Ai in Robotics

Artificial Intelligence
In Robotics
Assistant Prof. Dr. Aybars UGUR
EGE University
Computer Engineering Department
aybars.ugur@ege.edu.tr
Contents
Robotics
Genetic Algorithms
Artificial Neural Networks
EGE University Computer Engineering

September 2008 Department 2
AI
From Academia: AI is…
…modeling aspects of human cognition by computer
…the study of ill-formed problems
…"nothing more" than advanced algorithms research
…cool stuff! Machine learning, data mining, speech,
language, vision, web agents…and you can actually get
paid a lot for having fun!
…what other CS folks don’t yet know how to do, and
we AIers aren’t always too sure either

At least two major "consensus" definitions of intelligence have been
proposed.
1. American Psychological Association in 1995 :

Individuals differ from one another in their ability to understand complex ideas, to
adapt effectively to the environment, to learn from experience, to engage in various forms
of reasoning, to overcome obstacles by taking thought. Although these individual
differences can be substantial, they are never entirely consistent: a given person’s
intellectual performance will vary on different occasions, in different domains, as judged
by different criteria. Concepts of "intelligence" are attempts to clarify and organize this
complex set of phenomena

A second definition of intelligence comes from
"Mainstream Science on Intelligence", which was
signed by 52 intelligence researchers in 1994:
a very general mental capability that, among other things, involves the
ability to reason, plan, solve problems, think abstractly, comprehend
complex ideas, learn quickly and learn from experience. It is not merely
book learning, a narrow academic skill, or test-taking smarts. Rather, it
reflects a broader and deeper capability for comprehending our
surroundings—"catching on", "making sense" of things, or "figuring
out" what to do

Robotics
Robotics is the science and technology of
robots, their design, manufacture, and
application.
The Shadow robot hand

system holding a lightbulb.

Robot
"A reprogrammable, multifunctional
manipulator designed to move material, parts,
tools, or specialized devices through various
programmed motions for the performance of a
variety of tasks"
Robot Institute of America, 1979

Robot Definition
Webster says: An automatic device that performs
functions normally ascribed to humans or a machine in
the form of a human.
The term 'robotics' refers to the study and use of

robots. The term was coined and first used by the
Russian-born American scientist and writer Isaac
Asimov (born Jan. 2, 1920, died Apr. 6, 1992). Asimov
wrote prodigiously on a wide variety of subjects. He
was best known for his many works of science fiction.

Laws of Robotics
Asimov also proposed his three "Laws of Robotics", and he later
added a 'zeroth law'.
Law Zero: A robot may not injure humanity, or, through
inaction, allow humanity to come to harm.
Law One: A robot may not injure a human being, or, through
inaction, allow a human being to come to harm, unless this
would violate a higher order law.
Law Two: A robot must obey orders given it by human beings,
except where such orders would conflict with a higher order law.
Law Three: A robot must protect its own existence as long as
such protection does not conflict with a higher order law.

First Robots
The first industrial modern robots were the
Unimates developed by George Devol and Joe
Engelberger in the late 50's and early 60's. The
first patents were by Devol for parts transfer
machines. Engelberger formed Unimation and
was the first to market robots. As a result,
Engelberger has been called the 'father of
robotics.'

Shakey
1969-1971
Shakey the robot
(Fikes, Hart, Nilsson)

Robot Base:
Fixed v/s Mobile
Robotic manipulators used in Mobile bases are typically
manufacturing are examples of platforms with wheels or tracks
fixed robots. They can not attached. Instead of wheels or
move their base away from the tracks, some robots employ
work being done. legs in order to move about.

Robot Job Categories
Robots can be placed into roughly two categories
based on the type of job they do:
Jobs which a robot can do better than a human.
Here, robots can increase productivity, accuracy,
and endurance.
Jobs which a human could do better than a
robot, but it is desirable to remove the human
for some reason. Here, robots free us from dirty,
dangerous and dull tasks.
Applications of Robots
Space Exploration
Healthcare
Domestic
Military defense
Manufacturing
Social Use
Agriculture

Applications of Robots
Space Exploration

Robotics In Medicine
Robotic Assistant for Microsurgery

Robots In Home
The Roomba is a robotic vacuum cleaner made
and sold by iRobot. The Roomba was first
released in 2002 with updates and new models
released every year since.
As of January 2008, over 2.5
million units have been sold.

Roomba
The unit is a disc, 13.4 inches (34 cm) in diameter and less than
3.5 inches (9 cm) high. A large contact-sensing bumper is
mounted on the front half of the unit, with an infrared sensor at
its top front center.
Unlike the Electrolux Trilobite vacuuming robots, Roombas do
not map out the rooms they are cleaning. Instead, they rely on a
few simple algorithms such as spiral cleaning, wall-following and
random walk angle-changing after bumping into an object or
wall. This design is based on MIT researcher and iRobot CTO
Rodney Brooks' philosophy that robots should be like insects,
equipped with simple control mechanisms tuned to their
environments.

Military Defense

Manufacturing
Uses in Manufacturing:
Spray Coating
Assembling Operations
Material Removal
Cutting Operations
Part inspection, sorting, cleaning, polishing

Factory Robots
Car Production (automobile factories)
Packaging (end of a conveyor belt and placing them
rapidly into boxes)
Electronics: Mass produced printed circuit boards
(PCBs)
Automated Guided Vehicles: Mobile robots,
following markers or wires in the floor, or using vision
or lasers, are used to transport goods around large
facilities, such as warehouses, container ports, or
hospitals.

Robots in Hazardous
Environments
HAZBOT operating in
TROV in Antarctica atmospheres containing
operating under water combustible gases
Robot Examples
Transportation

Personal and Service Robotics
In this project, research focuses on the interaction
between a service robot and the people within a public
environment using speech conversation as a primary
communication tool. Due to the nature of human vocal
communication, it is also necessary to implement a
“body language” reading system – the ability of the
service robot to be able to tell the emotion of a person
speaking, and any hand gestures that may occur during
contact. The vocal communication method requires
various technologies, such as Speech Recognition,
Natural Language Processing, Conversational
Algorithms and Speech Synthesis. As the robot may
work in a noisy environment, a set of sound filtering
algorithms needs to be developed, along with a pattern
recognition system for identifying phonemes in speech.
A non-precise pattern recognition system, such as a
neural network, can be employed within the system, to
produce a speaker independent recognition system.
AIBO (Sony)
http://www.aibo.com
Motivation
Toward the development of soccer-playing
robots that will challenge humans in a match
around 2050.
QRIO is a humanoid robot, equipped with the
latest in advanced recognition, motional control
communications, information technology, and
artificial intelligence.

Qrio (Sony)
Entertaining people by interacting
with them through movements and
speech.
Underlying Qrio's ability to interact with
humans is software that performs face and voice
recognition and speech synthesis. The camera located
behind each of the robot's "eyes" collects the facial image
of a person and sends it on to the face-recognition software,
which compares it with all the faces a Qrio has
encountered. The two cameras also give Qrio stereoscopic
vision, allowing it to determine the location of objects in its
view and to navigate around them.

Qrio
By analyzing the information from the seven
microphones in its head, Qrio can also recognize
the voice of someone it knows and determine
which direction the voice is coming from. The
robot understands spoken commands, can say
more than a thousand words, and is able to learn
new ones.
Intelligent actuators enable Qrio to walk on two feet,
dance, climb and descend stairs, not fall over when
shoved, and even pick itself up when it takes a tumble.
http://www.robotadvice.com/sony-qrio_robot.html
Qrio
When Qrio recognizes words, it responds with an
appropriate answer through a speaker in its mouth, using
pitch and volume, as people do, to express emotion. It also
shows emotion by flashing colored lights that surround its
eyes.
The second problem was to ensure that Qrio can maintain
its balance as it walks, runs, hops, and dances. That was
solved by the development of a motion-control system that
combines information from the sensors with knowledge
about the position and motion of each limb. The system
also calculates what other motions have to occur to keep
Qrio upright and stable and issues commands to the limbs
to execute those motions.
Garbage Collection

Nanobots
Nanobot technology is officially known as Nanotechnology Robots.
Classifying Nanobots is difficult, as “Nanotechnology spans and merges
disciplines dealing with matter at the micro level (physics, chemistry, and
biology) with those dealing with matter at the macro level (engineering,
materials science and computer science).” (www.nanobot.info)
Nanobots are robots created molecule by molecule that execute specific
functions, as a larger robot. In essence, they are simply a robot built out of
atoms instead of other resources.
Nanobots are already in use today – one example is stain-resistant clothing.
Nanobots are what makes it stain resistant!
The future of nanobots holds many various and sundry uses from medical to
military to science fiction for nanobot technology.

Nanobots
Nanorobotics is the still largely hypothetical
technology of creating machines or robots at or
close to the scale of a nanometer (10-9 meters).
Also known as nanobots or nanites, they would
be constructed from molecular machines. So
far, researchers have mostly produced only
parts of these complex systems, such as
bearings, sensors, and Synthetic molecular
motors, but functioning robots have also been
made such as the entrants to the Nanobot
Robocup contest. Researchers also hope to be
able to create entire robots as small as viruses
or bacteria, which could perform tasks on a
tiny scale. Possible applications include micro A nanocar made from
surgery (on the level of individual cells), a single molecule
manufacturing, weaponry and cleaning.

Swarm Robots
Swarm robots: Inspired by colonies
of insects such as ants and bees,
researchers hope to create very large
swarms (thousands) of tiny robots
which together perform a useful
task, such as finding something
hidden, cleaning, or spying. Each
robot would be quite simple, but the
emergent behaviour of the swarm
would be more complex.
Components of robots – I
Actuators
The actuators are the
'muscles' of a robot; the
parts which convert stored
energy into movement. By
far the most popular
actuators are electric
motors, but there are many
others, some of which are
powered by electricity,
while others use chemicals,
or compressed air.
A robot leg,
powered by Air
Muscles.
Components of robots – II
Manipulation
Robots which must work in
the real world require some Grippers: A common effector is the
way to manipulate objects; gripper. In its simplest manifestation it
pick up, modify, destroy or consists of just two fingers which can
otherwise have an effect.
Thus the 'hands' of a robot open and close to pick up and let go of
are often referred to as end a range of small objects.
effectors, while the arm is Vacuum Grippers: Pick and place
referred to as a manipulator. robots for electronic components and
Most robot arms have
replacable effectors, each for large objects.
allowing them to perform
some small range of tasks.
Some have a fixed
manipulator which cannot be
replaced, while a few have Robot hand
one very general purpose holding an egg
manipulator, for example a
humanoid hand.

Manipulators
Material Handling Manipulator
Spot Welding Manipulator Assembly Manipulator

Locomotion - I
Rolling Robots
laser range
finder
sonars
touch sensors
Segway in the Robonaut resting
Robot museum on a Segway
in Nagoya. EGE University Computer Engineering
Locomotion - II
Intelligent RoboChair
Aim of this joint project is to explore and
develop the advanced technology needed for a
high performance low-cost RoboChair which
enables the elderly and disabled to gain
necessary mobility to live independently and
improve their quality of life in the society. This
RoboChair should have a user-friendly man-
machine interface and the ability of avoiding
collision and planning a path. It will be
equipped with a new vision system and a
wireless communication system so that its carer
or relative can monitor and tele-operate it when
necessary.
• intelligent control system to achieve good control stability, fast image

processing capability and autonomous navigation.
• interactive user interface for voice control, emotion and gesture detection,
as well as a 3G mobile phone for carers or relatives to monitor and
communicate remotely. EGE University Computer Engineering
Locomotion - III
Walking
Robots
RobotCub is a 5 years long project funded

by the European Commission. Started in
2004, its goal is to study cognition through
the implementation of a humanoid robot the
size of a 2 year old child: the iCub.
Locomotion - IV
Robotic Fish
The Robotic Fish Project at Essex.
undulation movement
The aim of project is to design and build autonomous robotic fishes

that are able to reactive to the environment and navigate toward the
charging station. In other words, they should have the features such
as fish-swimming behaviour, autonomously navigating ability
Locomotion - V
Two robot snakes. Left one has 32 motors, the
right one 10.

Human Interaction
If robots are to work effectively in
homes and other non-industrial
environments, the way they are
instructed to perform their jobs, and KISMET
especially how they will be told to
stop will be of critical importance. The
people who interact with them may
have little or no training in robotics,
and so any interface will need to be
extremely intuitive. Science fiction
authors also typically assume that
robots will eventually communicate
with humans by talking, gestures and
facial expressions, rather than a
command-line interface. Although
speech would be the most natural
way for the human to communicate, it
is quite unnatural for the robot.
KISMET
Kismet is a robot made in the late 1990's at MIT with
auditory, visual and expressive systems intended to
participate in human social interaction and to
demonstrate simulated human emotion and appearance.
Kismet simulates emotion through various facial

expressions, vocalizations, and movement. Facial
expressions are created through movements of the ears,
eyebrows, eyelids, lips, jaw, and head.

Personality
•Personality: Many of the robots of
science fiction have personality, and
that is something which may or may
not be desirable in the commercial
robots of the future. Nevertheless,
researchers are trying to create
robots which appear to have a
personality: i.e. they use sounds,
facial expressions and body
language to try to convey an internal
state, which may be joy, sadness or
fear. One commercial example is
Pleo, a toy robot dinosaur, which can
exhibit several apparent emotions.
Part - II
Genetic Algorithms and TSP
Neural Networks and Computer Vision
AI in Robotics
Autonomous Robots
Optimization
Path Planning
Learning
Recognition
Natural Language Processing

Genetic Algorithms
A genetic algorithm (GA) is a search
technique used in computing to find exact or
approximate solutions to optimization and
search problems.

Genetic Algorithms
Genetic algorithms are a particular class of
evolutionary algorithms (also known as
evolutionary computation) that use techniques
inspired by evolutionary biology such as
inheritance, mutation, selection, and
crossover (also called recombination).

Genetic algorithm outline
An outline of a genetic algorithm is the following:
Generate random population of n chromosomes
Repeat
Evaluate the fitness for all chromosomes
Select two parent chromosomes from the population
according to fitness
Generate new population using offsprings through
crossover and mutation
Until terminating condition
Population
Chromosome 1 : 1101100100110110
Chromosome 2 : 1101111000011110
Algorithm begins with a set of solutions
(represented by chromosomes) called
population. Solutions from one population are
taken and used to form a new population.

Crossover - I
Chromosome 1 11011 | 00100110110

Chromosome 2 10010 | 11000011110
Sonuç Birey 1 11011 | 11000011110
Sonuç Birey 2 10010 | 00100110110
Crossover operates on selected genes from parent

chromosomes and creates new offspring. The simplest
way how to do that is to choose randomly some crossover
point and copy everything before this point from the first
parent and then copy everything after the crossover point
from the other parent.
Crossover - II
Single point crossover
11001011+11011111 = 11001111
Two point crossover
11001011 + 11011111 = 11011111

Mutation - I
Chromosome 1 1101111000011110
Chromosome 2 1101100100110110
C1 (Mutation) 1100111000011110
C2 (Mutation) 1101101100110110
In case of binary encoding we can switch a few

randomly chosen bits from 1 to 0 or from 0 to 1.

TSP
Traveling Salesman problem (TSP) is one of the
best-known and most-studied combinatorial
optimization problems. Many researchers have
been studying to develop TSP solving methods
since 1950’s and trying to find optimal total
routing cost for a given collection of cities. Cost
can be distance, time, money, etc.

TSP Example
Shortest Route for

22 cities.
GA are used to
solve TSP since
1980’s.

TSP (Graph Theory)
Given a complete weighted graph finding the
Hamiltonian cycle with the least weight is the
equivalent formulation in terms of graph theory.
Vertices represent cities and edges represent
costs between cities.

Application Areas of TSP
The TSP is important not only from a
theoretical point of view, many practical
applications can be modeled as a traveling
salesman problem or as variants of it. A TSP
solver can find direct applications to a number
of fields: route planning (of airlines, buses,
delivery trucks and postal carriers, computer
networks), job scheduling, printed circuit boards
(PCB) drilling, cryptography, etc.

TSP In Robotics
Part Picking
Part Placing
PCB Drilling
…

Neural Networks and Machine
Learning
A neural network (NN) is a machine learning approach inspired
by the way in which the brain performs a particular learning task.
A NN is specified by:
an architecture: a set of neurons and links connecting neurons. Each link
has a weight,
a neuron model: the information processing unit of the NN,
a learning algorithm: used for training the NN by modifying the weights
in order to model the particular learning task correctly on the training
examples.
The aim is to obtain a NN that generalizes well, that is, that
behaves correctly on new instances of the learning task.

NN Definition
S. Haykin Definition :
A neural network is a massively parallel distributed processor
made up of simple processing units, which has a natural
propensity for storing experimental knowledge and making it
available for use. It resemble the brain in two respects:
Knowledge is acquired by the network from its environment
through a learning process
Interneuron connection strengths, known as synaptic
weights, are used to store the acquired knowledge.

Neuron

Artificial Neuron
Inputs Bias
b
x1 w1
Activation
Local function
Field Output
x2 w2 ∑ v ϕ(−) y
M M Summing
function
xm wm
weights

An Artificial Neural Network
Multi layer feed-forward
3-4-2 Network
Input layer Output layer
Gizli Katman
Hidden Layer
Computer Vision
Neural Networks (as an AI technique) are used to
Recognition: one or several pre-specified or
learned objects or object classes can be
recognized. (as classifier)
Identification: An individual instance of an
object is recognized.
Detection: the image data is scanned for a
specific condition.
Navigation
Navigation is the science (or art) of directing the
course of a mobile robot as it traverses the
environment (land, sea or air). Inherent in any
navigation scheme is the desire to reach a
destination without getting lost or crashing into
anything.

Three Tasks of Navigation
Navigation involves three tasks: mapping and
modeling the environment, path planning and
selection and path following. The path following
task is composed of trajectory planning, which
makes the generated path a time parameterized
line, and generating tracking control.

Path Planning
The path-planning problem is usually defined as
follows: “Given a robot and a description of an
environment, plan a path between two specific
locations. The path must be collision-free
(feasible) and satisfy certain optimization
criteria.”. Machalewicz first successfully applied
genetic algorithm in mobile robot navigation.

Autonomous Robots
An autonomous robot is a robot which can
perform desired tasks by perceiving its
environment without continuous human
guidance. Several techniques from artificial
intelligence, such as learning, neural networks
and genetic algorithms, can be applied to
autonomous robots in order to improve their
performance.

http://www.robocup.org/
USC robotics Lab

What is autonomous ground vehicle?
is a vehicle that navigates and drives entirely on
its own with no human driver and no remote
control. Through the use of various sensors and
positioning systems, the vehicle determines all
the characteristics of its environment required to
enable it to carry out the task it has been
assigned.

ALVINN
ALVINN (Autonomous Land Vehicle in a Neural
Network) is an artificial neural network based
perception system that learns to control vehicles by
watching a person drive.
ALVINN has many characteristics which make it
desirable as a robust, general purpose road following
system:
learns the features that are required for driving on the
particular road type for which it is trained.
is computationally simple.
learns features that are intuitively plausible when viewed
by a human.
has been shown to work in a variety of situations.
CMU NavLab

Data
Designed to work on three types of roads
single-lane dirt single-lane paved lined two-lane

access road bicycle path highway

ALVINN: Learning to Drive
(CMU)

Learning
ALVINN learned by watching how a human steered
used additional views of what the road would look like at
positions slightly off course.
After about three minutes of training, ALVINN was able
to take over and continue to drive.
ALVINN has successfully driven at speeds up to 70*mph
and for distances of over 90 miles on a public highway
north of Pittsburgh.
*Actually being inside the vehicle was a big incentive

for the researchers to develop a good neural network!

DARPA Grand Challenge
The DARPA Grand Challenge is a prize
competition for driverless cars, sponsored by the
Defense Advanced Research Projects Agency
(DARPA), the central research organization of
the United States Department of Defense.
Congress has authorized DARPA to award cash
prizes to further DARPA’s mission to sponsor
revolutionary, high-payoff research that bridges
the gap between fundamental discoveries and
their use for national security.
2004-2005
The first Grand Challenge event was held in
March 2004 and featured a 142-mile desert
course. Fifteen autonomous ground vehicles
attempted the course and no vehicle finished. In
the 2005 Grand Challenge, four autonomous
vehicles successfully completed a 132-mile
desert route under the required 10-hour limit,
and DARPA awarded a $2 million prize to
“Stanley” from Stanford University.

2005
The vehicle that completed the course in the shortest
amount of time was “Stanley,” entered by Stanford
University. The team wins the $2 million prize because
it finished the entire course in the shortest elapsed time
under 10 hours – six hours, 53 minutes and 58 seconds
(6:53:58).
Stanford Racing Team
http://www.darpa.mil/GRANDCHALLENGE/overview.asp
Darpa's Robot Car Race: Gentlemen,
Start Your Processors
Following a week of trials by an initial field of 36
autonomous vehicles, 20 finalists will compete
for $3.5 million in the Urban Challenge on
November 3, the third in a series of robot car
races sponsored by the Pentagon's wild-haired
research department, the Defense Advanced
Research Projects Agency, or Darpa.

2007
A driverless Chevrolet Tahoe developed by a Carnegie
Mellon University-based team won a 2 million U.S.
dollar top prize in a Pentagon-sponsored robot car race,
organizers announced Sunday. (2007)
In a race that began and ended in a casino parking lot
and traversed 132 miles of desert southwest of Las
Vegas on Oct. 8, the Stanford Racing Team's
autonomous robotic car, Stanley, won big. The
artificially intelligent car traversed the off-road course in
a little less than seven hours, yielding both a $2 million
payout and a lofty place in the history of robotics and
technology. (2005)

DARPA Grand Challenge : Future
In addition to simply finishing the 60-mile
racecourse in less than six hours, the cars will
have to follow all the rules of the road as
outlined in the California Driver Handbook.
The robot vehicles will not be allowed to blaze
through stop signs or cut off other cars, and
there certainly will be no speeding.

Conclusion
Robots are used in various areas today. AI
capabilities are not advanced level.
As technology increases, AI techniques (NN,

GA, ES, NLP,…) will be more important for
robotics.
Especially, in recognition, communication,

learning, optimization, planning, path planning,..
References
Aybars UGUR, Artificial Intelligence Course
Notes, EGE University Computer
Engineering Department, 2008
Russell, S.J. And Norvig, P., “Artificial
Intelligence : A Modern Approach, Second
Edition”, Prentice-Hall, 2003. (AIMA)
Sookram Sobhan, Introduction to Robotics,
Polytechnic University.
Wikipedia, 2008.

Ai in Robotics

Uploaded by

Copyright:

Available Formats

Ai in Robotics

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ai in Robotics

Uploaded by

Copyright:

Available Formats

Artificial Intelligence

EGE University Computer Engineering

EGE University Computer Engineering

1. American Psychological Association in 1995 :

EGE University Computer Engineering

EGE University Computer Engineering

The Shadow robot hand

EGE University Computer Engineering

Robot Institute of America, 1979

EGE University Computer Engineering

The term 'robotics' refers to the study and use of

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

Material Handling Manipulator

Spot Welding Manipulator Assembly Manipulator

EGE University Computer Engineering

• intelligent control system to achieve good control stability, fast image

RobotCub is a 5 years long project funded

The Robotic Fish Project at Essex.

The aim of project is to design and build autonomous robotic fishes

EGE University Computer Engineering

Kismet simulates emotion through various facial

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

Chromosome 1 11011 | 00100110110

Crossover operates on selected genes from parent

Two point crossover

11001011 + 11011111 = 11011111

EGE University Computer Engineering

In case of binary encoding we can switch a few

EGE University Computer Engineering

EGE University Computer Engineering

Shortest Route for

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering

EGE University Computer Engineering