hassan School of Computer Science and Engineering
University of New South Wales, Sydney, Australia
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Welcome to Internet of Things
Welcome to a Smart World
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Smart Home
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Smart Industry
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Smart Farm
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Smart Health
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Welcome to batteries?
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Outline
• Motivation for energy harvesting • Part 1 - Energy harvesting IoTs • Part 2 - Context detection from energy harvesting • Conclusion and future directions
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Typical Energy Harvesting IoT
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Energy harvesting sources Temperature sensor using solar EH
Wireless EEG using thermoelectric EH
Vibr
atio
n R
F (T
V si
gnal
)
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
More EH IoT products (from enocean – industry standard)
Occupancy Sensor (solar powered) Door/Window Sensor (solar powered)
Self-powered wireless switch (pressure powered)
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
High power consumption for Wearable IoTs • Continuous activity and context monitoring --- killer app • Continuous motion sensing à high power consumption • EH wearable IoTs more challenging
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Powering wearable IoT from energy harvesting convert motion energy to electrical energy
Piezoelectric Energy Harvester
Sensors
MCU
Radio
Harvested Power
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Scarcity of harvested energy
• Humans generate a small amount of kinetic energy, but • Accelerometer sampling is power consuming
Powering perpetual activity detection only using the harvested kinetic energy is a challenging problem
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Outline
• Motivation for energy harvesting • Part 1 - Energy harvesting IoTs
• Part 2 - Context detection from energy harvesting • Conclusion and future directions
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Our Idea – KEH as a sensor for context detection
KEH
walking
running
Harvested Power
• Harvested energy is influenced by human activity, so activity should be detectable
• Power saving potentials: unlike accelerometer, KEH does not consume power
• KEH proxying as a motion sensor may have other benefits (simpler, smaller form factor for wearables)
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Piezoelectric Energy Harvester (vibration-based)
Mide.com
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
KEH Data Logger
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
KEH Data Logger
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Exp #1 [IC2015] Human activity recognition – KEH Data Collection
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
KEH Time Series
Walking
Running
Standing
accelerometer KEH
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Does KEH contain information for activity recognition In
form
atio
n G
ain
Features
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Activity Recognition Accuracy – 5 activities (W, R, S, SU, SD)
Classifier Activity Recognition Accuracy (%)
Accelerometer-based
KEH-based
Hand Waist Hand Waist
K-nearest neighbour 95.01 98.70 81.13 87.01
Decision Trees 87.91 91.02 79.74 79.86
Multilayer Perceptron
88.25 96.39 78.29 85.52 Low
er th
an a
ccel
erom
eter
, bu
t not
too
bad
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Exp #2 [iThings2015] Step counting
4 su
bjec
ts, s
trai
ght a
s w
ell
as tu
rnin
g pa
ths,
pea
k de
tect
ion
algo
rithm
, 570
st
eps,
96%
acc
urac
y
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Exp #3 [BODYNET2015]
Calorie burning estimation • 10 subjects, 4 male, 6 female: waist mounted • Anthropometric data: Age (26-35 years, mean = 29, s.d.
= 3.06), Weight (58-91 Kg, mean = 69.3 s.d. = 10.21), Height (154-185 cm, mean = 168.5, s.d. = 9.98)
• Two activities: walking and running • Linear regression model to estimate calorie burning from
energy harvesting samples and anthropometric data • Leave-one-out cross validation (1 for testing, 9 for
training)
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Calorie estimation results - running
Close to accelerometer
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Exp #4 [PerCom-WIP2016]
Transport mode detection
Train Car Bus Walking
Running
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Pause periods of train (pauses are removed)
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
2-layer classification
Peak analysis
Mean analysis
Summary of Trace Data
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Results
93% 77% 88%
Bus and Car are confused, but not train
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Exp #5 [WOWMOM2016]
Hotword detection
3cm
Quiet Room Conditions
Hotword: “OK Google” Non-hotwards: “Good morning”, “how are you”, “fine, thank you” 8 subjects: 4 m, 4 f 60 instances (30 hotword 30 non-hotword) per subject
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Speaker orientation
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Results
Flat Vertical
Speaker Independent
78% 62%
Speaker Dependent
85% 73%
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Conclusion
• Vibration energy harvesting can detect a wide range of human contexts à power saving opportunity
• Further research is required to improve context detection accuracy and reduce system power consumption
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
References 1. [IC2015] Energy-Harvesting Wearables for Activity-Aware Services," IEEE
Internet Computing, 9(5), 2015 2. [iThings2015] Step detection from power generation pattern in energy-
harvesting wearable devices,” IEEE iThings 2015
3. [BODYNET2015] Estimating Calorie Expenditure from Output Voltage of Piezoelectric Energy Harvester - an Experimental Feasibility Study" BODYNETS 2015
4. [PerCom_WIP2016] Transportation Mode Detection using Kinetic Energy Harvesting Wearables, PerCom Work-in-Progress, 2016
5. [WOWMOM2016] Feasibility and Accuracy of Hotword Detection using Vibration Energy Harvester, WOWMOM 2016 (accepted).
Mahbub Hassan IEEE Distinguished Lecture, Malaysia, 26 May 2016
Questions?