1. SNOOZE ALLEN presents
“NOVI LIFE”
HOW TO PREVENT FALLING ASLEEP AT THE WHEEL
By: Nanfei Liu, Nicholas Goryachev, & Sean Brown
2. Issue of drowsy driving
1,550 Deaths!
100,000
Accidents!
$12.5 Billion in
Monetary Losses!
No test to determine sleepiness
3. GOAL: To decrease the number of
road accidents caused by falling
asleep behind the wheel.
OBJECTIVE: To create a wearable
device that detects fatigue by
alerting the wearer before he/she
dozes off.
4. Target Audience
○ Young people
■ Especially men (aged 18-29 years old)
○ Adults with children
○ Shift workers
○ Long distance commuters and professional
drivers; truckers
5. Elon Musk’s comments on Autopilot (Tesla S Model)
"AUTOPILOT IS GETTING BETTER ALL THE
TIME, BUT IT IS NOT PERFECT AND STILL
REQUIRES THE DRIVER TO REMAIN ALERT."
7. Components of NOVI Band
Thermoregulation
Electrodermal
Response
Heart Rate
Accelerometer &
Pedometer
➔ Button to connect Bluetooth
➔ Waterproof & Shockproof
➔ Lightweight & durable
➔ Electronic motor for vibrations
➔ Speaker for digital beeps
➔ Charging port
NOVI
8. Electrodermal Activity (EDA)
The skin becomes a better conductor of electricity when there is
internal/external stimuli
Body Heat
Prior to drowsiness our bodies begin to lose some heat (helps to induce sleep)
reduced by 1 to 2°F.
Motion
Accelerometer utilizes electromagnetic forces
1- Detect a person’s activity levels
2- Sensitive to motion
Heart Rate
Slows down as drowsiness increases
12. Why should invest?
Familiar and easy market entry
$34 billion worth
Lowered accident costs
as incentive for
government client
Insurance companies
as distributors
15. Appendix
● Target Audience
● Symptoms
● Sensor for Conductance
● Sensor for Motion
● Sensor for Temperature
● Sensor for Heart rate
● Additional App features
● Machine learning further explained
16. Symptoms of drowsiness
● Difficulty focusing, frequent blinking, or heavy eyelids
● Daydreaming; wandering/disconnected thoughts
● Trouble remembering the last few miles driven; missing exits or traffic
signs
● Yawning repeatedly or rubbing your eyes
● Trouble keeping your head up
● Drifting from your lane, tailgating, or hitting a shoulder rumble strip
● Feeling restless and irritable
17. Sensor for Conductance
● The skin conductance response/ electrodermal response (and in older
terminology as "galvanic skin response"), is the phenomenon that the
skin momentarily becomes a better conductor of electricity when either
external or internal stimuli occur that are physiologically arousing.
● As it pertains to drowsiness, skin becomes less conductive as drowsiness
increases/arousal decreases.
● The sensor would be able to measure electrodermal changes on the skin
over periods of time and reduced activity could indicate drowsiness
18. HOW do you measure Skin Conductance?
● 8mm diameter silver/silver chloride electrodes
● As perspiration increases, more sweat glands begin to conduct electricity in
a given area of skin.
BUT
● Fatigue leads to lowered body temp and less sweating
● Activity can affect conductance
Lowered activity->decreased temp->sent alert to app/device
19. Sensor for Motion
● An accelerometer utilizes electromagnetic forces such as the amount of
static acceleration due to gravity in order to measure the speed and
direction of an object’s travel.
● We will utilize two different accelerometers.
○ One accelerometer is already in the smartphone, which can detect if someone is driving in
a car
○ The second accelerometer is in the band itself, and it senses sensitive motion such as the
motion (or lack of) of one’s hand in order to gauge if someone is using a steering wheel,
running, sleeping, etc.
20. Sensor for Temperature
● Body temperature tends to drop when people become more drowsy (less
active)
● Just before we fall asleep, our bodies begin to lose some heat to the
environment (helps to induce sleep), reduced by 1 to 2°F.
● A decrease in body temperature may signal that there is a reduction in
metabolic activity, and thus an increase in drowsiness.
● Thermometer sensors relatively cheap, ~$5
21. Sensor for Heart Rate
● Heart Rate typically slows down as drowsiness increases, as shown by the
example graph.
● Heart rate sensors, like the previous sensors, are cheap to produce stand
alone.
● ~ $5
22. Function of APP
Analyze data
collected from
device
Create awake
and sleep
trend
Produce Sleep
report, and
thresholds
Able to alert drive.
More notifications
during circadian hours
GPS: pick up if
car is swerving,
speeding, driving
long disances
Able to connect with
friends (choose
people to alert)
User able to
interact and
personalize
Give suggestion of
coffee shops, sleep
clinics, relaxation
techniques
23. Machine Learning
The band itself doesn’t store or compile data, the app does. The sensors on the band send information to the phone which utilizes its
hardware and in-app software in order to store and process the biometric data. The band itself doesn’t need much memory or
power(battery) because all it has to do is take small-data interval snapshots of the biometrics and then send it to the phone (via
bluetooth); it can forget this information later or overwrite it with new data on the next snapshot it takes. When data is received by the
phone, the app compiles it and creates historical data out of it. For example, it will create running averages of heart rate during certain
parts of the day. It will do the same with every type of data from the sensors. The idea is that this app will be able to learn each
individual’s own circadian patterns (of sleep and day processes) and also how one’s biometric data is different every hour of the day.
One concern to this is that moving averages may also account for data coming together across various activity levels, and compile
both someone’s average heart rate at 8:32 pm on a friday once while they were running and another friday at 8:32 pm while they were
sitting behind the wheel. Obviously heart rate is higher when someone’s running, and compiling both 8:32 pm friday averages
generates unreliable data because of different circumstances. One deep learning process example is the app’s ability to also discount
misleading information, such as running, when formulating averages of people at rest. The app will see abnormally high heart rate, but
at the same time it will notice the accelerometer(or pedometer)’s readings to tell that someone is running, and then place that data
separate from behind the wheel data. Similarly a hyper-sensitive accelerometer in the band could notice and learn typical hand-on
steering wheel motion patterns . Ultimately, if the sensors detect behavior or data that is unusual for a certain hour of the day (usually
at night), and when the accelerometer in the phone says that the user is going fast enough to be in a car (and that maybe it is
potentially swerving), the phone makes an audible noise through the car’s bluetooth or aux connected speakers.
