China
Africa
Explore chapters and articles related to this topic
Introduction
Published in Cong Wang, David J. Hill, Deterministic Learning Theory, 2018
Intelligent control was originally developed to motivate discussion of several areas related to learning control, with the emphases on problem solving or high-level decision capability [57]. Compared with learning control, intelligent control is a more general term describing the intersection of the fields of automatic control systems and artificial intelligence. The motivation of intelligent control lies in the attempt by control engineers to design more and more human-like controllers with adaptation and learning capabilities. On the other hand, many research activities in artificial intelligence, including machine learning and pattern recognition, might usefully be applied to solve learning control problems. This overlap of interest between the two areas has created many points of interest for control engineers. Furthermore, it was proposed that intelligent control should analytically investigate control systems with cognitive capabilities that could successfully interact with the environment. Therefore, in the early 1980s intelligent control was considered as a fusion of research areas in systems and control, computer science, and operations research, among others [197,198].
Fuzzy Logic and Fuzzy Systems
Published in Yi Chen, Yun Li, Computational Intelligence Assisted Design, 2018
With the development of computational intelligence, human inference‐oriented fuzzy systems (FS) and fuzzy logic control (FLC) have received increasing attention world‐wide. For example, a fuzzy controller incorporates uncertainty and abstract nature inherent in human decision‐making into intelligent control systems. It tends to capture the approximate and qualitative boundary conditions of system variables (as opposed to the probability theory that deals with random behavior) by fuzzy sets with a membership function. Such a system flexibly implements functions in near human terms, i.e., IF‐THEN linguistic rules, with reasoning by fuzzy logic, which is a rigorous mathematical discipline. Hence, it is termed a type of expert systems that handles problems widespread with ambiguity. It is well known for its capability in dealing with non‐linear systems that are complex, ill‐defined or time‐varying. In addition, fuzzy systems are reliable and robust and are straightforward to implement [Ng, et al. (1995)].
Fuzzy Control and Stability
Published in Ali Zilouchian, Mo Jamshidi, Intelligent Control Systems Using Soft Computing Methodologies, 2001
The aim of this chapter is to define fuzzy control systems and cover relevant results and development. Traditionally, an intelligent control system is defined as one in which classical control theory is combined with artificial intelligence (AI) and possibly OR (Operations Research). Stemming from this definition, two approaches to intelligent control have been in use. One approach combines expert systems in AI with differential equations to create the so called expert control, while the other integrates discrete event systems (Markov chains) and differential equations [1]. The first approach, although practically useful, is rather difficult to analyze because of the different natures of differential equations (based on mathematical relations) and AI expert systems (based on symbolic manipulations). The second approach, on the other hand, has well developed and solid theory, but is too complex for many practical applications. It is clear, therefore, that a new approach and a change of course are called for here. We begin with another definition of an intelligent control system. An intelligent control system is one in which a physical system or a mathematical model of it is being controlled by a combination of a knowledge-base, approximate (humanlike) reasoning, and/or a learning process structured in a hierarchical fashion. Under this simple definition, any control system which involves fuzzy logic, neural networks, expert learning schemes, genetic algorithms, genetic programming or any combination of these would be designated as intelligent control.
Height control in GMA-AM using external wire as controlling variable
Published in Materials and Manufacturing Processes, 2023
Yiyang Zhang, Jun Xiong, Dayong Li, Guangjun Zhang
After establishing the dynamic model, the next step is to design a reasonable controller to adjust the process parameters to keep the NTSD of each layer consistent with the given value. Considering that the GMA-AM process is nonlinear and time-variant, a parameter self-tuning fuzzy controller with automatically adjustable parameters is designed to control the NTSD for obtaining the optimal control performance. Fuzzy control theory is an intelligent control method based on human experience, and performs intelligent control via imitating human reasoning process and decision-making processes such as fuzzification, fuzzy reasoning, fuzzy decision-making, and defuzzification. The parameters of the traditional fuzzy controller are fixed, which cannot simultaneously meet the system performance requirements such as fast dynamic response, high steady-state accuracy, and small overshoot. Compared with the traditional fuzzy controller, the parameter self-tuning fuzzy controller possesses variable parameters such as quantification and scale factor. The parameters can be dynamically adjusted according to the correction rules to meet the requirements of various performance indicators. The parameter self-tuning fuzzy controller possesses the merits of simple structure, strong robustness, and strong dynamic response capability. The structure of the parameter self-tuning fuzzy controller is presented in Fig. 10.
Predictive control of coke oven flue temperature based on orthogonal neural network
Published in Australian Journal of Electrical and Electronics Engineering, 2020
At present, the control method of coke oven flue temperature is mainly focused on PID algorithm. The controller designed by PID algorithm can only meet the control performance or the stability of the controlled object under a certain operating conditions. PID algorithm cannot make the controlled object have good control performance in a large range. Therefore, flue temperature control of coke oven based on PID has great limitations. With the wide application of computer in the field of industry, many scholars put forward many new methods to control the flue temperature of the coke oven. These control methods include fuzzy control (Feng et al. 2016; Wang et al. 2010; Lei, Wu, and She 2015), fuzzy PID control (Gao et al. 2006; Li et al. 2012a, 2012b), intelligent control (Li and Zhang 2016; Appari et al. 2015; Mozaffari et al. 2016), dynamic matrix control (Li, Zou, and Lu 2015; Wu, Zhang, and Lu 2014), expert system (Mitra and Gangadaran 2016; Liu and Liu 2014), and etc. Although the control effect of these methods has been improved compared with the conventional PID algorithm, there are also some disadvantages. The performance of fuzzy control and fuzzy PID is restricted by the fuzzy rules and membership function. Intelligent control algorithm is complex, which requires a lot of computing time. Therefore, intelligent control algorithm is too difficult to be used in the actual control application. Dynamic matrix control has very excellent robustness, but it needs a long computation time to solve inverse operation of the matrix. The expert system has no self-learning and adaptive ability. At the same time, the required numbers of rules are too large; the maintenance and control process is difficult. The expert system is not particularly applicable to the complex coke oven production process.