Intelligent Control

Fuzzy Logic Applications

Clarence W. de Silva

• Publisher: CRC Press
• Year: 2018
• eBook ISBN:
  9781351437684
• Topic: Tecnología e ingeniería
• Subtopic: Ingeniería computacional

The emergence of fuzzy logic and its applications has dramatically changed the face of industrial control engineering. Over the last two decades, fuzzy logic has allowed control engineers to meet and overcome the challenges of developing effective controllers for increasingly complex systems with poorly defined dynamics. Today's engineers need a working knowledge of the principles and techniques of fuzzy logic-Intelligent Control provides it. The author first introduces the traditional control techniques and contrasts them with intelligent control. He then presents several methods of representing and processing knowledge and introduces fuzzy logic as one such method. He highlights the advantages of fuzzy logic over other techniques, indicates its limitations, and describes in detail a hierarchical control structure appropriate for use in intelligent control systems. He introduces a variety of applications, most in the areas of robotics and mechatronics but with others including air conditioning and process/production control. One appendix provides discussion of some advanced analytical concepts of fuzzy logic, another describes a commercially available software system for developing fuzzy logic application. Intelligent Control is filled with worked examples, exercises, problems, and references. No prior knowledge of the subject nor advanced mathematics are needed to comprehend much of the book, making it well-suited as a senior undergraduate or first-year graduate text and a convenient reference tool for practicing professionals.

1	CONVENTIONAL AND INTELLIGENT CONTROL

INTRODUCTION

Fuzzy logic control belongs to the class of “intelligent control”, “knowledge-based control”, or “expert control”. Even though it is incorrect to claim that fuzzy logic control is always better than the conventional types of “crisp” (“hard” or “inflexible”) control, it has several advantages. Our objective in this chapter is to outline several of the prominent techniques of control, both modern and classical, and indicate drawbacks of these methods which make “soft” intelligent control more appropriate in some applications.

With steady advances in the field of artificial intelligence (AI), especially pertaining to the development of practical expert systems or knowledge systems, there has been considerable interest in using AI for controlling complex processes (Francis and Leitch, 1984; Goff, 1985; Isik and Mystel, 1986; Saridis, 1988; de Silva, 1991). The rationale for using intelligent control systems may be related to difficulties which are commonly experienced by practicing control engineers. It is generally difficult to accurately represent a complex process by a mathematical model or by a simple computer model. Even when the model itself is tractable, control of the process using a hard control algorithm might not provide satisfactory performance. Furthermore, it is commonly known that the performance of some industrial processes can be considerably improved through the higher-level control actions (tuning actions, in particular) made by an experienced and skilled operator, and these actions normally cannot be formulated by hard control algorithms.

There are many reasons for the practical deficiencies of hard (crisp) control algorithms. If the process model itself is inaccurate, a conventional, model-based control can generate unsatisfactory results. Even with an accurate model, if the parameter values are partially known, ambiguous, or vague, appropriate estimates must be made. Crisp control algorithms that are based on such incomplete information will not usually give satisfactory results. The environment with which the process interacts may not be completely predictable, and it is normally not possible for a hard algorithm to accurately respond to a condition that it did not anticipate and that it could not “understand”. A hard control algorithm will need a complete set of data to produce results, and indeed, program instructions and data are conventionally combined into the same memory of the controller. If some data were not available, say, unexpectedly as a result of sensing problems, the inflexible algorithm will naturally fail.

Humans are flexible (or soft) systems. They can adapt to unfamiliar situations, and they are able to gather information in an efficient manner and discard irrelevant details. The information which is gathered need not be complete and precise and could be general, qualitative, and vague because humans can reason, infer, and deduce new information and knowledge. They have common sense. They can make good decisions, and also can provide logical explanations for those decisions. They can learn, perceive, and improve their skills through experience. Humans can be creative, inventive, and innovative. It is a very challenging task to seek to develop systems that possess even a few of these simple human abilities. This is the challenge faced by workers in the field of AI and knowledge-based systems. However, humans have weaknesses too. For example, they tend to be slow, inaccurate, forgetful, and emotional. In other areas, of course, one might argue that these human qualities are strengths rather than weaknesses.

In a knowledge-based system one seeks to combine the advantages of a computer with some of the intelligent characteristics of a human for making inferences and decisions. It is clearly not advisable to attempt simply to mimic the human thought process without carefully studying the needs o...

Table of contents

• Cover
• Title Page
• Copyright Page
• Dedication
• Table of Contents
• 1. Conventional and Intelligent Control
• 2. Knowledge Representation and Processing
• 3. Fundamentals of Fuzzy Logic
• 4. Fuzzy Logic Control
• 5. Knowledge-Based Tuning
• 6. Knowledge-Based Control of Robots
• 7. Servo Motor Tuning
• 8. Hierarchical Fuzzy Control
• 9. Intelligent Restructuring of Production Systems
• 10. Future Applications
• Appendix A: Further Topics on Fuzzy Logic
• Appendix B: Software Tools for Fuzzy Logic Applications
• Index