The block diagram shown in Figure 4- 1 illustrates the basic concept of the control system. After the action signal passes through (via) the control system component, it provides an indication, and the purpose of the system is to control the variable c within the indication. Generally speaking, the controlled variable is the output of the system, and the action signal is the input of the system. For a simple example, the steering control of a car, the direction of the two front wheels can be regarded as the controlled variable, that is, the output; The position of its steering wheel can be regarded as the input, that is, the action signal E. For another example, if we want to control the speed of the car, the total pressure of the throttle is the action signal, and the speed is the controlled variable.
Figure 4- 1 Block diagram of basic control system
We call the physical quantity to be controlled as the controlled quantity or output of the system. The excitation signal used to make the system have expected performance or output is defined as the control quantity or input quantity of the system. The factor that makes the controlled quantity deviate from the expected value is called disturbance quantity. Automatic control process is a process of trying to eliminate the influence of interference factors to keep the controlled quantity changing as expected.
Therefore, the automatic control system can understand that any system, without direct participation, makes the controlled object or process run automatically according to the predetermined law through the control device.
After mastering the simple manufacturing technology, in order to reduce or replace their own labor, human beings came up with the idea of creating automation devices, which is the original source of the idea of automatic control. The development process of automatic control technology has roughly experienced four stages: ancient stage,17-19th century stage,19th century stage, World War II stage and post-world war II stage. During this period, classical control theory and modern control theory developed from scratch.
2 early development of automatic control technology
In ancient times, from about14th century BC to 1 1 th century BC, ancient civilizations in the world, including China, Egypt and Babylon, all appeared leaky pots that could automatically time due to the needs of production development. Zhang Heng, a scientist in Han Dynasty, invented armillary sphere and seismograph. The outline of the model is shown in Figure 4-2, and the idea of automatic control is applied to astronomical observation instruments and earthquake observation instruments. During the Three Kingdoms period, the south-pointing car appeared, which is a successful example of using automatic control thought in azimuth measuring instruments. During the Northern Song Dynasty in China (A.D. 1086- 1089), as shown in Figure 4-3, Su Song and Han Gonglian made a water transport instrument platform, which was an instrument integrating an astronomical observation instrument, an astronomical demonstration instrument and an automatic timing device. The whole system is a closed-loop nonlinear automatic control system based on negative feedback principle.
Figure 4-2 Outline Diagram of Scale Seismograph Model
Figure 4-3 Water Instrument Image Platform
Semi-automatic simple machines appeared in ancient Egypt and Greece, such as the automatic opening device of the church gate, the bronze priest who sprinkled holy water automatically, the coin-operated holy water jar, and the bronze bird who sang automatically at the church gate. These are some unrelated primitive automatic devices and some personal inventions. /kloc-after the 0/7th century, with the development of production and the progress of science, a variety of automatic devices appeared in Europe, including: 1642, French physicist Pascal invented an adder that can carry automatically; 1657 Dutch mechanic huygens invented the clock; 1745, the British mechanic e Li Faming invented an airflow mill with wind direction control, which uses the steering function of the tail wing to make the main wing aim at the wind direction; 1765, Russian mechanic Polzunov invented the float valve water level regulator, which can automatically control the water level of the steam boiler. During this period, due to the needs of production development, automatic control technology was produced.
1788, British scientist Watt (Figure 4-4) invented the centrifugal throttle valve, also known as the flying ball governor. As shown in Figure 4-5, it is used to control the steam valve of the steam engine to form a closed-loop automatic control system for the speed of the steam engine, thus realizing the control of the speed of the steam engine by the centrifugal throttle valve. Watt's invention promoted the wide application of modern automatic regulating device, and had an important influence on the first industrial revolution brought by steam engine and the development of control theory after that. Inventions in other countries include: 1854 electromagnetic governor invented by Russian mechanic and electrician Constantine; 1868, French engineer Falco invented the feedback regulator to adjust the steam valve and manipulate the rudder of the steam ship. This is the servo mechanism that was widely used later. Before 1868, automation technology was only a few individual inventions and simple applications, so it was called the first stage. After 1868, the theoretical analysis and large-scale wide application of automatic control system began gradually, so it was called the second stage.
Figure 4-4 Watts
Figure 4-5 Diagram of Steam Engine Controlled by Watt Centrifugal Throttle Valve
1- steam engine; 2- steam valve; 3- Governor; 4- Load
3 the early development of automatic control theory
Various simple automatic control devices can improve the production process and improve the production efficiency. Although the invention of this technology went through a long historical process before18th century, there was no theoretical analysis and mathematical description, but they played a leading role in the formation of automation technology, and they were all summarized from practical experience. 17- 18 century is the gradual formation period of automation technology, followed by the development period of modern automation technology, and mathematical description and theoretical analysis have played a vital role. The first problem people encounter is the stability of the automatic regulator, because the centrifugal governor invented by Watt sometimes causes instability of the system and makes the steam engine oscillate violently. In the19th century, the stability of the ship's autopilot was discovered again. These problems have aroused widespread concern, and some mathematicians try to describe and analyze the stability of the system with differential equations. The original mathematical description of the automatic control system is the British physicist Maxwell (Figure 4-6), who published an article "On the Governor" in 1868, summarizing the theory of the governor without static error.
1877, the British mathematician E.J.Routh put forward the famous Routh stability criterion, which is an algebraic stability criterion, and can judge the stability of the control system according to the coefficients of the differential equation. 1895, German mathematician A. Hurwitz (Figure 4-7) put forward the famous Hurwitz stability criterion, which is another form of algebraic stability criterion. Routh-leonid hurwicz stability criterion is an important criterion to predict the stability of the regulator in advance according to the transfer function or differential equation. 1892, Russian mathematician Lyapunov published a monograph on general problems of motion stability, gave a strict definition of the concept of motion stability in the form of mathematical language, and gave two methods to judge the stability of the system.
Figure 4-6 Maxwell
Figure 4-7 Hurwitz
After entering the 20th century, due to the need of industrial revolution, people began to adopt automatic control devices to solve the control problems raised in industrial production. The application of automatic controller marks that automation technology has entered a new historical period. In the meantime, the controller is a device to track the given value, so that some physical quantities can be kept near the given value. The wide application of various automatic control devices in industrial production has promoted the analysis and comprehensive research of the regulating system. After the 1920s, the United States began to adopt proportional, integral and differential regulators, referred to as PID regulators. PID regulator is an analog regulator, which is now used in many factories. In the first 20 years of the 20th century, the feedback control structure has been widely used in automatic controllers. Since the 1920s, more and more people began to study the feedback control system in theory.
1925, British electrical engineer O. Hevesey applied Laplace transform to solve the problem of electric network, and obtained the transient process by calculus. During the period of 1927, when Bell Telephone Laboratory solved the distortion problem of electron tube amplifier, electrical engineer H.S. Black introduced the concept of feedback from the angle of electrical signal. 1932, American telecom engineer Nyquist (Figure 4-8) put forward the famous Nyquist stability criterion, which can directly draw the Nyquist diagram according to the transfer function of the system to judge the stability of the feedback system. 1938 mikhailov, a Soviet electrical engineer, applied frequency method to study the stability of automatic control system and put forward the famous mikhailov stability criterion.
Figure 4-8 Nyquist
With the development of automatic control theory, the ideas of program control, logic control and automata have been developed. 1833, the British mathematician C. Babbage first put forward the concept of program control when designing analytical automata. He tried to use the punched card method designed by the French inventor J.M. jacquard to realize the program control of analytical automata. 1936 British mathematician Turing developed the famous Turing machine, which became the prototype of modern digital electronic computer. He defined the computable function class with Turing machine, and established the algorithm theory and automaton theory. 1938, Shannon, an American electrical engineer, Nakajima, a Japanese mathematician, and Shestakov, a Soviet scientist 194 1 year independently established the theory of logical automata, which was composed of relays with only two working states to realize logical control. In addition, Shannon also established the information theory.
4 the formation of classical control theory
The development history of automatic control technology is the history of human beings using their own intelligence to extend and expand organ functions. Automation is the crystallization of modern science and technology and modern industry, and its development fully embodies the comprehensive application of science and technology. Automatic control technology is developed with the needs of society, especially the control of production equipment and military equipment, as well as the needs of aerospace industry. The classical control theory formed during the Second World War has played an important role in promoting the development of postwar automatic control technology. During World War II, Germany's air superiority and Britain's defensive position forced scientists from the United States, Britain and Western Europe to concentrate on solving military technical problems such as air defense fire control system and aircraft automatic navigation system. In the process of solving these problems, classical control theory has been formed, and various precise automatic adjustment devices have been designed, which has created a new scientific field of system and control.
During World War II, feedback control method was widely used to design and develop military systems such as aircraft autopilot, artillery positioning system and radar antenna control system (Figure 4-9). The complexity of these systems and the pursuit of high performance for fast tracking and precise control urgently require the expansion of existing control technologies, which leads to the emergence of many new ideas and methods, and also promotes the research on nonlinear systems, sampling systems and stochastic control systems.
1945, American mathematician wiener (Figure 4- 10) extended the concept of feedback to all control systems. 1948, wiener published the book cybernetics, which laid the foundation of cybernetics. In the same year, Shannon, an American telecommunications engineer, published Mathematical Theory of Communication, which laid the foundation of information theory. Wiener and Shannon study the motion of the system from two aspects: control and information. Wiener also studied the essence of feedback control from the angle of information. Since then, people have a deeper understanding of feedback and information. From 65438 to 0954, China system scientist Qian Xuesen comprehensively summarized the classical control theory, and further promoted it to a higher theoretical level, and published the book "Engineering Cybernetics" in the United States. The purpose of engineering cybernetics is to study those parts of cybernetics that can be directly used in engineering design controlled systems. Engineering cybernetics makes it possible for people to have a broader vision and observe related problems in a more systematic way, so they can often get more effective new methods to solve old problems, and may also reveal new prospects that have never been seen before.
Figure 4-9 Radar during World War II
Figure 4- 10 sausage
At that time, this new discipline was called servo mechanism theory in the United States and automatic adjustment theory in the Soviet Union, mainly to solve the control problem of single variable. At that time, the concepts of transfer function and frequency response were widely used in the analysis and design of feedback servo system. The most commonly used methods are Nyquist method (1932), Bode method (1945) and Evans method (1948). Evans method, also called root locus method, was put forward by American telecom engineer W.R.Ewans in 1948. The frequency method developed in 1930s and 1940s, which is suitable for single variable regulation and servo system design, laid the foundation of classical control theory. Later, frequency method became the main method to analyze and design linear automatic control system. This method can not only determine the design direction qualitatively, but also be a simple approximate calculation tool. Therefore, this method is particularly effective and popular for the engineering design of control systems that still need to rely on experience and trial and error to a great extent.
The name of classical control theory was put forward at the first national joint automatic control conference in 1960. At this meeting, the topic of studying univariate control in the field of system and control is called classical control theory; The subject of studying multivariable control in the field of system and control is called modern control theory.
1952, the first CNC machine tool was born. The application of CNC machine tool technology not only brought revolutionary changes to the traditional manufacturing industry, but also made the manufacturing industry a symbol of industrialization. The NC machine tool is shown in Figure 4- 1 1.
The electronic digital computer invented in the mid-1940s ushered in a new era of digital program control. Although it was limited to automatic calculation at that time, it laid the foundation for the rapid development of automation technology in the 1960 s and 1970 s.
196 1 year, the world's first industrial robot (Figure 4- 12) was born, which greatly promoted the automation of industrial production lines.
Figure 4- 1 1 CNC machine tool
Figure 4- 12 industrial robot
Transistor computers appear in 1958, integrated circuit computers in 1965, and single-chip computers in 197 1. The appearance of microprocessor has a great influence on control technology. Control engineers can easily use microprocessors to realize all kinds of complex control, making integrated automation a reality.
1957, the inaugural meeting of IFAC was held. Delegations from 18 countries attended the meeting, and China was one of the sponsors. Since 1960, the international conference on automatic control has been held every three years, and periodicals such as Automation and IFAC Newsletter have been published. The establishment of IFAC marks the maturity of automatic control and promotes the new development of system and control field through international cooperation.
5 the formation and development of modern control theory and technology
Since 1950s, there have been many new developments in classical control theory. Various new theories and methods have gradually penetrated into the research of control theory. However, by the end of 1950s, it was found that when the classical control theory was extended to multivariable systems, wrong conclusions would be drawn, and the classical control theory had its limitations. In order to solve and overcome the problems encountered, modern control theory came into being.
1) system identification, modeling and simulation
In modern control theory, the design of optimal controller, the design of observer and the configuration of zero poles are all carried out on the premise of knowing the dynamic equation or state equation of the system. These system synthesis methods often choose a convenient description form, regardless of how to obtain these mathematical models. In practical application, the model of the system is often unknown. For complex systems, it is often difficult to establish a model with known physical laws. Therefore, the method of establishing mathematical model according to the input and output data of the system has been developed, and the theory and method of system identification have gradually formed.
In the process of analysis, synthesis and design of automatic control system, besides theoretical calculation, it is often necessary to carry out experimental research on the characteristics of the system. Obviously, it is impossible to test the system before it is established. For the existing system, if the system is very complex, it is not allowed, and sometimes it is even impossible to carry out experiments on the actual system, regardless of economy or safety. Therefore, it is necessary to test the system on the simulation equipment, including establishing, modifying and reproducing the model of the system. This testing process is usually called system simulation. At present, system identification, modeling and simulation have become very active and important topics in the field of system and control.
2) adaptive control and self-tuning regulator
In the early 1950s, people began to pay attention to the research of adaptive control in order to design the automatic navigation system of aircraft and make it fly in a wide range of speed and altitude. The development of control theory in 1960s deepened the understanding of adaptation process. Adaptive control can be described by a stochastic recursive process. In 1970s, due to the new breakthrough in microelectronics, adaptive control can be realized by simple and economical methods. At present, three methods of parameter adaptive control have been developed, namely, gain adjustment method, model reference method and self-tuning regulator.
3) Telemetry, remote control and remote sensing
At the end of 19, there was an attempt of remote measurement and control. In the 1920s, telemetry and remote control began to enter the practical stage, which was used to control signals and turnouts on railways. 1930, the world's first radio altimeter was sent to measure atmospheric meteorological data. This is the first relatively complete radio telemetry equipment. In the 1940s, large-scale power systems, oil and gas pipeline transportation systems and urban public utilities systems all needed centralized monitoring of geographically dispersed objects through remote sensing, remote communication, remote control and remote dispatching, which promoted the development of remote sensing and remote control systems. The Soviet Union and Eastern European countries call this system telecontrol system.
Telemetry is a long-distance measurement of some parameters of the measured object. Generally, some parameters of the measured object are measured by sensors and converted into electrical signals, and then these electrical signals are transmitted to remote telemetry terminals for processing, display and recording by multi-channel communication and data transmission technology. Telecommunication is to measure the working limit state (whether it works or not) of the long-distance measured object. Remote control is the remote control of the controlled object. Remote control technology comprehensively applies automatic control technology and communication technology to realize remote control and monitoring of the controlled object. Among them, the adjustment of the working state of the remote control object is called remote control adjustment. Remote control of the controlled object moves according to certain guidance laws, which is called guidance, that is, control and guidance, and is widely used in aerospace, aviation, navigation and other fields.
Since 1960s, remote sensing technology has developed rapidly. Remote sensing is a sensor installed on aircraft, satellites and other vehicles, which collects electromagnetic waves reflected or emitted by ground targets and uses these data to obtain target information. After the development from aerial remote sensing with airplane as the main carrier to space remote sensing with earth satellite and space shuttle as the main carrier, people can observe various phenomena and their changes on the earth periodically and quickly from the height of space, thus making the exploration of earth resources and the study of some natural phenomena on the earth enter a new stage and have been applied to agriculture, forestry, geology, geography, ocean, hydrology, meteorology, environmental protection and military reconnaissance.
4) Integrated automation
From the late 1950s to the early 1960s, automatic factories controlled by electronic digital computers began to appear. At the end of 1960s, many automatic production lines appeared in manufacturing industry (Figure 4- 13), and industrial production began to develop from local automation to comprehensive automation. Since 1970s, great breakthroughs in microelectronics, computer and robotics have promoted the rapid development of integrated automation. In process control, 1975 began to appear distributed control system, which made the process automation reach a high level. In manufacturing industry, flexible manufacturing system (FMS), computer-aided design (CAD) and computer-aided manufacturing (CAM) developed on the basis of adopting group technology, CNC machine tools, machining centers and group control have become the basis of factory automation. Many industrial robots, inductive unmanned trolleys, automated warehouses and unmanned forklifts developed in 1970s have become powerful tools for integrated automation. The flexible manufacturing system was developed in 1960s, and the first flexible manufacturing system in America was put into production in 1972. From the late 1970s to the early 1980s, flexible manufacturing systems developed rapidly, and handling robots and assembly robots were widely used. 1982 10 the first international conference on flexible manufacturing systems was held in Brighton, England.
Figure 4- 13 Automatic Production Line
5) The birth of large-scale system theory.
Since the mid-1960s, the application of system and control theory has gradually penetrated into agriculture, commerce and service industry, as well as biomedicine, environmental protection and social economy. Due to the high development of modern social science and technology, many large-scale systems need comprehensive management, and modern control theory can't solve such complicated problems, so the system and control theory urgently need new breakthroughs. In terms of computer technology, database technology began to develop in the early 1960s, and 1970 proposed a relational database. By the 1980s, database technology had reached a considerable level. At the end of 1960s, the combination of computer technology and communication technology produced data communication. From 65438 to 0969, the first phase of ARPA (Advanced Research Agency of US Department of Defense) was successfully put into use, which opened a new era of computer network. In 1980s, database technology and computer network created favorable conditions for realizing management automation. One of the core issues of management automation is office automation, which is a comprehensive technology developed from 1970s and matured in 1980s. Office automation has laid a good foundation for management automation.
The International Federation of Automatic Control (IFAC) held the first large-scale system conference in Udina in 1976, and the second large-scale system conference in Toulouse in 1980. The Institute of Electrical and Electronics Engineers (IEEE) held an international symposium on large-scale systems in Virginia Beach, Virginia, USA on June 1982+00. 1980, the international journal Large-scale System-Theory and Application was officially published in the Netherlands. These activities mark the birth of large-scale system theory.
6) artificial intelligence and pattern recognition
Although it is a long-standing wish of human beings to use machines to simulate human intelligence, it was not realized until the appearance of electronic computers. 1936, Turing put forward the idea of logical reasoning with machines. Since 1950s, the research of artificial intelligence has been based on giving full play to the role of computers.
The early research on artificial intelligence began with exploring people's problem-solving strategies, that is, starting with puzzles, playing chess and proving theorems with little difficulty, summarizing the laws of human psychological activities when solving problems, and then using computer simulation to make computers show some intelligence. 1948, American mathematician Weiner first put forward the problem of making chess machines in Notes on Cybernetics. From 65438 to 0954, Samuel, an engineer of International Business Machines Corporation (IBM), compiled a checkers program with a heuristic program and stored it in an electronic digital computer, making a chess machine that can accumulate chess experience. 1959 The chess machine defeated its designer. 1956, herbert simon and Allen Newell developed a program called logic theorist, and proved 33 theorems in 52 theorems in chapter 2 of Whitehead and Russell's famous book Principles of Mathematics with electronic digital computer. 10 In 956,10 scientists, including M.L. Minsky, J. McCarthy, Newell and Simon, launched a seminar on artificial intelligence at Dartmouth University, marking the formal birth of the subject of artificial intelligence. 1960 The four founders of artificial intelligence, namely McCarthy of Stanford University, Minsky of MIT, Newell and Simon of Carnegie Mellon University, formed the first artificial intelligence research group, which effectively promoted the development of artificial intelligence. Since 1967, Machine Intelligence has been published regularly, with a total of 9 episodes. The journal Artificial Intelligence was published in 1970. The International Conference on Artificial Intelligence (IJCAI) has been held every two years since 1969. These activities further promoted the development of artificial intelligence. The rapid development of microelectronics and microprocessors since 1970s has combined artificial intelligence with computer technology. On the one hand, the achievements of artificial intelligence are widely used in the design of advanced computers. On the other hand, artificial intelligence is realized by using super microprocessors, which greatly accelerates the research and application of artificial intelligence. The foundation of artificial intelligence is knowledge acquisition, representation and reasoning technology. The commonly used artificial intelligence languages are LISP and PROLOG. The research fields of artificial intelligence involve natural language understanding, natural language generation, machine vision, machine theorem proving, automatic programming, expert system and intelligent robot. Artificial intelligence has developed into a frontier field of system and control research.
1977 E.A. feigenbaum put forward the problem of knowledge engineering at the fifth international conference on artificial intelligence. Knowledge engineering is a branch of artificial intelligence, and its central task is to construct an expert system. 1973 ——1975 feigenbaum led a research team of Stanford University to successfully develop a medical expert system MYCIN for the diagnosis and treatment of blood infectious diseases and meningitis. This system can learn the knowledge of expert doctors, imitate doctors' thinking and diagnostic reasoning, and give reliable diagnosis and treatment suggestions. 1978 feigenbaum and others have successfully developed an advanced chemical expert system, DENDRAL. 1982, American scholar W.R. Nelson developed REACTOR, an expert system for diagnosing and handling nuclear reactor accidents. China has also successfully developed an expert system of traditional Chinese medicine and an expert system of silkworm breeding. Now expert system has been applied to medicine, machine fault diagnosis, aircraft design, geological exploration, molecular structure and signal processing.
In order to expand the application of computers, computers can directly accept and process all kinds of natural mode information, that is, language, characters, images, scenery and so on. The research of pattern recognition has attracted people's attention. 1956 selfridge and others developed the first character recognition program, followed by a character recognition system and an image recognition system, forming a pattern recognition theory with statistics and structural methods as the core. The research of speech recognition and natural language understanding has also made great progress, providing a new interface for direct communication between people and computers.
From the late 1960s to the early 1970s, Massachusetts Institute of Technology, Stanford University and Edinburgh University in the United Kingdom conducted many theoretical studies on robotics, and noticed that all technologies of artificial intelligence were integrated to develop intelligent robots, such as the hand-eye devices of MIT and Stanford University and the robots with vision and touch of Hitachi. Robots have attracted people's attention because of their great advantages in improving productivity, replacing people from dangerous and harsh working environment and expanding the scope of human activities. Robot technology has developed rapidly and has been widely used, which has played a great role in industrial production, equipment inspection and maintenance of nuclear power plants, marine survey, underwater oil exploitation, space exploration and so on. The military robot being studied also has great potential application value. The design, manufacture and application technology of robots form robotics.
Summarizing the experience and lessons of artificial intelligence research, people realize that machines must have the knowledge of human experts to solve problems, and the essence of artificial intelligence should be how to transfer human knowledge to machines. 1977, feigenbaum initiated expert system and knowledge engineering, and then developed knowledge engineering with knowledge acquisition, representation and application as its core. Since 1970s, artificial intelligence scholars have developed various expert systems for medical diagnosis, geological exploration, chemical data interpretation and structure interpretation, oral and image understanding, financial decision-making, military command, large-scale integrated circuit design and so on. The development of intelligent computers, new sensors and large-scale integrated circuits provides new control methods and tools for advanced automation.
Since 1950s, some progress has been made in exploring the mechanism of feeling and thinking of living things and human beings, as well as simulating with machines, such as self-organizing system, neuron model, neuron network brain model and so on. It is enlightening to the development of automation technology. The general system theory, dissipative structure theory, synergetics and hypercycle theory developed at the same time provide new theories and methods for the development of automation technology.