�ݺ�ߣ

AI: Home Automation
Patricio Ricaud A01192626
Oscar Romero A01192355
Mauricio Diaz A01192389

Routine Recognition
Humans are noted to having
well organized routines or
patterns in their lives. It is
then easy to assume that
these routines could be
hardwired into code so that
they could be memorized or
applied by an algorithm.

Routine Recognition
In a smart home, or a living place with an intelligent and learning
algorithm, it is possible to record and learn from certain routines a
person may have such as:
?? A morning routine: The smart home can prepare things such as
alarms, coffee, breakfasts.
?? Afternoon routine: The home will detect when you arrive home,
will prepare any activities you have planned, such as meals,
showers, reminders of future plan, such as work or hobbies.
?? Night routine: Reminder of nightly activities such as brushing your
teeth or even setting up a channel you want to watch on TV or a
form of entertainment.
The home does this with a supervised approach, meaning that it
monitors when the person enters or leaves a room, turns on a light
or a device, or made automatically by time.

Person Recognition
Person recognition involves identifying a person or a
characteristic of a human through a frame often
featuring facial structures. This can be known as
biometrics, or the recognition of humans through
their traits.At the moment it is common for this type
of technology to be present through security systems
such as fingerprint or iris recognitions. This, though,
can be taken to another level in order to recognize
someone not only for security reasons, but for
household reasons. Full facial recognition is not
something new in the present, as it can be seen used
in Facebook while tagging pictures of others, or
while using Kinect in your XBOX.

Person Recognition
?? Taken a step further, person recognition can be
able to determine a person from a still frame of
his body or given a video of his or her body
language. It can help tremendously as the
��mind�� of the house can learn likes and dislikes
of recursive persons of this home and know if
they like a cold room or a hot one. It can also
detect new people, that if accompanied by the
owner pose no threat, but if they enter alone
the home knows it is an uninvited stranger.
?? Together with routine recognition, a smart
house would be able to deduce the activity that
the person wants to perform. For example, if
the owner of the house enters the kitchen at
8:00 P.M. the smart house can know that he is
going to cook dinner, so it can start preparing
the oven, stove or other kind of appliances.

Person Recognition
?? Another use can be in the energy saving department,
since the house can turn on and off air conditioning in
rooms it knows it is most needed. If the owner leaves
his bedroom after his morning routine the house can
know that it can safely shut down the air conditioning
in that room. If it is late in the evening and the owner
steps into the living room, the house can assume this
person will watch TV or stay there for a relatively long
time, so it can turn on the AC to the appropriate
temperature.

Voice Control
?? Tech to control home automation
with voice.
?? Current examples: VoicePod, Tasker
+VeraLite, enBlink.
?? Now: mobile device as the medium.
?? Future: no mobile device (ex. Moto X)

CS Behind Voice Control
?? Natural Language Processing
?? Attempt to understand language through
probability.
?? Example: ��I ate cherry�� vs. ��Eye eight Jerry��
o? Uses context to understand.
?? Language Modeling: N-Grams
?? Vector Space Model

Nest thermostat
?? Programmable and self-
learning thermostat
?? After being set for certain
temperatures at certain
times
o? The thermostat starts adjusting
itself
o? Develops a temperature schedule

Intelligence in Nest Thermostat
?? Records an Away temperature
o? Records variations made by the user during
the day
o? After a couple of days an algorithm develops
a schedule based on user preferences
?? Ability to be modified from mobile devices
o? Via the internet the thermostat can be
modified
o? Stores energy usage so the user can have
access to it
o? Modifies its schedule and behavior in order
to conserve energy and reduce electricity

Intelligence in Cleaning Robots
?? Ex: iRobot, Infinuvo, Hom-bot
?? Computer Vision
o? Localization
o? To understand surroundings
?? Search
o? Cost function
o? Most efficient path
o? Decision making

ASIMO
?? Advanced Step in Innovative
MObility
?? Designed to resemble humans and
help them in their tasks
?? Functions
o? Hand and Arm mobility with 34 degrees of
freedom
o? Ability to carry trays
o? Ability to push carts
o? Human recognition

Intelligence in ASIMO
?? Learns constantly from his
environment
o? Capable of choosing the best route
to a point
o? Avoid movable and static obstacles
?? Facial recognition and voice
recognition
o? Personalized interactions with
different people
?? Identify unknown sounds and turn
towards the source of sound
o? Take decisions of action concerning

Uses for ASIMO
?? Dangerous jobs for humans
o? Dealing with toxic materials
?? Helping in hospitals
o? Its ability to carry trays, and facial recognition allow him to
take medicines or other objects to specific patients.
o? Also could push around wheelchairs without crashing and
deciding the best route to follow.
?? Home help
o? This humanoid can perform human daily chores

The Future for ASIMO
?? Honda is looking forward to develop its
intelligence even further
o? Ability to make judgements when confronted with a
given situations
��?? Could take a good or bad choice
?? Implications of bad choices could be great, given he is
dealing with humans
o? i.e. If a medicine ends he could give more importance
to administering a medicine than administering the
RIGHT medicine

References
Hara, Yoshiko. "'My Name is Asimo, I'Ll be Your Server Tonight'." Electronic Engineering Times.1402 (2005): 8. ProQuest.Web. 12
Mar. 2014.
"Honda Upgrades Asimo Robot into Speedy Errand Assistant." TechWeb Dec 13 2005: 1. ProQuest. Web. 12 Mar. 2014 .
Hara, Yoshiko, and Hiroaki Kitano. "'Personal Robots' Get Ready to Walk on the Human Side / Comment." Electronic
Engineering Times (2002): 157-62. ProQuest. Web. 12 Mar. 2014.
"Lowe's Rolls Out Nest Learning Thermostat." Manufacturing Close - Up (2012)ProQuest. Web. 12 Mar. 2014.
"Teaching Nest to save Energy." Nest Labs. Web. 12 Mar. 2014. <
https://nest.com/blog/2011/11/08/teaching-nest-to-save-energy/>.
"Asimo, The World's Most Advanced Humanoid Robot." ASIMO by Honda. Honda. Web. 12 Mar. 2014. <
http://asimo.honda.com/>.
King, Rawlson. "Explainer: Retinal Scan Technology." BiometricUpdate.com. N.p., 12 July 2013. Web. 12 Mar. 2014.
Collins, Michael. "Language Modeling." Cs.columbia.edu. Columbia University, n.d. Web. 3 Feb. 2014.
Manning, Christopher. "Natural Language Processing." Stanford University, 2012. Web. 3 Feb. 2014. Lecture.
Mooney, Raymond J. ��N-Gram Language Models.�� The University of Texas at Austin. PPT file.
Chua, Sook-Ling, Stephen Marsland, and Hans W. Guesgen. "Behaviour Recognition in Smart Homes." Massey University,
n.d. Web. 11 Mar. 2014.
"The International Biometric Society ? Definition of Biometrics." The International Biometric Society. IBS, n.d. Web. 13 Mar.
2014.

�ݺ�ߣ

Ai home automation

Recommended

More Related Content

Viewers also liked (9)

Similar to Ai home automation (20)

Recently uploaded (20)

Ai home automation