CONTRIBUTORS TO THIS VOLUME TAEH.CHO I J CONN ELL NDYN EKERE P M FINNIGAN PAUL M.FRANK JOSEPH C GIARRATANO ROGER G HANNAM A F HATHAWAY W E LORENSEN V N PARTHASARATHY J.B.ROSS JERZY W ROZENBUT RALF SELIGER FREDERICKE SISTLER DAVID M.SKAPURA STEUOS C A THOMOPOULOS IAN WHITE BERNARD P ZIEGLER CONTROL AND DYNAMIC SYSTEMS ADVANCES IN THEORY AND APPLICATIONS Edited by C T LEONDES School of Engineering and Applied Science University of California, Los Angeles Los Angeles, California and College of Engineering University of Washington Seattle, Washington V O L U M E : MANUFACTURING AND AUTOMATION SYSTEMS: TECHNIQUES AND TECHNOLOGIES Part of ® ACADEMIC PRESS, INC Harcourt Brace Jovanovich, Publishers San Diego New York Boston London Sydney Tokyo Toronto ACADEMIC PRESS RAPID MANUSCRIPT REPRODUCTION This book is printed on acid-free paper @ Copyright © 1991 by ACADEMIC PRESS, INC All Rights Reserved No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher Academic Press, Inc San Diego, California 92101 United Kingdom Edition published by Academic Press Limited 24-28 Oval Road, London NW1 7DX Library of Congress Catalog Number: 64-8027 International Standard Book Number: 0-12-012749-0 PRINTED IN THE UNITED STATES OF AMERICA 91 92 93 94 CONTRIBUTORS Numbers in parentheses indicate the pages on which the authors' contributions begin Tae H Cho (191), AI Simulation Group, Department of Electrical and Computer Engineering, The University of Arizona, Tucson, Αήζοηα 85721 I J Connell (289), GE Corporate Research and Development Center, Schenectady, New York 12301 Ndy N Ekere (129), University ofSalford, Salford, United Kingdom P M Finnigan (289), GE Corporate Research and Development Schenectady, New York 12301 Center, Paul M Frank (241), Department of Measurement and Control, University of Duisburg, Ψ-4100 Duisburg 1, Germany Joseph C Giarratano (37), University of Houston—Clear Lake, Houston, Texas 77508 Roger G Hannam (129), University of Manchester, Institute of Science and Technology (UMIST), Manchester M601QD, United Kingdom A F Hathaway (289), GE Corporate Research and Development Schenectady, New York 12301 Center, W E Lorensen (289), GE Corporate Research and Development Schenectady, New York 12301 Center, V N Parthasarathy (289), GE Corporate Research and Development Center, Schenectady, New York 12301 J B Ross (289), GE Aircraft Engines, Cincinnati, Ohio 45215 vii viii CONTRIBUTORS Jerzy W Rozenblit (191), AI Simulation Group, Department of Electrical and Computer Engineering, The University of Arizona, Tucson, Arizona 85721 Ralf Seliger (241), Department of Measurement and Control, University of Duisburg, W-4100 Duisburg 1, Germany Frederick E Sistler (99), Department of Agricultural Engineering, Louisiana State University Agricultural Center, Baton Rouge, Louisiana 70803 David M Skapura (37), Loral Space Information Systems, Houston, Texas 77058 Stelios C A Thomopoulos (339), Decision and Control Systems Laboratory, Department ofElectncal and Computer Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802 Ian White (1), Defence Research Agency, Portsdown, Cosham P06 4AA, England Bernard P Ziegler (191), AI Simulation Group, Department of Electrical and Computer Engineering, The University of Arizona, Tucson, Arizona 85721 PREFACE At the start of this century, national economies on the international scene were, to a large extent, agriculturally based This was, perhaps, the dominant reason for the protraction, on the international scene, of the Great Depression, which began with the Wall Street stock market crash of October 1929 In any event, after World War II the trend away from agriculturally based economies and toward industrially based economies continued and strengthened Indeed, today, in the United States, approximately only 1% of the population is involved in the agriculture industry Yet, this small segment largely provides for the agriculture requirements of the United States and, in fact, provides significant agriculture exports This, of course, is made possible by the greatly improved techniques and technologies utilized in the agriculture industry The trend toward industrially based economies after World War II was, in turn, followed by a trend toward service-based economies; and, in fact, in the United States today roughly 70% of the employment is involved with service industries, and this percentage continues to increase Nevertheless, of course, manufacturing retains its historic importance in the economy of the United States and in other economies, and in the United States the manufacturing industries account for the lion's share of exports and imports Just as in the case of the agriculture industries, more is continually expected from a constantly shrinking percentage of the population Also, just as in the case of the agriculture industries, this can only be possible through the utilization of constantly improving techniques and technologies in the manufacturing industries As a result, this is a particularly appropriate time to treat the issue of manufacturing and automation systems in this international series Thus, this is Part of a five-part set of volumes devoted to the most timely theme of "Manufacturing and Automation Systems: Techniques and Technologies." The first contribution to this volume is "Fundamental Limits in the Theory of Machines," by Ian White This contribution reviews some of the fundamental limits of machines that constrain the range of tasks that these machines can be made to undertake These include limitations on the computational process, limitations in physics, and limitations in the ability of their builders to define the ix x PREFACE processes to be undertaken This contribution seeks to relate several developments in the theory of computer science, physics, and philosophy to the question of what machines can and cannot The question is appallingly hard, so the reader should seek in this contribution some insights rather than total enlightenment Nonetheless, the question is more than a philosophical one The relationship of machines to intelligent functioning is a central question in machine theory and is one that brings philosophy into the mainstream of what is far more empirical science than is generally acknowledged As more and more is expected of increasingly capable manufacturing systems, the many fundamental issues raised in this contribution need to be recognized and taken into account The next contribution is "Neural Network Techniques in Manufacturing and Automation Systems," by Joseph C Giarratano and David M Skapura A neural net is typically composed of many simple processing elements arranged in a massively interconnected parallel network Depending on the neural net design, the artificial neurons may be sparsely, moderately, or fully interconnected with other neurons Two common characteristics of many popular neural net designs are that (1) nets are trained to produce a specified output when a specified input is presented rather than being explicitly programmed and (2) their massive parallelism makes nets very fault tolerant if part of the net becomes destroyed or damaged This contribution shows that neural networks have a growing place in industry by providing solutions to difficult and intractable problems in automation and robotics This growth will increase now that commercial neural net chips have been introduced by vendors such as Intel Corporation Neural net chips will find many applications in embedded systems so that the technology will spread outside the factory Already, neural networks have been employed to solve problems related to assembly-line resource scheduling, automotive diagnostics, paint quality assessment, and analysis of seismic imaging data These applications represent only the beginning As neural network technology flourishes, many more successful applications will be developed While not all of them will utilize a neural network to solve a previously intractable problem, many of them will provide solutions to problems for which a conventional algorithmic approach is not cost-effective Based on the success of these applications, one looks forward to the development of future applications The next contribution is "Techniques for Automation Systems in the Agriculture Industry," by Frederick E Sistler The agriculture industry encompasses the growth, distribution, and processing of food and fiber, along with related suppliers of goods and services This contribution presents techniques and control systems used in on-farm agriculture It is applications-oriented rather than mathematically oriented because the primary contribution is seen to be in the unique applications of existing sensors, systems, and techniques to biological systems The properties and behavior of plants and animals vary greatly both among and within species The response of a biological system is greatly dependent upon its PREFACE XI environment (moisture, temperature, relative humidity, soil, solar radiation, etc.), which itself can be highly variable and difficult to model All of this makes biological systems more difficult to model than inorganic systems and materials Automation is used in agriculture for machine control, environmental (building) control, water management, sorting and grading, and food processing Farming has traditionally been associated with a very low level of automation However, as noted at the beginning of this preface, more and more is expected of a diminishing percentage of the population, which can only be achieved through constantly improving automation techniques and technologies such as are presented in this contribution The next contribution is "Modeling and Simulation of Manufacturing Systems," by Ndy N Ekere and Roger G Hannam A manufacturing system generally includes many linked processes, the machines to carry out those processes, handling equipment, control equipment, and various types of personnel A manufacturing system for an automobile could include all the presslines to produce the body panels; the foundries to produce the engine blocks and transmission housing; forge shops to produce highly stressed parts such as suspension components and crankshafts; the machine shops that convert the forgings, castings, and other raw material to accurately sized components; and the subassembly and final assembly lines that result in the final product being produced Many writers call each of these subsections a manufacturing system, although each is also a constituent of a larger manufacturing system The machines and processes involved in manufacturing systems for mass production are dedicated to repetitive manufacture The majority of products are, however, produced by batch manufacturing in which many different parts and products are produced on the same machines and the machines and processes are reset at intervals to start producing a different part The techniques presented in this contribution apply to manufacturing systems that extend from a few machines (that are related—generally because they are involved in processing the same components) up to systems that might comprise the machines in a complete machine shop or complete processing line The characteristics of batch manufacturing are often analyzed by simulation; mass production systems are analyzed more by mathematical analysis This contribution is an in-depth treatment of these issues of modeling and simulation that are of major importance to manufacturing systems The next contribution is "Knowledge-Based Simulation Environment Techniques: A Manufacturing System Example," by Tae H Cho, Jerzy W Rozenblit, and Bernard P Zeigler The need for interdisciplinary research in artificial intelligence (AI) and simulation has been recognized recently by a number of researchers In the last several years there has been an increasing volume of research that attempts to apply AI principles to simulation This contribution describes a methodology for building rule-based expert systems to aid in discrete event simulation (DEVS) It also shows how expert systems can be used in the XU PREFACE design and simulation of manufacturing systems This contribution also presents an approach to embedding expert systems within an object-oriented simulation environment, under the basic idea of creating classes of expert system models that can be interfaced with other model classes An expert system shell for the simulation environment (ESSSE) is developed and implemented in DEVSscheme knowledge-based design and simulation environment (KBDSE), which combines artificial intelligence, system theory, and modeling formalism concepts The application of ES models to flexible manufacturing systems (FMS) modeling is presented The next contribution is "Fault Detection and Isolation in Automatic Processes," by Paul M Frank and Ralf Seliger The tremendous and continuing progress in computer technology makes the control of increasingly complex manufacturing and automation systems readily possible Of course, the issues of reliability, operating safety, and environmental protection are of major importance, especially if potentially dangerous equipment like chemical reactors, nuclear power plants, or aircraft are concerned In order to improve the safety of automatic processes, they must be supervised such that occurring failures or faults can be accommodated as quickly as possible Failures or faults are malfunctions hampering or disturbing the normal operation of an automatic process, thus causing an unacceptable deterioration of the performance of the system or even leading to dangerous situations They can be classified as component faults (CF), instrument faults (IF), and actuator faults (AF) The first two steps toward a failure accommodation are the detection and the isolation of the fault in the system under supervision The term detection denotes in this context the knowledge of the time at which a fault has occurred, while isolation means the determination of the fault location in the supervised system (i.e., the answer to the question "which instrument, actuator, or component failed?") This contribution is an indepth treatment of this issue of fault detection and isolation and the role it can play in achieving reliable manufacturing and automation systems The next contribution is "CATFEM—Computer Assisted Tomography and Finite Element Modeling," by P M Finnigan, A F Hathaway, W E Lorensen, I J Connell, V N Parthasarathy, and J B Ross Historically, x-ray computed tomography (CT) has been used for visual inspection of cross-sectional data of an object It has been successfully applied in the medical field as a noninvasive diagnostic tool and in industrial applications for quality evaluation This contribution presents a conventional look at CT and, in addition, details revolutionary approaches to the use of computed tomography data for engineering applications, with emphasis on visualization, geometric modeling, finite element modeling, reverse engineering, and adaptive analysis The concept of a discrete solid model, known as a digital replica TM, is introduced The digital replica possesses many of the same attributes intrinsic to a conventional CAD solid model, and thus it has the potential for broad applicability to many geometry-based ap- PREFACE Xlll plications, including those that are characteristic of steps that are involved in many manufacturing processes This contribution discusses three-dimensional imaging techniques for the CT slice ensemble using surface reconstruction Such capability provides the user with a way to view and interact with the model Other applications include the automatic and direct conversion of x-ray computed tomography data into finite element models The notion of reverse engineering a part is also presented; it is the ability to transform a digital replica into a conventional solid model Other technologies that support analysis along with a system architecture are also described This contribution provides sufficient background on CT to ease the understanding of the applications that build on this technology; however, the principal focus is on the applications themselves The final contribution to this volume is "Decision and Evidence Fusion in Sensor Integration," by Stelios C A Thomopoulos Manufacturing and automation systems will, in general, involve a number of sensors whose sensed information can, with advantage, be integrated in a process referred to as sensor fusion Sensor integration (or sensor fusion) may be defined as the process of integrating raw and processed data into some form of meaningful inference that can be used intelligently to improve the performance of a system, measured in any convenient and quantifiable way, beyond the level that any one of the components of the system separately or any subset of the system components partially combined could achieve This contribution presents a taxonomy for sensor fusion that involves three distinct levels at which information from different sensors can be integrated; it also provides effective algorithms for processing this integrated information This volume concludes this rather comprehensive five-volume treatment of techniques and technologies in manufacturing and automation systems The authors of this volume and the preceding four volumes are all to be commended for their splended contributions, which will provide a uniquely significant reference source for workers on the international scene for years to come 410 STELIOS C A THOMOPOULOS Let (u, = u for some u ε S w , i Φ k, i ε S , and u = p, p ε S_.) i i i M N k r r M be in the set N Due to Lemma 1, the right hand side of (a 21) equals M Σ P(kj,l) The existence of an optimum quantization implies that there J=P exist probabilities P(i,2,0), , P(i,M,0), for every i ε S , such that the false alarm requirement (a 10) is satisfied Hence, if there exists a quantization which attains the largest possible values for the M probabilities { Σ P(kj,l), p = 2, M, k ε S }, consistent with the J=P false alarm requirement a then such a quantization is optimum Notice that the LRQ satisfies Lemma (see Lemma 2) We complete the proof of the theorem by showing that the LRQ in (a 13) achieves the largest possible values for the probabilities v· * M { Σ P(kj,l), p = J=P M, k ε Λ Consider the LRQ T and any other quantization T such that r ε C implies u = i forT , Λ r ε C implies u = i for T , Pr(r ε C I H ) = P (i,i,m), i t m Pr(r ε C I H ) = p A (U,m), i i m and P (i,1.0) = P (U,0) (a.22) DECISION AND EVIDENCE FUSION IN SENSOR INTEGRATION 411 * In the above expressions, i ε S,„ i ε S^T, and C M N Λ (C ) are i t mutually exclusive and collectively exhaustive subsets of R for every I, i.e they form a coverage of R Denote the likelihood function dP (r.) under H as L , the integral m m of two sets S and S dP (r J as J ^ m i J R V as S S R m i L , the intersection m , and the compliment of a set S as S Consider the difference P*(i,M,l) -p A (i,M,l)= J L - J L (a.23) Λ C M Upon adding and subtracting ι * C °M L to the right hand side of (a.23), Λ MCM P*(i,M,l)-pA(i,M,l)= J L - J * —Λ C 1^ (a.24) Λ —* MCM C MCM For the M-th quantization level threshold at the 1-th sensor, L L l l λΛ^Μ * — < oo holds in Cmir, and — < λ, w holds in C T Hence, by making i,M M i.M L L L o o use of the (a 16), the right hand side of (a.24) is bounded from below by W ί Lo - ί L o ] · * —Λ C MCM Λ —* C MCM Upon adding and subtracting j C L * to the above bound, we obtain X MCM 412 STELIOS C A THOMOPOULOS Ρ*(1,Μ.1)-ΡΑ(1,Μ,1) ^ λ Μ ( j ' LQ - j c * L0liO c M (a.25) Λ M where the last inequality follows from the requirement that P (i.M.O) = Λ P (i.M.O) Along similar lines, we could show that the following relations are true: M „ M A Σ P(i.j.l)-Z P(i.j.l)= J=P J=P f J L M u c f * k= p k for p = 2, - J M L ;> (a.26) A u c k= p k M-l where "U" stands for set union in the above Furthermore, the inequalities in (a.25) and (a.26) are satisfied for eveiy i ε S This completes the proof of Theorem ° INDEX A Absolute Absolute limits, limits, defined, defined, 2-3 2-3 Acoustic Acoustic signals, signals, seismic seismic imaging, imaging, 84-88 84-88 Activities Activities in in discrete discrete simulation simulation modeling, modeling, 139 139 world views, 141 Activity Activity cycle cycle diagrams diagrams (ACD), (ACD), simulation simulation modeling modeling components diagrams, 157-158 157-158 Extended Extended Control and Simulation Simulation Language Language (ECSL), (ECSL), 150-151,155-159 150-151,155-159 human human operators, operators, 159 159 pallet diagrams, pallet diagrams, 158-159 158-159 Actuator faults (AF), 241 Actuator 241 Adaline neural 46 Adaline neural net net model, model, 46 Adaptive analysis, analysis, CATFEM CATFEM modeling, modeling, 331-333 331-333 Adaptive Adaptive control Adaptive control neural 54-56 neural nets, nets, 54-56 robotics applications, applications, 61 61 robotics Adaptive critic critic methods, methods, neurocontroller neurocontroller Adaptive architectures, 53 53 architectures, Adaptive Resonance Resonance Theory Theory (ART) (ART) models models Adaptive neural 47 neural nets, nets, 47 unsupervised learning, 50 50 unsupervised learning, Agriculture industry industry automation automation systems systems Agriculture background, 99-100 background, 99-100 chemical application application automation, automation, 110-112 110-112 chemical orchard orchard spraying, spraying, 111-112 111-112 electrical management, 123-125 electrical energy energy management, 123-125 demand demand control, 123-125 123-125 environmental environmental control control automation, automation, 121-123 121-123 potato storage, potato storage, 121-122 121-122 poultry housing, housing, 122-123 poultry 122-123 grading grading and and sorting sorting automation, automation, 115-119 115-119 peach grading, peach grading, 115-117 115-117 small small fruit fruit sorting, sorting, 117-118 117-118 tomato tomato grading, grading, 118-119 118-119 greenhouse and nursery operations, 119-121 container container handling, 119-120 119-120 greenhouse greenhouse control, control, 120-121 120-121 irrigation irrigation and and drainage drainage automation, automation, 112-115 112-115 center-pivot center-pivot and and surface surface irrigation, irrigation, 113-114 113-114 lysimeter lysimeter control, control, 112-113 112-113 solid-set solid-set irrigation irrigation control, control, 114-115 114-115 tractor tractor operations operations automation, automation, 100-110 100-110 apple harvester harvester guidance, guidance, 102-103 102-103 apple ballast position control, 106-107 106-107 ballast position control, combine control, control, 103-105 103-105 combine guidance systems, systems, 100-102 100-102 guidance location sensing, sensing, 105-106 105-106 location planter depth control, control, 107-110 107-110 planter depth Algorithmic information, information, communication communication and, and, 29 29 Algorithmic Algorithms Algorithms fault fault detection detection and and isolation, isolation, 272-281 272-281 neural nets, nets, 39 39 neural paradigm of interpreted interpreted mathematics, mathematics, 13-15 13-15 paradigm of Analogical reasoning, reasoning, limits limits of of logic logic and, and, 24 24 Analogical Analytical knowledge, knowledge, knowledge-based knowledge-based fault fault Analytical detection and and isolation, isolation, 271-272 271-272 detection Analytical redundancy redundancy Analytical fault fault detection detection and and isolation isolation models, models, 242-243 242-243 basic concepts, 247-248 247-248 basic concepts, robustness, 245 robustness, 245 Apple harvester, harvester, guidance guidance systems systems for, for, 102-103 102-103 Apple ART-ROSS system, system, simulation simulation systems systems modeling, modeling, ART-ROSS 193 193 Artificial Artificial intelligence intelligence (AI) (AI) fundamental fundamental limits limits and, and, 32-33 32-33 perceptron research, perceptron research, 44-49 44-49 simulation simulation systems systems modeling, modeling, 191-193 191-193 Neural nets Artificial Artificial network network systems systems (ANS) (ANS) See See Neural nets Associative recall, Associative recall, neural neural nets, nets, 41 41 Atomic-expert model model Atomic-expert DESE-KBDSE DESE-KBDSE interface, interface, 204-208 204-208 KBDSE techniques, techniques, 196-199 196-199 KBDSE type router version, version, 228, 236-239 type router 228, 236-239 413 413 INDEX 414 Attributes, in discrete simulation modeling, 138-141 Automated analysis, CATFEM modeling, 333-334 Automated test facility (ATF) architecture for, 214 system entity structure, 214-215 Automated Test Facility (ATF) model, 209-213 distributed architecture, 227-228 simulation results, 224-226 Automatic guidance systems apple harvester, 102-103 tractor operations, 100-102 Automotive systems, robotic neural nets, - Autotuned controllers, neural nets, 54 B Backpropagation Bayesian/N-P DDF, 392-398 neural net models, - paint quality assurance, 6 - Backpropagation-through-time, neurocontroller architectures, 52 Balancing, mathematical models for, 183-184 Ballast position control, agricultural automation, 106-107 BAM neural nets, unsupervised learning, 50 Batch production, inventory with, 132 Bayesian distributed decision fusion (DDF) binary hypothesis testing multi-level local logic, 353-375 single-level local logic, - Dempster-Shafer's (D-S) theory compared with, 375-388 Neyman-Pearson DDF combined with, 388-401 optimal linear configuration, 389-390 Bayesian theory characteristics, 23 distributed decision fusion (DDF), 342-343 generalized evidence processing (GEP), 354-375 Beam hardening, CT imaging, 296-297 Bellman equation, adaptive critic methods, 53 Best-fit linear models, CATFEM modeling, 316-317 Blackboard data structure, expert systems (ES) modeling, 229 Boundary representation (B-rep), CATFEM techniques, 300-301 Brain-State-in-a-Box neural net model, 47 CAN-Q (Computer Analysis of Network of Queues), in systems modeling, 181-184 CAPS (Computer-aided programming of simulation), 151-152 Carrier sense, ATF design model, 212-213 Carrier sense medium access (CSMA), ATF design model, 212-213 Categories, set theory, 17 CATFEM (computer-assisted tomography and finite element modeling) background, 289 digital replica, 302-308 geometric operators, 306-308 normal surface calculation, 305 terminology, 303-305 FEM/FEA process and, 319-335 adaptive analysis, 331-333 automated analysis integration, 333-334 automatic mesh generation, 319-323 geometry and, 323-324 integration of, 324-331 CATFEM-2D, - CATFEM-3D, 326-331 geometry, 0 - reverse engineering, 312-318 CATRE-2D, 314-317 CATRE-3D, 317-318 visualization, 308-312 examples, 311-312 oblique reformating, 311 volume modeling, 309-311 volume rendering, 308-309 opaque, 312 translucent, 312 CATRE (Computer-Assisted Tomography and Reverse Engineering), modeling with, 313-318 three-dimensional techniques, 317-318 two-dimensional techniques, 314-317 Center-pivot irrigation systems, automated design for, 113-114 Cerebellar model articulation controller, 60-61 Chemical application, automation of, in agricultural systems, 110-112 Cholesky factorization, identification-based FDI hypothesis testing, 267 Circumscription, characteristics, 22 Classic inverted pendulum, neural net applications, 58 Classical logic (CL) common sense reasoning (CSR) and, 21 defined, 20 INDEX Code generators, characteristics of, 142-143 Collision detection, ATF design model, 212-213 Color analysis automation, peach grading with, 116 Combine control, guidance systems for, 103-105 Common Sense Reasoning (CSR) limits of logic and, - logic and, - Communication denned, 2 - information, - inter-element, 10-11 ISO OSI seven layer model, - Complexity limits, defined, - Component-faults, 241 Compositing, CATFEM modeling visualization, 309 Computed tomography (CT) beam hardening, 296 detector rings, 296 digital radiograph (DR) and image generation, 293-294 edge detection methods, - 9 edge extension grooves, - elements of, 291-300 history of, 291-293 image blurring, - industrial applications, 290 lack of penetration, 298 limits of, 9 - 0 smearing, 296-297 three-dimensional image generation, - Computer-aided design (CAD), computed tomography and, 290 Computer-aided engineering (CAE), computed tomography and, 290 Connectionism neural nets, 37 rational machines, 32 Constructive Solid Geometry (CSG), CATFEM techniques, 300-301 Container handling, automation of, 119-120 Continuous systems, simulation modeling of, 136 Control issues, expert systems (ES) modeling, 229-230 Cost factors, generalized evidence processing (GEP) theory, 371-374 Coupled models DESE-KBDSE interface, 206-207 KBDSE techniques, 196-199 Credit Assignment problem, neural nets applications, 58 models, - Curve, defined, for CATFEM techniques, 301 415 D Dead-beat observers, robust fault detection and isolation models, 253 Decision rules generalized evidence processing (GEP), 357-375 combining rule, - single and multi-level logic, DDF, 341-342 Decorrelation techniques, fault detection and isolation, 269 Dedicated observer schemes, fault detection and isolation models, 249 Definite decision regions, generalized evidence processing (GEP), decision rules, 360-361 Delaunay triangulation automatic mesh generation, 321-322 three-dimensional CATFEM modeling, 328-331 Dempster-Shafer's (D-S) theory Bayesian DDF, statistically independent sensors, 350-351 combining rule, - 8 distributed decision fusion (DDF), - , 375-388 Depth sensing device, automated plant depth control with, 108-110 Detection concealed faults, robust fault detection and isolation models, - Detector rings, CT imaging, 296-297 DetermineNormal(x,y,z) operator, CATFEM modeling, 307 DetermineUniverse operator, CATFEM modeling, 306 Deterministic systems, simulation modeling of, 137 Diagnostic models, fault detection and isolation, 243 Digital radiography (DR), CT image generation and, - Digital replica CATFEM modeling, - geometric operators, 306-308 inside/outside, 304-305 interpolation, 305 normal surface calculations, 305 terminology, 303-305 three-dimensional CATFEM, 326-331 two-dimensional CATFEM, 325-326 computed tomography with, 291 Digital Replica Geometric Modeling Utility (DRGMU), - Direct inverse control, neurocontroller architectures, 51 INDEX 416 Discrete event-system (DEVS) formalism, KBDSE techniques, 195-199 Discrete systems simulation modeling, 136 elements of, 138-141 languages and packages, 142-147 SLAM (Simulation Language for Alternative Modeling), 170-171 Discrete time models, fault detection and isolation models, 245 Distributed Decision (Evidence) Fusion (DD(E)F), 340-343 Distributed decision fusion (DDF) Dempster-Shafer's (D-S) theory, - 8 sensor integration Bayesian DDF, binary hypothesis testing, 345-353 evidence processing and, 3 - Distributed Expert System Environment (DESE) KBDSE interface with, - Object-oriented programming (OOP), 0 - Distributed processing, expert systems (ES) modeling, 2 - 2 Disturbance-modeling, fault detection and isolation models, - Dividing cubes algorithm, CATFEM modeling visualization, 310-311 Dynamic models, fault detection and isolation, 242 E Edge, defined, CATFEM techniques, 302 Edge detection techniques, CT imaging, - 9 Edge extension grooves, CT imaging blurring, 297-298 Edge profiles CT image blurring, - smearing, 296-297 Efficiency, manufacturing systems, 131-132 Einstein-Podolski-Rosen (EPR) effects, - Electrical energy management automation of, 123-125 demand control, 123-125 Element generation, automatic mesh generation, 321-322 Embedding simulation, expert systems, 194-195 Energy less computers, reversible and irreversible limits, - Engine diagnostics block diagram, - neural nets in robotics, - Entities, in discrete simulation modeling, 139 Entity structure base (ENBASE), KBDSE techniques, 196 Environmental control, automation, 121-123 in greenhouses, 120-121 potato storage, 121-122 poultry housing, 122-123 Epoch numbers, Bayesian/N-P DDF, - Evapotranspiration (ET), lysimeter control systems for, 112-113 Event modeling discrete simulation modeling, 139-140 world views, 141 SLAM (Simulation Language for Alternative Modeling), 170-171 specific routines, 179-180 Experimental frame conditions, expert-system (ES) modeling, 221-222 Expert system (ES) models controllers, neural nets, - flexible manufacturing systems (FMS) modeling, 213-231 frame conditions, 221-222 interruptibility of, - material flow models, 217-219 neural nets, 40 Expert system shell for the simulation environment (ESSSE), - Expert-core models, interruptibility of, - Extended Control and Simulation Language (ECSL), simulation modeling activities section, 162-165 activity cycle diagrams (ACD), 150-151, 155-159 AFIX and BFIX activities, 163-164 case study, 149-168 data section, 165-166 definition statements, 159-161 DYNAMICS sector statements, 162 finalization section, 165 flow and control, 152-153 initialization routine, 161-162 library functions, 161 procedure and results, 166-168 RECORDING Sector statements, 162-163 RTLOAD activity, 164-165 Extended Kaiman filter (EKF), Bayesian/N-P DDF training, 400-401 External agent (EA), communication, - F Face, defined, CATFEM techniques, 302 Failures defined, 241 See also Fault detection INDEX Fault detection and isolation (FDI) applications, 272-281 robot Manutec R , 274-281 steam generator, 280-281 background, - , - knowledge-based methods, - modeling, - analytical redundancy and, - fault and disturbance-modeling, - fault detection filter, 250 frequency domain methods, - generalized dedicated observer scheme, 249-250 identification-based methods, 264-267 hypotheses testing, 265-267 parameter estimation, 264-265 nonlinear fault detection filter, 260-261 nonlinear systems, observer-based models, 259-264 nonlinear unknown input observers, 261-264 observer eigenstructure assignment, 256-257 parity space approach, - process model description, 244-245 robustness, 250-260 unknown input observers, - observer-based, in nonlinear systems, - residual evaluation, 267-270 decorrelation method, 269 generalized likelihood ratio technique, 267-268 multiple hypotheses testing, 268 threshold selector, 269-271 Fault detection filter fault detection and isolation models, 250 observer-based FDI, nonlinear systems, 260-261 Fault tolerance, neural nets, 41 Feature extraction techniques, peach grading with, 116 Finite element modeling and analysis (FEM/FEA) CATFEM system and, 319-335 adaptive analysis, 331-333 automated analysis, 333-334 automatic mesh generation, 319-320 geometry and, 323-324 integration of, 324-331 smoothing, 322 tree building, 320-321 Finite limit theory, - First arrival (FAR) signals, seismic imaging, 86-88 First break picking, neural nets in robotics, - 8 First order logics, limits of logic and, - FISE (finite integral squared error) criteria, - 417 Flexible manufacturing systems (FMS), neural nets, - Flow-line systems, mathematical models for, 183-184 Forward-expert models, DESE-KBDSE interface, 204-208 Frequency domain methods, robust fault detection and isolation models, - Frobenius theorem, robust FDI, - Fundamental limit theory background, 1-2 classification, 2-12 absolute limits, - complexity limits, - infinite and finite limits, - physics-based machine limitations, 5-12 communication, - models and meaning, 12-16 interpretive process, 15-16 mathematics, 14-15 rational machine building, - sets and logic, 16-25 logic of logics, - Putnam's theorem, 18-20 Fusion error, Bayesian/N-P DDF backpropagation, 398 Fuzzy decision region, generalized evidence processing (GEP) combining rule, 364 decision rules, 361-362 G Gaussian channels Dempster-Shafer's (D-S) theory, 385-388 generalized evidence processing (GEP) theory, 370 signal-to-noise ratio, - Generalization, logic and, 20 Generalized dedicated observer scheme, FDI models, - Generalized delta equation LMS error signal, - neural net models, - Generalized Evidence Processing (GEP) theory Bayesian DDF, 354-375 combining rule, 362-374 multi-level local logic, 354 D-S theory compared, 366-367 decision boundaries, - Generalized likelihood ratio technique, FDI residual evaluation, 267-268 steam generator applications, 281 INDEX 418 Geometrie operators, CATFEM modeling, 306-308 Geometry automatic mesh generation, 323-324 CATFEM techniques, 300-302 Godel's theorem, - mathematics as paradigm for physics, 15 Graceful degradation, in neural nets, 41 Gradient descent, LMS error signal, - Grading and sorting automation agricultural systems, 115-119 peach grading, 115-117 Graphical models, of networks, 184 Graphics in simulation modeling, 144 Greenhouse operations, automation of, 120-121 H H-method of adaptive analysis, CATFEM modeling, 332 Harvesting, automation of, 106-107 Heat dissipation, massively parallel irreversible computers, 11 Hebb, Donald O., 45 Hedges, in set theory, 17 Heuristics, knowledge-based fault detection and isolation, 271-272 Hopfield nets development of, - unsupervised learning, 50 Hypotheses testing identification-based FDI, 265-267 residual evaluation of FDI, 268 binary Bayesian distributed decision fusion (DDF) multi-level local logic, 353-375 single-level local logic, - Dempster-Shafer's (D-S) theory, 378-388 I Idealized Cognitive Models (ICMs), set theory, 17 Identification-based FDI hypotheses testing, 265-267 parameter estimation, - 6 techniques, 264-267 Identity observers, observer-based FDI, 261 IF THEN reasoning, neural nets, 40 Image blurring, CT image generation, 295-296 Image metrology, three-dimensional CT image generation, 294-295 Implication, of logic, 20-21 Indecision regions, generalized evidence processing (GEP), 358-359 Inference engines (IE) Distributed Expert System Environment (DESE), 0 - fault detection and isolation, 271-272 neural nets in robotics, - Infinity axiom of, limit theory, - Information flow models, printed circuit boards (PCBs), 219-224 InOut Test (x,y,z) operator, CATFEM modeling, 307 Instance variables, Distributed Expert System Environment (DESE), 2 - Instinct net, robotics applications, 60 Instrument-faults, 241 Intensity vs thickness plot, CT imaging, 296 Inter-element communication, 10-11 Inter-element spacing, massively parallel irreversible computers, 11 Interpolate^, y,z) operator, CATFEM modeling, 306-307 Interpretation, of logic, - Interpreted mathematics paradigm, fundamental limits, 12-13 IntersectLineWithModel (pi, p2) operator, CATFEM modeling, 308 IntersectOctantEdgeWithModel operator, CATFEM modeling, 307 IntersectRayWithModel (pO, direction) operator, CATFEM modeling, 308 Irreversible computers massively parallel machines, 9-12 physical limits, Irrigation, automated systems for, 112-115 center-pivot and surface irrigation, 113-114 lysimeter control, 112-113 solid-set control, 114-115 ISI system, simulation modeling, 145 K Kaiman filter Bayesian/N-P DDF training, 400-401 fault detection and isolation decorrelation methods, 269 generalized likelihood ratio technique, 267-268 Knowledge-based design and simulation environment (KBDSE) DEVS formalism, 196 INDEX 419 Distributed Expert System Environment (DESE) interface, - manufacturing systems modeling, 195-199 SES formalism, 196-199 Knowledge-based expert systems (KBES), flexible manufacturing systems (FMS), - Knowledge-based systems (KBS) Distributed Expert System Environment (DESE), 0 - environment techniques, 191-195 expert systems (ES) modeling, 229 fault detection and isolation, 4 - , 270-272 modularity in, 20 Kronecker's delta function, generalized evidence processing (GEP) combining rule, - 6 LocateVoxel(x,y,z) operator, CATFEM modeling, 306 Location sensing, irrigation systems, 105-106 Logic of logics, - limits of, - outline of, 21-23 rational machines, 30 set theory, 16 Loop, defined, CATFEM techniques, 302 Loweheim Skolem Theorem, 18-19 L efficiency, modeling and simulation techniques, 131-132 mathematical models, 181-184 modeling and simulation techniques, 129-188 networks and graphical models, 184 Petri net models, 185-188 simulation techniques background, 133-134 discrete simulation model, 138-141 fundamentals, 134-138 continuous and discrete systems, 136 stochastic and deterministic systems, 137-138 languages and packages, 142-147 code generators, 142-143 ECSL case study, 149-168 graphics, 144 SLAM case study, 168-180 workbench, 144-146 world views of, 141-142 Marching cubes algorithm Language, rational machines, 32 Laplace transforms fault detection and isolation models, 246 frequency domain analysis, - Laplacian-of-Gaussian (LoG) operator CATFEM modeling, two-dimensional CATRE, 315-317 CT imaging edge detection, - 9 Laser technology, agricultural automation, 105-106 Leak detection, fault detection and isolation techniques, - Learning theory, neural nets, - Least-mean-square error signal, neural nets, 95-98 Level of Confidence (LOC), generalized evidence processing (GEP) theory, - Levesque's Logic, characteristics, 22 Likelihood Logic, characteristics, 22 Likelihood ratio test (LRT) Bayesian DDF multi-level local logic, - single-level local logic, 345-347 distributed decision fusion (DDF), - monotone property, 408-412 Neyman-Pearson DDF, - monotone property, - Linear process modeling, fault detection and isolation, 272 Linear time-invariant systems, fault detection and isolation models, 246 Linguistic qualifiers, set theory, 16-17 Local area network (LAN), Automated test facility (ATF) design model, 212 M Madaline neural net model, 46 Manufacturing systems CATFEM modeling visualization, 309-311 three-dimensional CATRE techniques, 317-318 Markov theory, fault detection and isolation, 270-271 Massively parallel irreversible computer fundamental limit machine, 11-12 heat dissipation, 11 inter-element communication, 10-11 inter-element spacing, 11 limits, 9-12 power considerations, 10 Massively parallel machines, propagation time limits, Material flow models, expert system (ES) models, 217-219 INDEX 420 Mathematical modeling, manufacturing systems, 181-184 Matrix controlled inference engine (MACIE), 40-41 Maximum intensity projection, CATFEM modeling visualization, 309 McDermott and Doyle logic, characteristics, 21 Mean value analysis (MAVQ), systems modeling, 181 Mean-squared error, Bayesian/N-P DDF, - Medium access, automated test facility (ATF) design model, 212-213 Mesh generation automatic CATFEM system and, 319-320, 324-331 element generation, 321-322 smoothing, 322 three-dimensional approach, 326-331 tree building with, 320-321 two-dimensional approach, - geometry in, 3 - computed tomography with, 291 Meta logics, limits of logic and, - Metynomic models, 16 Minimal rationality, rational machines, - Minsky, Marvin, - frame problem, 24 Mobotics, rational machines, 31-32 Modal logics, characteristics, 21-22 Model base (MBASE), KBDSE techniques, 197-197 composition tree, ATF design model, 210 defined,130-131 Modularity, logic and, 20 Moore's Autoepistemic Logic characteristics, 21-22 Mother concept, set theory, 16 Multiple hypothesis Dempster-Shafer's (D-S) theory, - 8 generalized evidence processing (GEP) theory, 367-370 Multiplication, external agent (EA) and, 27 N Natural language, rational machines and, 32 Network-of-queues (NOQ) analysis, systems modeling, 181-184 Network/process modeling graphical models, 184 SLAM (Simulation Language for Alternative Modeling) block creation, 175-177 case study, 169-170 specific features, 174-179 workstation blocks, 177-179 Neural adaptive control, neurocontroller architectures, 52 Neural nets advantages of, 41-42 applications, 39-41 architectures and learning, - commercial vendors, - 4 first break picking, - 8 function and components, - generic neurocontroller architectures, 51-53 history of, 4 - industrial applications, 61-62 nonlinearities, process control, - process controllers, - robotics applications, - automotive systems, - engine diagnostics, - paint quality assurance, - scheduling systems, - flexible manufacturing systems (FMS), 79-83 workstation maintenance expert system, 74-79 Neural networks, Bayesian/N-P DDF, 388-401 equivalent networks, - training rules, 392-401 Neurocontroller architectures, neural nets, 51-53 NeuroShell neural nets, - 4 Neyman-Pearson (N-P) test Bayesian DDF neural networks, 388-401 parallel sensor topology, 348 single-level local logic, 345-347 statistically independent sensors, 348-351 Dempster-Shafer's (D-S) theory, - 8 distributed decision fusion (DDF), 343-345 monotone property, - nonrandomized testing appendix, - training based on, 9 - 0 Noise signals, FDI, decorrelation methods, 269 Non-contact sensing, apple harvester, 102-103 Non-Destructive Evaluation (NDE) techniques, 290 Non-monotonicity, logic, 20 Non-output units, LMS error signal, - Nondeterministic polynomial/exponential complexity, - Nonlinear systems, neural nets, - Numbers, paradigm of interpreted mathematics, 14 Numerical quantifier logic, characteristics, 23 Nursery operations, automation of, 119-121 INDEX O Object-oriented programming (OOP), Distributed Expert System Environment (DESE), 200-204 Oblique reformating, CATFEM modeling visualization, 311 Observer eigenstructure assignment, robust fault detection and isolation models, 256-257 Observer-based FDI dedicated schemes, 249 eigenstructure assignment, 256-257 nonlinear systems, - fault detection filter, 260-261 nonlinear unknown input observer, 261-264 unknown inputs, - nonlinear, robust FDI via, 261-264 Opaque volume rendering, CATFEM modeling visualization, 312-313 Operational analysis, systems modeling, 181 Orchard spraying, automated systems for, 111-112 Output units, LMS error signal, - P Paint quality assurance, neural nets, - Papert, Seymour, - Paradigm of interpreted mathematics, 13-15 Paradox of zero utility, defined, 26 Parallel distributed processing See Neural nets Parallel sensor topology, Bayesian DDF, 347-348 Parameter estimation, FDI technique and, 264-266 Parity space approach geometric interpretation, 253 robust fault detection and isolation models, 250-254 Pattern matching, neural nets in robotics, 81-83 pcb-info message, information flow models, 220 Perceptrons, in neural nets, 4 - Petri-net modeling, manufacturing systems, 185-188 Physics, of machine limitations, 5-12 irreversible and reversible machines, massively parallel irreversible computer, 9-12 power dissipation, - propagation time limits, 8-9 quantum computers, - Piecewise linear models, CATFEM modeling, 316-317 Planting depth, automatic control of, 107-110 Plasticity, neural nets, 42 Plowing, automation of, ballast position control, 106-107 421 P-method of adaptive analysis, CATFEM modeling, 332 Pneumatic down-pressure system, automated plant depth control with, 109-110 Point, defined, CATFEM techniques, 301 PopPointToModel operator, CATFEM modeling, 308 Possibilistic logic, characteristics, 2 - Potato storage, automated environmental control, 121-1222 Poultry housing, automated environmental control, 122-123 Power limits massively parallel irreversible computers, 10 reversible and irreversible machine limits, - Pressure sensors, agriculture combine control with, 104-105 Primality, external agent (EA) and, - Printed circuit boards (PCBs) expert system (ES) models, 214-217 information flow models, 219 material flow models, 217-219 Process controllers neural nets, - nonlinearities, - Process-oriented languages, in discrete simulation modeling, 141-142 Processes, in discrete simulation modeling, 140-141 Programming languages, simulation techniques with, 135-136 Propagation time limits, theories, - Proportional-integral-derivative (PID) in neural nets, 54 Pruning, automated test facility (ATF) design model, 210 Putnam's theorem, set theory, 18-20 Q Quadtree/octree technique, automatic mesh generation CATFEM systems, 320-321 examples, 322-323 geometry and, 323-324 integration with CATFEM, 324-331 three-dimensional CATFEM, 326-331 two-dimensional CATRE, 314-317 Qualification, logic and, 20 Quantum computers, reversible and irreversible limits, - Queues, discrete simulation modeling, 140 See also Network of queues (NOQ) analysis INDEX 422 R Radial categories in set theory, 16 Rational machines connectionism, 32 language, 32 logic alternatives, 30 minimal rationality, 30-31 mobotics, 31-32 Ray tracing, CATFEM modeling visualization, 309 Rayleigh channels Dempster-Shafer's (D-S) theory, 385-388 generalized evidence processing (GEP) theory, 370-371 signal-to-noise ratio, 370-373 Reason neural net, in robotics applications, 60 Receiver Operating Characteristics (ROCs), Bayesian/N-P DDF, 393-398 Reduced-coulomb energy (RCE), neural nets in robotics, - Reference signals, fault detection and isolation, 269 Region, defined, CATFEM techniques, 302 Reiter's default logic, 21 Relational information, communication and, 28-29 Representative models, fault detection and isolation, 243-244 Residual evaluation, fault detection and isolation, 267-270 decorrelation method, 269 generalized likelihood ratio technique, 267-268 multiple hypotheses testing, 268 threshold selector, 269-271 Resource competition, in manufacturing systems, 132 Reverse engineering, CATFEM modeling visualization, 312-318 CATRE-2D, 314-317 CATRE-3D, 317-318 Reversible computers physical limits, quantum computers, - R-method of adaptive analysis, CATFEM modeling, 332 Robotics Manutec R3 robot fault detection and isolation techniques, 274-281 DC-motor and centrifugal pump, 275, 278 physical parameter estimation, 278-279 residual values, 275, 277 neural net control, - Robustness fault detection and isolation models, 245, 248-267 frequency domain methods, 257-259 observer eigenstructure assignment, 256-257 parity space approach, 250-254 unknown input observers, 254-256 identification-based FDI, hypotheses testing, 265-267 neural nets, 55 in robotics, 68 observer-based FDI, nonlinear unknown input observers, 261-264 Rosenblatt, Frank, perceptron theory, 4 - routing-infor message, information flow models, 221 Sample point, defined, 303 Scale models, characteristics of, 131 Scheduling systems mathematical models for, 183-184 neural nets in robotics, - flexible manufacturing systems (FMS), 79-83 work maintenance expert system (WMES), 74-79 SCOOPS (Scheme's Object Oriented Programming System) Distributed Expert System Environment (DESE), 200-204 type router ES model, 2 - 2 , 236-239 Screw conveyors, automated container handling with, 119 Segmentation automation, peach grading with, 116 Seismic imaging, neural nets and, - 8 Semantics, of mathematics paradigms, 15-16 Sensor failure detection, backpropagation neural net, 57 Sensor integration Bayesian DDF in binary hypothesis testing multi-level local logic, 353-375 single-level local logic, - Bayesian / N-P DDF and neural networks, 380-401 backpropagation based on mean-squared error, 392-398 Kaiman filter training, 400-401 Neyman-Pearson based training, 9 - 0 Dempster-Shafer's theory, 357-388 distributed decision fusion and evidence processing, 339-343 INDEX likelihood ratio and Neyman-Pearson test, 343-345 Set theory categories, 17 in discrete simulation modeling, 140 hedges, 16-17 idealized cognitive models (ICMs), 17 linguistic qualifiers, 16-17 mathematical paradigms, 16-20 metynomic models, 16 Putnam's theorem, 18-20 radial categories, 16 Shell, defined, CATFEM techniques, 302 Signal-to-noise ratio (SNR) Dempster-Shafer's (D-S) theory, 386-388 generalized evidence processing (GEP) theory, 371-374 Simulation Craft, environmental modeling, 192-193 Simulation packages, characteristics of, 142 Simulation techniques, for manufacturing systems background, 133-138 modeling fundamentals, 134-138 variability in, 137-138 Simulation workbench, simulation modeling, 144-146 SLAM (Simulation Language for Alternative Modeling) case study, 168-180 discrete event and combined models, 170-171 discrete event portion of model, 179-180 network portion of model, 174-179 network/process modeling, 169-170 nodes and parameters for, 170 processor components, 171-172 Slice, defined, digital replica, 303 Small fruit sorting, automation of, 117-118 Smearing, CT imaging, 2962-97 Smoothing automatic mesh generation, 322 three-dimensional CATFEM modeling, 330 Solid modeling, CATFEM techniques, 300-301 Solid-set irrigation control systems, automated design for, 114-115 Space complexity, defined, Space limits, Special-purpose languages, discrete simulation modeling, 145, 147 State space equations fault detection and isolation models, 244-245 nonlinear unknown input observer, robust FDI via, 261-264 States, in discrete simulation modeling, 140-141 Static models, fault detection and isolation, 242 423 Statistically independent sensors, Bayesian DDF, 348-351 Steam generator, fault detection and isolation techniques, 280-281 Stochastic systems, simulation modeling of, 137 Structured uncertainties, fault detection and isolation models, 246 Sum projection, CATFEM modeling visualization, 309 Super-sensors, Bayesian DDF, 351-353 Supervised control, neurocontroller architectures, 51 Support function, Dempster-Shafer's (D-S) theory, 376-377 Surface, defined, CATFEM techniques, 301 Surface irrigation systems automation, 113-114 sw-download message, information flow models, 220 sw-request message, information flow models, 220 Symbolic theory of mathematics, 19 System architectures, neural nets in robotics, 68 System entity structure (SES) formalism, KBDSE techniques, 195-199 entity, aspect and specialization nodes, 197-198 multiple decomposition, 198 T TAGUS (Topology and Geometry Utility system), 302-303 Taylor series approximation fault detection and isolation models, 246 observer-based FDI, 260-261 Temporal logic, characteristics, 23 Three-dimensional CATFEM technique, 326-331 Three-dimensional CT image generation, 294-296 metrology, 294-295 Three-tank systems, fault detection and isolation, 273-274 Threshold selector, fault detection and isolation, 269-271 Thresholding technique CATFEM modeling digital replica, 304-305 CT imaging, edge detection, 298-299 generalized evidence processing (GEP) theory, 366-367 Time horizons, robust fault detection and isolation models, 251 Time resource limits, Tomato grading, automation of, 118-119 INDEX 424 Topological entities, CATFEM techniques, 301-302 Tractor operations, automation of, 100-110 Trained data interpolation, neural nets, 41-42 Transformational information, communication and, - Translucent volume rendering, CATFEM modeling visualization, 312-313 Tree building, automatic mesh generation, 320-321 Turing machine, reversible and irreversible limits, Turing's theorem, infinite and finite, - Two-dimensional CATFEM technique, - Vision systems peach grading with, 116 tractor operations, 101-102 Visual inspection systems, neural nets, - 6 Visualization, in CATFEM modeling, 308-312 volume modeling, 309-312 volume rendering, 308-309 Volume defined, digital replica, 303 modeling, CATFEM visualization, 309-312 rendering, CATFEM visualization, 309-313 von Neumann machine, propagation time limits, 8-9 Voxel defined, digital replica, 3 - three-dimensional CATRE techniques, 318 U W Universal quantum computer (UQC), reversible and irreversible limits, Unknown input observer nonlinear, robust FDI, 261-264 robust fault detection and isolation models, 254-256 Widrow, Bernard, 46 Workstation maintenance expert system (WMES) block diagram, - 7 neural nets in robotics, - Z V Variability, in simulation modeling, 138 Vertex, defined, CATFEM techniques, 302 Zero energy thesis, reversible and irreversible limits, ... California and College of Engineering University of Washington Seattle, Washington V O L U M E : MANUFACTURING AND AUTOMATION SYSTEMS: TECHNIQUES AND TECHNOLOGIES Part of ® ACADEMIC PRESS, INC... result, this is a particularly appropriate time to treat the issue of manufacturing and automation systems in this international series Thus, this is Part of a five -part set of volumes devoted... applications of existing sensors, systems, and techniques to biological systems The properties and behavior of plants and animals vary greatly both among and within species The response of a biological