robotics and artificial intelligence, or artificial...

Robotyka i Sztuczna Inteligencja

Robotics and Arti�cial Intelligence, orArti�cial Intelligence as a Goal and Method of Robotics

Part 1

Jacek Malec

Robotics and Semantics Systems,

Department of Computer Science,

LTH, Lund University,

Lund, Sweden

March 29-30, 2010

Jacek Malec, RSS, CS. LTH 1(92)


Umowa

B�ed�e (próbowa�) mówic po polsku.

Slajdy b�ed �a po angielsku.

Pytania mo�zna zadawac w dowolnym momencie.

W dowolnym zrozumia�ym mi j �ezyku:-)

Najpewniejszy kontakt: [email protected].



Outline

1 Introduction: Who am I?

2 What is arti�cial intelligence? De�nitions. Discussion.

3 Autonomous robots as a tool for AI research.

4 Problems with autonomy: thinking vs. reacting. Brooks.

5 Agent-based models.

6 Arti�cial intelligence: state of the art, challenges.

7 Progress stimulated by competitions: RoboCup, FIRA,

ELROB, Grand Challenge ...

8 Advanced applications of industrial robots, AI in automation.

9 Industrial robots with intelligent extensions.

10 Service robots: demands, expectations, today.

11 Should robots be intelligent? Discussion.



Outline

1 Introduction

2 What is arti�cial intelligence?

3 Autonomous robots in AI

4 Thinking vs. acting

5 Agent-based models

6 AI: state of the art, challenges

7 Robotic competitions

8 AI and robotics in automation

9 Industrial robots with (some) intelligence

10 Service robots

11 Should robots be intelligent?



Who am I?

1981, mgr in�z., Elektronika, PWr

1987, dr in�z., ICT, PWr

promotor: prof. Jerzy Jaron

1990�98, Linköping

1995�99, Västerås

1999� , Lund

2001, docent



My University

Lund, 110k inhabitants

University grounded in 1666

40k undergraduate students (largest in Scandinavia)

2300 PhD students

6000 employees

Computer Science belongs to Technical Faculty



Research Groups @ CS, LU

Computer Graphics (Tomas Akenine-Möller)

3D Graphics for Mobile PlatformsShadow Rendering

Embedded Systems Design (Krzysztof Kuchcinski)

Design MethodsConstraining ProgrammingRecon�gurable computing

Robots and Semantic Systems (Klas Nilsson)

Arti�cial IntelligenceNatural Language ProcessingRobotics and Automation Software



Research Groups @ CS, LU (cont'd)

Software Development and Environments (Boris Magnusson)

Pervasive systemsLanguage EnvironmentsSoftware Con�guration ManagementReal-time programming

Software Engineering (Per Runeson)

Empirical Software EngineeringRequirements EngineeringSoftware Process QualityVeri�cation and Validation



Outline

1 Introduction









10 Service robots




Intelligence

inteligencja [�ac. intelligentia �zdolnosc pojmowania�, �rozum�],

zool. umiej �etnosc niektórych zwierz �at szybkiego znajdywania

najtrafniejszego post �epowania w nowej dla nich sytuacji.

Zród�o: Encyklopedia PWN



Intelligence

inteligencja [�ac. intelligentia �zdolnosc pojmowania�, �rozum�],

psychol. cecha umys�u warunkuj �aca sprawnosc czynnosci

poznawczych, takich jak myslenie, rozwi �azywanie problemów; od

inteligencji zale�zy sprawne korzystanie z nabytej wiedzy, a tak�ze

skuteczne zachowanie si �e wobec nowych sytuacji i zadan.




Intelligence

inteligencja emocjonalna, psychol. zdolnosc rozumienia i kontroli

w�asnych emocji; skutkiem rozwoju i.e. jest m.in. zwi �ekszona

wra�zliwosc na emocje prze�zywane przez inne osoby; poj �ecie

�inteligencja emocjonalna� rozpropagowa� D. Goleman w pracy

Inteligencja emocjonalna.




Intelligence quotient

iloraz inteligencji, ang. Intelligence Quotient (IQ ), psychol.

liczbowy wskaznik, miara poziomu inteligencji; uzyskiwany na

podstawie badania testowego inteligencji; ok. 68% populacji

osi �aga iloraz inteligencji w przedziale wartosci 85�115.




Arti�cial Intelligence

sztuczna inteligencja (SzI), ang. Arti�cial Intelligence (AI), dzia�

informatyki, którego przedmiotem badan s �a regu�y rz �adz �ace

inteligentnymi zachowaniami cz�owieka (np. postrzeganiem,

uczeniem si �e), tworzenie modeli formalnych tych zachowan i

symuluj �acego je oprogramowania.





Polish Wikipedia: terrible! Don't read! Forget, if you already have

read it! Edit!

English Wikipedia: good, extensive overview.

Check also: AI Topics at http://aaai.org.




Arti�cial intelligence (AI) is the intelligence of machines and the

branch of computer science that aims to create it. Textbooks de�ne

the �eld as "the study and design of intelligent agents,"where an

intelligent agent is a system that perceives its environment and

takes actions that maximize its chances of success. John

McCarthy, who coined the term in 1956, de�nes it as "the science

and engineering of making intelligent machines."



What is AI?

Systems that think like humans

Systems that act like humans

Systems that think rationally

Systems that act rationally



Acting humanly: The Turing test

Turing (1950) �Computing machinery and intelligence�:

�Can machines think?� ! �Can machines behave

intelligently?�

Operational test for intelligent behavior: the Imitation Game

Predicted that by 2000, a machine might have a 30% chance

of fooling a lay person for 5 minutes

Anticipated all major arguments against AI in following 50

years

Suggested major components of AI: knowledge, reasoning,

language understanding, learning

Problem: Turing test is not reproducible, constructive, or amenable

to mathematical analysis



Thinking humanly: Cognitive Science

1960s �cognitive revolution�: information-processing psychology

replaced prevailing orthodoxy of behaviorism

Requires scienti�c theories of internal activities of the brain

What level of abstraction? �Knowledge� or �circuits�?

How to validate? Requires1 Predicting and testing behavior of human subjects (top-down)

or2 Direct identi�cation from neurological data (bottom-up)

Both approaches (roughly, Cognitive Science and Cognitive

Neuroscience) are now distinct from AI

Both share with AI the following characteristic: the available

theories do not explain anything resembling human-level

general intelligence

Hence, all three �elds share one principal direction!



Thinking rationally: Laws of Thought

Normative (or prescriptive) rather than descriptive

Aristotle: what are correct arguments/thought processes?

Several Greek schools developed various forms of logic: notation

and rules of derivation for thoughts; may or may not have

proceeded to the idea of mechanization

Direct line through mathematics and philosophy to modern AI

Problems:

1 Not all intelligent behavior is mediated by logical deliberation

2 What is the purpose of thinking? What thoughts should I have

out of all the thoughts (logical or otherwise) that I could have?



Acting rationally

Rational behavior: doing the right thing

The right thing: that which is expected to maximize goal

achievement, given the available information

Doesn't necessarily involve thinking�e.g., blinking re�ex�but

thinking should be in the service of rational action

Aristotle (Nicomachean Ethics):

Every art and every inquiry, and similarly every action and pursuit,

is thought to aim at some good



AI prehistory

Philosophy logic, methods of reasoning

mind as physical system

foundations of learning, language, rationality

Mathematics formal representation and proof

algorithms, computation, (un)decidability,

(in)tractability

probability

Psychology adaptation

phenomena of perception and motor control

experimental techniques (psychophysics, etc.)

Economics formal theory of rational decisions

Linguistics knowledge representation

grammar

Neuroscience plastic physical substrate for mental activity

Control theory homeostatic systems, stability

simple optimal agent designs



A short history of AI

McCulloch & Pitts: Boolean circuit model of brain (1943)

Turing's �Computing Machinery and Intelligence� (1950)

Early euphoria (1952�69)

Early AI programs, including Samuel's checkers program, Newell &

Simon's Logic Theorist, Gelernter's Geometry Engine (1950s)

Dartmouth meeting: �Arti�cial Intelligence� adopted (1956)

Robinson's complete algorithm for logical reasoning (1965)

AI discovers computational complexity

Neural network research almost disappears (1966�74)

Early development of knowledge-based systems (1969�79)



A short history of AI

Expert systems industry booms (1980�88)

Expert systems industry busts: �AI Winter� (1988�93)

Neural networks return to popularity (1985�95)

Resurgence of probability; general increase in technical depth

�Nouvelle AI�: ALife, GAs, soft computing (1988� )

Agents, agents, everywhere . . . (1995� )

Human-level AI back on the agenda (2003� )



The Chinese Room

INTELLIGENT MACHINES ARE NOT POSSIBLE!

or, more exactly, conscious machines are impossible.

John Searle, philosopher, UC Berkeley



Outline

1 Introduction









10 Service robots




Prehistory

Karel �Capek, R.U.R. (Rossum's Universal Robots), 1921

Norbert Wiener, Cybernetics, 1940

William Grey Walter, turtles, 1950s



The turtle



Shakey

Stanford Research Institute, 1969



Movie Time

Shakey

Flakey

RoboCup cases

Kismet

Stanley

Cogniron

The Bar-Bot



Outline

1 Introduction









10 Service robots




Autonomy

Autonomia (z greckiego) � zgodnie ze swym zród�os�owem

oznacza tworzyc dla siebie w�asne prawo.

W naszej dziedzinie oznacza to samodzielne podejmowanie

decyzji przez system, na podstawie wiedzy o otoczeniu i

aktualnego stanu.

Autonomia cz�esto zwi �azana jest z celowosci �a dzia�an: system jest

wtedy okreslany mianem teleologicznego.



How to achieve autonomy?

What mechanisms constitute autonomous behaviour?

What is a necessary condition for intelligent behaviour?

How to implement rationality?

The �rst attempt: SEARCH!



Autonomous robots

Can make decisions on their own.

Why do they need to? Because of the following properties of real

environments (cf. Russell and Norvig):

the real world is inaccessible;

the real world is nondeterministic;

the real world is nonepisodic;

the real world is dynamic;

the real world is continuous.



Sense, Think, Act

1 First, do perception;

2 Then, update the world model;

3 Then, plan what to do next;

4 Then, translate your plan to motion commands;

5 Finally, execute the commands.



Physical symbol system, 1/2

A physical symbol system consists of a set of entities, called

symbols, which are physical patterns that can occur as

components of another type of entity called an expression (or

symbol structure).

Thus, a symbol structure is composed of a number of instances (or

tokens) of symbols related in some physical way (such as one

token being next to another).

At any instant of time the system will contain a collection of these

symbol structures.



Physical symbol system, 2/2

Besides these structures, the system also contains a collection of

processes that operate on expressions to produce other

expressions: processes of creation, modi�cation, reproduction and

destruction.

A physical symbol system is a machine that produces through time

an evolving collection of symbol structures.

Such a system exists in a world of objects wider than just these

symbolic expressions themselves.



The physical symbol system hypothesis

A physical symbol system has the necessary and suf�cient means

for general intelligent action.



Requirements

Multiple goals

con�ictscontext-dependent interdependencies

Multiple sensors

Robustness

Extensibility

Purposefulness

to cope appropriately and in timely fashion with changes in the

dynamic environment



Rodney Brooks, 1985



System architectures

sense - think - act (serial decomposition, functional

decomposition);

parallel decomposition (e.g. subsumption, more general:

behaviour-based control);

hybrid, mixed, layered.



Subsumption

horizontal vs. vertical decomposition

a system is more than a sum of its parts (emergent

intelligence)

each behaviour can sense the environment and generate a

physical action



Assumptions behind subsumption architecture

complex behavior needs not be a product of complex control

system

things should be simple

map making is important

map should be 3D

map should be relational

the world does not consist of polyhedra

sonar data does not lead to rich descriptions of the world

failure recovery should be quick

robots should be self-sustaining



Behavioural decomposition

Avoid contact with objects

Wander aimlessly around without hitting things

�Explore� the world by seeing places in the distance that look

reachable and heading for them

Build a map of the environment and plan routes from one

place to another

Notice changes in the �static� environments

Reason about the world in terms of identi�able objects and

perform tasks related to certain objects

Formulate and execute plans that involve changing the state of

the world in some desirable way

Reason about the behaviour of objects in the world and modify

plans accordingly



Layered Architectures



Symbol System Hypothesis

Main problems:

interface between perception and symbols

inadequacy of symbols

symbol systems rely on emergent properties (search)



Physical Grounding Hypothesis

situatedness

�the world is its own best model�

embodiment

intelligence

�intelligence is determined by the dynamics of

interaction with the world�

emergence

�intelligence is in the eye of the observer�



Open Source

ROS 1.0 (Willow Garage)

Tekkotsu 4.0.1 (CMU)

Player/Stage/Gazebo 3.0

GenoM 3.0 (LAAS)

Orocos 2.0 RTT

...



Collective intelligence

An example: warrior ants, army ants




A soldier:




An intelligent entity: the nest



Outline

1 Introduction









10 Service robots




An agent

[Wooldridge, Reasoning about Rational Agents, MIT Press, 2000]

Agents are active, purposeful originators of action. These

actions are performed in order to modify and shape the

environment inhabited by the agent.

Our focus: computer systems capable of independent,

autonomous action in order to meet their design objectives or,

in other words, capable of deciding for themselves what to do

in any given situation.



A rational agent

[Wooldridge, 2000]

An agent is said to be rational if it chooses to perform actions that

are in its own best interests, given the beliefs it has about the world.

Properties of rational agents:

Autonomy (they decide);

Proactiveness (they try to achieve their goals);

Reactivity (they react to changes in the environment);

Social ability (they negotiate and cooperate with other agents).



Agent (Ferber, 1/2)

An agent is a physical or virtual entity

which is capable of acting in an environment,

which can communicate directly with other agents,

which is driven by a set of tendencies (in the form of individual

objectives or of a satisfaction/survival function which it tries to

optimise),

which possesses resources of its own,

which is capable of perceiving its environment (but to a limited

extent),



Agent (Ferber, 2/2)

An agent is a physical or virtual entity

which has only a partial representation of its environment (and

perhaps none at all),

which possesses skills and can offer services,

which may be able to reproduce itself,

whose behaviour tends toward satisfying its objectives, taking

into account of the resources and skills available to it and

depending on its perception, its representations and the

communication it receives.



ROBOT

An active, arti�cial agent whose environment is the physical world.



Goal-based agent



Utility-based agent



Logical characterisation of agents

How can we characterize rationality?

Crucial terms: beliefs and best interests;Decision theory, game theory;Logic;Epistemic logics;Temporal logics, dynamic logics, ...

How can we build rational agents?

Software architectures;Executable speci�cations;Logic;



Logic

Syntax:

theoremhood, proof theory;

Semantics:

model, validity;

Correspondence: soundness, completeness.

Computational aspects:

Complexity of proofs;Automatisation of theorem proving:

Automatisation of inference;

Model checking.

STATIC!



The omniscience problem

All normal modal logics obey the following:

K axiom:

j= �(�! �)! (��! ��)

necessitation rule:�

��

Consequences?

In�nite set of beliefs;

Consistency of beliefs;

Equivalence of beliefs.



Multi-agent systems

Important in large number of application domains

logistics

auctions, trade

e-commerce

computer games

(military) robotics

...



Multi-robot systems

Interesting for a number of reasons:

Performance - many robots may do the job faster, with less

effort.

Sometimes only many robots can do the job (if they are

heterogenous or if the deadline is hard);

Reliability and robustness - when one robot fails, the rest may

do the job;

Adaptivity - robots exposed to different environmental

conditions can learn appropriately (and even communicate the

results to others);

Special case of faults: communication faults not occuring in a

single-robot case.



Interaction, Coordination, Cooperation

Interaction: common resources

antagonisticnon-antagonistic

Coordination: planning for use of common resources

Cooperation: planning for maximisation of utility

eusocial behaviourcooperative behaviour



Taxonomy

cooperation

knowledge

coordination

organization

communication (direct, indirect - stigmergy)

team composition (homogenous, heterogenous)



Control system design

behaviour-based control

sense-model-plan-act

layered (hybrid)

i.e., as usual, but now we have many of them!



Designing MRS

1 Multi-robot system design is inherently harder than design of

single robots;

2 Multiple robots may distract activities of each other, in the

extreme precluding the team from achieving the goal of the

mission;

3 A team may have problems with recognizing the case when

one or more team members, or the team as a whole, become

unproductive;

4 The communication among the robots is a nontrivial issue

(what, to whom and when to communicate?);

5 The �appropriate� level of individualism and cooperation within

a team is problem-dependent.



Cooperation without communication

A �rst proof-of-concept:

Barry Brian Werger: Cooperation without deliberation: A minimal

behavior-based approach to multi-robot teams

Published in Arti�cial Intelligence Journal, vol. 110 (1999),

293�320

A team of behaviour-based robots, playing football

cooperating without any explicit communication



Cooperation without deliberation



Outline

1 Introduction









10 Service robots




State of the art

Which of the following can be done at present?

Play a decent game of table tennis T

Drive safely along a curving mountain road T

Drive safely along Kazimierza Wielkiego N

Buy a week's supply of groceries on the web T

Buy a week's supply of groceries at Pasa�z Grunwaldzki N

Play a decent game of bridge T

Discover and prove a new mathematical theorem M

Design and execute a research program in molecular biology

M



State of the art

Which of the following can be done at present?

Write an intentionally funny story N

Give competent legal advice in a specialized area of law T

Translate spoken English into spoken Polish in real time M

Converse successfully with another person for an hour N

Perform a complex surgical operation M

Unload any dishwasher and put everything away N



The future

it took 40 years to play chess better than humans.

RoboCup and similar initiatives take this challenge to

mechatronics area

Military-driven research: grand challenges � push for more

autonomy

language: Loebner prize (Turing test of today)

application areas that drive development: war, intelligence,

demining, medicine, rehabilitation, services, social interaction,

sex

neural nets get close to human brain (counting neurons)

SINGULARITY

ethics (machine ethics vs. ethics of its creator)

creativity, perception

simulation, hypothetical reasoning



Rodney Brooks

Roughly every �fty years humanity solves a great mystery of

science. We have a chance to solve such a mystery now:

How does the human mind work?

There are many corollary questions:



Corollary questions:

Where does the mind reside?

What is the nature of memory?

What is the role of emotions?

What sort of representations does the brain use?

What does our visual system compute?

How did the evolution shape us?

How do we learn?

What is consciousness?

...



Keys to intelligence

self-adapting perceptual systems, motor systems and

language-related modules

(in contrast to reasoning, planning and knowledge representation)



Outline

1 Introduction









10 Service robots



robotics and artificial intelligence, or artificial...

Documents