Transcript Lesson14: Critterbot for Science 8 - cmaste

Critterbot – Lesson
1 (Sensors)
Science 8 Outcomes
This lesson will address the following outcomes from:
Science 8 Program-Of-Study:
Unit D: Mechanical Systems (Science and Technology Emphasis)
Outcomes for Science, Technology and Society (STS) and Knowledge
2. Analyze machines by describing the structures and functions of the overall system, the
subsystems and the component parts
• Analyze a mechanical device, by:
• Describing the overall function of the device
• Describing the contribution of individual components or subsystems to the
overall function of the device
Introduction: Critterbot
Do you have a pet?
Have you ever tried to train your
pet to sit? To fetch? To lie down?
Each time it learns a new trick or does
something good you would probably
give them a treat to indicate they did a
‘good job’!
Critterbot
What if your pet wasn’t a typical toilet
bowl drinking, peeing/pooping all over
your house pet…
…instead it is a robotic ‘critterbot’!
Critterbot
Just like a pet, you’ll
need to train your
critterbot…
BUT unlike a robot,
critterbot will learn new
tricks on its own.
Critterbot
Critterbot is a learning robot that uses robotic sensors
to gain information about its surroundings. It uses this
information to create higher-level knowledge (such as
what to do when it comes up to a wall) which allows it
to be a life-long learner.
Critterbot
Imagine a baby crawling up to a wall, the baby has a
couple of options:
(a) turn to the right or left
(b) stay at the wall
(c) back away from the wall
Eventually the baby will learn to do something after
they get to the wall, but how do they learn what to do?
Critterbot
Critterbot attempts to learn these seemingly natural
actions through reinforcement learning. When
critterbot is in front of the wall it may learn to back
away from the wall in which case the researcher will
give it positive reinforcement through computer code,
like the 'yummy treats' you give your pet for their good
behaviour.Critterbot will
eventually learn that when it
backs away from the wall it
is doing something good and
continue to back away from
walls each rolls up to a wall.
Critterbot – Reinforcement
Learning
For more information regarding reinforcement learning,
please refer to the AICML lesson ‘Reinforcement
Learning’ located either on your computer or at
Reinforcement Learning for Critterbot Science 8.
Critterbot – Machine Learning
1232746809.240 0 0 0 0 0 0 0 0 0 -32 -32 976 16 2 2 2 2 2 18 2 63
2 10 308
120 120
1232746809.250
0 0of0the
0 0 0 0 0 0 -32 -32
The172
Critterbot
Project
is an initiative
992 4 Reinforcement
2 3 3 3 5 17 4 64
3 19 308
168
124 120
1232746809.260 0 0
Learning
and
Artificial
Intelligence
0 0 0 0(RLAI)
0 0 0 -32
-32the
992
8 3 2 4 5of3 Alberta.
17 2 65 In
2 19
lab at
University
that308
lab172
they124 120
1232746809.270
0 0 0critterbot
0 0 0 0 0to0 be
-32an
-32
992 pet!
4 2 2This
2 2 2 17 2 66 5
build and teach
‘ideal’
21 312learning
172 120aspect
120 1232746809.280
0 0 0 0 0to
0 Machine
0 0 0 -32 -32 992
of critterbot is attributed
12 3 2Learning,
2 2 2 22 3where
65 2 17
304 172codes
124 120
computer
are 1232746809.290
programmed to 0 0 0
0 0 0 0review
0 0 -32
-32
976collected
0 2 2 2 2and
2 21
3 66what
3 18to
312
the
data
learn
do 168
with120
the 124
1232746809.300
0 0own.
0 0 0These
0 0 0 are
0 -32
-32‘thinking
976 12 3robots’!!!
2 2 2 2 21 2 49
data all on its
truly
2 18 308 168 124 120 1232746809.310 0 0 0 0 0 0 0 0 0 -32 -32
992 4 3 3 2 2 2 21 2 49 4 19 308 172 124 120 1232746809.320 0 0
0 0 0 0 0 0 0 -32 -32 992 12 2 2 3 3 5 24 2 48 2 18 308 172 120
BUT, how does critterbot sense walls and treats?
120 1232746809.330 0 0 0 0 0 0 0 0 0 -32 -32 992 8 2 2 2 2 2 23 2
Critterbot has a variety of sensors on it to detect a
49 2 17 308 172 120 120 1232746809.340 0 0 0 0 0 0 0 0 0 -32 -32
variety of objects and rays like ‘treats’.
992 12 3 3 3 6 4 25 2 58 3 19 312 168 120 124 1232746809.350 0
0 0 0 0 0 0 0 0 -32 -32 992 4 2 2 2 2 2 25 4 58 3 16 304 172 124
Critterbot
Next we will discuss the variety of sensors in detail:
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Light Sensor (visible light)
Infrared Sensor (distance detector)
Infrared Sensor (heat detector)
Gyroscope Sensor (balance detector)
Accelerometer (acceleration detector)
Sound Sensor (detects sound waves)
Bump Sensor (detect how something feels)
Magnetic Field Sensor
Battery Level Sensor
Motor Temperature Sensor
Critterbot – Light Sensor
Light Sensor
Is used to detect the visible
range of the Electromagnetic
Spectrum which includes the
colors of the rainbow. Not
only does it detect the color,
it can also detect the
brightness of the light source.
Critterbot – Infrared Sensor
Infrared Sensor 1
Critterbot has two types of
infrared sensors, this one uses
a pulse of light to detect the
distance critterbot is to a
specific object.
Critterbot – Infrared Sensor
Infrared Sensor
Used to detect the infrared
range of the Electromagnetic
Spectrum which represents
the amount of heat given off
by an object. They can
detect where any object that
gives off heat (people) are
located.
Critterbot – Gyroscope Sensor
Gyroscope Sensor
Device used for measuring the orientation of critterbot. It
allows critterbot to sense its balance, whether it’s sitting
flat on the ground or travelling up a ramp.
Critterbot – Accelerometer
Accelerometer
Used to detect how fast critterbot is increasing its speed.
Critterbot – Sound Sensor & Bump
Sensor
Sound Sensor
To allow critterbot to ‘hear’ its surroundings. It helps
critterbot gather more information about its
surroundings. Eventually critterbot may be
able to decipher individuals from each other
Bump Sensor
Allows critterbot to detect the amount of pressure
someone is pushing against critterbot with.
Critterbot – Magnetic Field Sensor &
Battery Level Sensor
Magnetic Field Sensor
Critterbot uses this sensor to detect the Earth’s magnetic
field. This sensor can be used as a compass for
critterbot to help figure out its location.
Battery Level Sensor
Used to detect the amount of battery left in
critterbot, a measure if critterbot’s life line.
Critterbot – Motor Temperature
Sensor
Motor Temperature Sensor
To determine how hot the motor is. This sensor can also
be used to determine the ‘tiredness’ of critterbot.
Critterbot
Perhaps one day critterbot will be trained to be useful in
our everyday lives and critterbot will learn to bring you
your newspaper or slippers…
Field Trips for Critterbot
For information regarding field trips to the ‘University of
Alberta – Computer Science Centre’ to see critterbot and
other exciting new research on robotics and Machine
Learning please contact:
Sheryl Mayko
[email protected]
Practice Questions
1. Why does critterbot need so many different types of
sensors?
2. How can critterbot help you with your daily life?
3. Do you think critterbot is the ‘ideal pet’? Why or why not?
4. What other functions could we train critterbot to do to
make it more useful to society?
5. What improvements could be made to critterbot to make
it more useful?
Critterbot
Resources:
http://www.cs.ualberta.ca/~sokolsky/critterbot/
http://rlai.cs.ualberta.ca/RLAI/rlai.html
http://www.cs.ualberta.ca/~sutton/book/ebook/node7.ht
ml
Centre for Mathematics Science and Technology Education (CMASTE)
382 Education South
University of Alberta
Edmonton AB T6G 2G5
www.CMASTE.ca
To download: select Outreach, Alberta Ingenuity Resources and Centre for Machine Learning
Filename: AICML6BrainTumourAnalysis
Reinforcement Learning and Artificial Intelligence
Department of Computing Science
University of Alberta
2-21 Athabasca Hall
Edmonton AB T6G 2E8
(780) 492-5640
http://rlai.cs.ualberta.ca/RLAI/ualberta.html
Centre for Machine Learning
Department of Computing Science
University of Alberta
2-21 Athabasca Hall
Edmonton AB T6G 2E8
(780) 492-4828
www.machinelearningcentre.ca
Alberta Ingenuity
2410 Manulife Place, 10180-101 Street
Edmonton AB T5J 3S4
(780) 423-5735
www.albertaingenuity.ca