Intelligent robotics

224 Pages·1989·11.377 MB·English

by  Mark H. Lee (auth.)

Intelligent robotics
Open University Press Robotics Series
Edited by
P'G. Davey CBE MA MIEE MBCS C.Eng
This series is designed to give undergraduate, graduate and practising
engineers access to this fast developing field and provide an understanding
of the essentials both of robot design and of the implementation of
complete robot systems for CIM and FMS. Individual titles are oriented
either towards industrial practice and current experience or towards those
areas where research is actively advancing to bring new robot systems and
capabilities into production.
The design and overall editorship of the series are due to Peter Davey,
Managing Director of Meta Machines Limited, Abingdon; Fellow of St
Cross College, Oxford University; and formerly Co-ordinator of the UK
Science and Engineering Research Council's Programme in Industrial
Robotics.
His wide ranging responsibilities and international involvement in robotics
research and development endow the series with unusual quality and
authority.
TITLES IN THE SERIES
Industrial Robot Applications E. Appleton and D.J. Williams
Robotics: An Introduction D. McCloy and M. Harris
Intelligent Robotics M.H.Lee
Printed Circuit Board Assembly P.J.W. Noble
Robots in Assembly A. Redford and E. Lo
Robot Sensors and Transducers R. Ruocco
Intelligent robotics
MarkH. Lee
HALSTED PRESS
John Wiley & Sons .
New York - Toronto
and
OPEN UNIVERSITY PRESS
Milton Keynes
To all those who have been neglected
Open University Press
12 Cofferidge Close
Stony Stratford
Milton Keynes MKlllBY
First Published 1989
Copyright © Mark H. Lee 1989
Sof'tcover reprint of the hardcover 1st edition 1989
All rights reserved. No part of this publication may be
reproduced, stored in a retrieval system or transmitted in
any form or by any means, without written permission from the
publisher.
British Library Cataloguing in Publication Data
Lee,MarkH.
Intelligent Robotics.
1. Robots. Development of applications of
artificial intelligence
I. Title
629.8'92
ISBN-13: 978-1-4684-6239-5 e-ISBN-13: 978-1-4684-6237-1
DOl: 10.1007/978-1-4684-6237-1
Published in the USA, Canada and Latin America by
Halsted Press, a division of John Wiley & Sons, Inc., New York
Library of Congress Cataloging-in-Publication Data
Lee,MarkH.
Intelligent Robotics.
(Open University Press robotics series)
1. Robotics. 2. Artificial intelligence. I. Title.
II. Series
TJ211.IA4 1989 629'8'92 88-30417
ISBN-13: 978-1-4684-6239-5
Contents
Series editor's preface ix
Preface xi
Chapter 1 Setting the scene 1
1.1 Aiming for realistic goals 1
1.2 Artificial intelligence 2
1.3 Industrial emphasis 4
1.4 Robots: what are they? 5
1.5 Sensing, thinking and acting 7
1.6 A future scenario 8
1.7 Summary 10
1.8 Further reading material 12
Chapter 2 Sensing the world 13
2.1 Sensor selection 13
2.2 Sensor deployment and integration 20
2.3 Control and coordination 22
2.4 The importance of constraints 24
2.5 Summary 25
2.6 Further reading material 26
Chapter 3 Artificial sight 27
3.1 A salient characteristic 27
vi Intellige~t robotics
3.2 Computer vision applications 28
3.3 Industrial requirements 29
3.4 Research fields 29
3.5 Pattern recognition 31
3.6 Basic techniques -local operators and segmentation 35
3.7 Industrial vision systems 45
3.8 Future developments 50
3.9 Summary 51
3.10 Further reading material 52
Chapter 4 The problem of perception 53
4.1 Perception involves interpretation 53
4.2 The analysis of three-dimensional scenes 58
4.3 Blocks worlds 61
4.4 Current research directions 63
4.5 The problem of understanding 65
4.6 Summary 65
4.7 Further reading material 65
Chapter 5 Building a knowledge base 67
5.1 Introduction 67
5.2 Knowledge representation schemes 69
5.3 Representation review 89
5.4 Knowledge integrity properties 91
5.5 Organization and control 95
5.6 Summary 96
5.7 Further reading material 97
Chapter 6 Machinery for thinking about actions 98
6.1 Introduction 98
6.2 Searching for solutions 99
6.3 Goal directed planning 118
6.4 Rule based planning 123
6.5 Blackboard systems 126
6.6 Summary 128
6.7 Further reading material 129
Chapter 7 Speech and language: from mouse to man 131
7.1 The nature of the problem 132
7.2 Speech processing 135
Contents vii
7.3 Text and language analysis 141
7.4 Robotics and factory systems 142
7.5 Summary 147
7.6 Further reading material 147
Chapter 8 Emulating the expert 148
8.1 The basic expert 148
8.2 A few difficulties and the need for research 152
8.3 Expert systems in industrial automation 155
8.4 The generate-and-test approach 162
8.5 Towards deeper levels of understanding 164
8.6 New techniques for mechanical systems 169
8.7 Summary 174
8.8 Further reading material 174
Chapter 9 Errors, failures and disasters 175
9.1 The importance of automatic error diagnosis and recovery 175
9.2 Classes of errors 176
9.3 Observed behaviour and internal states 177
9.4 Failures in the assembly world 179
9.5 Coping with errors 182
9.6 Building a world model 190
9.7 Summary 194
Chapter 10 Better by design 195
10.1 A proposed assembly system 195
10.2 Summary 199
10.2 Further reading material 200
Chapter 11 Towards a science of physical manipulation 201
11.1 Introduction 201
11.2 The mathematical world 202
11.3 The industrial world 202
11.4 The human world 203
11.5 A case study 204
11.6 Feasibility and maturity 206
Index 208
Series editor's preface
An industrial robot routinely carrying out an assembly or welding task is an impressive
sight. More important, when operated within its design conditions it is a reliable production
machine which - depending on the manufacturing process being automated - is relatively
quick to bring into operation and can often repay its capital cost within a year or two. Yet
first impressions can be deceptive: if the workpieces deviate somewhat in size or position,
or, worse; if a gripper slips or a feeder jams the whole system may halt and look very
unimpressive indeed. This is mainly because the sum total of the system's knowledge is
simply a list of a few variables describing a sequence of positions in space; the means of
moving from one to the next; how to react to a few input signals; and how to give a few
output commands to associated machines.
The acquisition, orderly retention and effective use of knowledge are the crucial missing
techniques whose inclusion over the coming years will transform today's industrial robot
into a truly robotic system embodying the 'intelligent connection of perception to action'.
The use of computers to implement these techniques is the domain of Artificial Intelligence
(AI) (machine intelligence). Evidently, it is an essential ingredient in the future
development of robotics; yet the relationship between AI practitioners and robotics
engineers has been an uneasy one ever since the two disciplines were born.
To make his models at all tractable, the AI scientist has so far had to use so many
simplifying assumptions that the industrial robot engineer tends to scorn such work as naive
or irrelevant. On the other hand, to make a cost-effective production system the engineer is
often forced to tailor it so specifically to the detailed requirements of his production process
that the scientist tends to ignore it as totally ad-hoc, unworthy of his attention because
seemingly showing no generic attributes whatever.
This book occupies a key position in the architecture of our Open University series
because the author is an engineer whose own research field - applying knowledge-based
systems to industrial automation - lies exactly in between these two extreme viewpoints.
x Intelligent robotics
He shows how essential it is for the viability of AI methods in real-world situations that
constraints be applied upon searching, perception, and reasoning by using the maximum
possible knowledge about the process being automated; yet also how vital it is for the
engineer to ensure that he uses computing techniques which are sufficiently generic to have
a 'transfer value' into different robot applications, as well as into harder situations that may
unexpectedly crop up in the same task.
The author describes important concepts of sensor processing, perception, knowledge
bases, planning and expert systems at a non-trivial level-yet with the use of a minimum
of jargon - all of these being seen as crucial to future robot systems in manufacture. Other
major topics in AI such as the understanding and synthesis of natural speech are carefully
described even though they are not yet seen as important for robotics .
The book uses examples from industrial-robot tasks, but also shows clearly how the same
AI techniques will benefit the entire process of Computer Integrated Manufacture extending
from product design through process planning to automatic manufacture, and beyond that to
the harder problems in applying so-called 'advanced' and mobile robots to less structured
environments outside the factory.
It will, I hope, come to be seen both as good AI and good robotics, and so do much to
foster the mutual support that each discipline can give the other.
P.G.Davey