Human Vs. Robot Intelligence Workshop

embarrassedlopsidedΤεχνίτη Νοημοσύνη και Ρομποτική

14 Νοε 2013 (πριν από 3 χρόνια και 5 μήνες)

62 εμφανίσεις

Human Vs. Robot Intelligence

Workshop

“Exploring Roomba, the Vacuum Robot”

Objective
:

1.


Practice problem s
olving
,

team work
, and the scientific method.

2.

Learn about algorithms and compare the process in

computers with the human brain
.

3.

Observe and talk abou
t the real applications of robots
.

Supplies
:

1.

Roomba

2.

Swivel Sweeper Vaccum

3.

C
onfetti

4.

Blindfold

5.

Barrier:

8

1

x6
” boards
cut into 3 ft lengths

6.

Poster

7.

Tape to hang poster

For each student:

1.

Clip Board

2.

Pen or Pencil

3.

Laboratory paper

4.

Pre
-
quiz handout

5.

Post
-
quiz h
andout



Figure 1: Supplies

Estimated time
: 30
-
45

minutes

Size of Audience
:
10
-
20

Targeted Age Group:

11
-
17 years

Parts of the Presentation
:

1.

Pre
-
Quiz
-

give audience the attached quiz before the presentation, to compare to the

quiz after the
presentation, so that you can judge what they learned from the presentation.

2.

Explanation of the parts of the Roomba

using the poster

(See Figure 2
)

**make sure to cover the search pattern answer on the poster
a.

Hardware

i.

Embedded computer
-

the

“brain” of the robot, where the algorithm, or
instructions that the robot uses is stored and read.

ii.

Wheel Motors
-

the “legs” of the robot, each of the wheels have their own
motors
.

iii.

Brushes
-

“tools of the robot”;

how the Roomba does work. Big brushes in th
e
middle pick up most of the dirt, but there are small brushes on the side to pick
up dirt along the wall. The Roomba does have vacuuming capabilities, but
mostly just sweeps up the dirt.

iv.

Battery
-
energy source for the robot

“food for the robot”
.
There a
re

12 cells and
each cell holds 1.25 volts
. Battery is recharged by the docking station.

b.

Software

i.

D
irt detection
-

There is a larg
e metal plate inside the robot
that
makes a sound

when pieces of dirt/debris are picked up and hit against the metal. This i
s how
the Roomba can
“hear”

when it is in an area of heavy dirt.

ii.


Search

pattern
-

explain the concept of the robot having a search patter
n; that
is, that the Roomba must vacuum the entire room without knowing the size,
shape or the placement of furniture i
n the room.

Don’t

tell

details of the
pattern;

part of the lab
will
be
to
observ
e and understand

the search pattern.

c.

S
ensors

i.

B
umper
-

There are two bumper sensors on the front of the Roomba. It can
differentiate between the left and the right side and mov
e accordingly. If hit on
the left side, usually it will turn a random angle to the right, and vice
-
versa for
the right side.

ii.

D
irt
-

described above

iii.

C
liff
-

Two censors on either side of the front of the Roomba, so that it will stop if
it no longer senses th
e floor under it. So if it is picked up or comes to a ledge it
will stop.

iv.

I
nfrared
-

can detect infrared signals, so that a remote can be used to control it,
or it can sense
the infrared “walls”. When it comes to one of the infrared wall
beacons, it will
stop about a foot from the center of the beacon and treat it as if
it was a
w
all.
3.

Experiment (students will follow
the laboratory worksheet
)
.

a.

Algorithm

i.

Have the audience read the definition of an algorithm.

ii.

Then explain how a robot has to be told how to be
have, and that is done with an
algorithm.

iii.

Have the students record a description of the algorithm a blindfolded person would
follow to vacuum an unknown room

iv.

Have the students record how a robot, with the hardware capabilities discussed on the
poster, woul
d vacuum the same room.

b.

Procedure and
Observations

Notes

i.

For the blindfolded person, make sure that the rest of the group is not just giggling at
the awkwardness of a blindfolded person. Remind them to observe in what way the
person goes about vacuuming a
nd what spots the person gets and what spots they miss.

ii.

Discuss how the person might have done better.

iii.

Observe the search pattern of the Roomba.

c.

Comparisons

i.

How did the person know where they had been and where they hadn’t?

1.

Used the boards as a guide for t
he shape of the room. Similar to the
wall
following
behavior

of the Roomba search pattern.

2.

Listened for the sound of confetti being vacuumed. Similar to the dirt detect on
the Roomba.

ii.

What spots did both the Roomba and the person miss?

1.

Corners, because of
the difficulty of reaching.

2.

Places in the middle, the robot is programmed to sweep the
middle of the room
in a random criss
-
cross method, so it is likely to miss spots in the middle.
Hopefully the human was able to remember where in the room he/she had
be
en.

d.

Discuss results

i.

Search pattern (See Figure
3
)

1.

The search pattern

for the Roomba

is mostly
random
. When placed in the
center of a room it
spirals out
until it hits something. Then it can either
criss
-
cross

though the center of the room, or
hug the wal
ls
using the w
all sensors on
its right side.

2.

The human had many similar techniques to the Roomba. The Roomba had to be
told what to do for the search pattern, but the human could think on its own.

ii.

Post
-
Quiz


Figure 2: Poster





Figure
3: Close up of
Search Pattern on Poster


Results f
rom Trial Exp
e
riment:

(April 10
th

2010, 4 groups of 10
-
15 girls ages 11
-
17)

1.

Hypotheses

a.

Example Human Hypotheses

i.

“They would feel around to find where to vacuum at.”

ii.

“People could tell them where to go.”

iii.

“Guess by feel
ing with the vacuum cleaner to know if they ran into anything.”

b.

Example Robot Hypothese

i.

“Go around edge, then go a little closer to the middle each time.”

ii.

“Use a camera or find where the dirt is.”

iii.

“Uses the bumpers to make a map of the room and use the inf
rared to see where it’s
been.”

c.

In general the audience had a clear idea of
how a human might behave

when trying to clean a
room without sight, but had vague ideas
of how the robot might behave
. Although there was
some understanding of an algorithm beforeh
and.

2.

Observations during the experiment

a.

Participants were
more into the humor of bumping into things, than in finding the best way to
vacuum the room. Perhaps offer a prize if they vacuum all of the space.

b.

The
participants

recording the observations would
n’t get involved in the experiment and would
get left out. Perhaps provide a clip board and encourage them to be involved.

c.

The
participants

enjoyed the interaction with the robot. Perhaps pass around the Roomba so
that they can see the underside. Point
out where the different sensors are, and where the
batteries and brushes are.

3.

Results from the quiz

a.

Question 1:
What is an algorithm?

i.

“A” was given if there was more than a regurgitation of the definition on the paper
given.

Ex: “A certain patter or sequen
ce taken to complete a task.”

ii.

“C” was given if some understanding of the definition was shown, but could have just
been regurgitation.

Ex. “A simple set of rules.”

iii.

“F” was given in no understanding of the definition of what an algorithm is.

Ex. “A rule.”

b.

Q
uestion 2:
What is the brain of the robot?

i.

“A” was given if something about a computer was mentioned.

Ex. “embedded computer”

ii.

“C” was given if there was some ambiguity in the answer.

Ex. “What tells the robot what to do, allows the robot to function”

iii.

“F” w
as given if nothing was mentioned about the computer

Ex. “The batteries”

c.

Question 3:
How would you tell a robot in human form to find the base of the stairs, go up a
flight of stairs and stop at the top

if the robot was just an empty shell and didn’t know
what a
set of stairs were or even how to walk?

i.

“A” was given if all of the aspects of the problem were covered, and a full understanding
of the concept of an algorithm was demonstrated. This usually involved some type of
repeating in the instructions.

Ex.

“1) Lift one leg up then the other. 2) Feel for the rail. 3) Grab the rail. 4)
Continue to do #1 until the rail stops”

ii.

“C” was given if some understanding of an algorithm was demonstrated and most of the
aspects of the problem were covered.

Ex. “View Stai
rs. Lift foot to 1
st

stair. Lift other foot to meet the first. Repeat
until no more steps are seen.”

iii.

“F” was given if no understanding of an algorithm was demonstrated.

Ex. “He has a camera cause that way the robot knows when to move and where
somethin
g is at.”

d.

Results are summarized in Figure
4

below.

e.

Overall Results

i.

A majority of the audience gained knowledge about the hardware of the Roomba and
robots in general

ii.

A majority of the audience gained some basic idea of how to write an algorithm or
“pseudo

code”. The robot lost its “magic” thinking qualities.

Figure
4
:

Quantitative

Post

Quiz Results

(no pre
-
quiz was given)