FEEDBACK CONTROL is a notoriously interdisciplinary field, being applied to an enormous variety of dynamic processes and objects. The disciplinary range is further extended by the use of digital control logic either in general purpose computers or in special purpose digital signal processors which have been recently introduced. As a consequence, the controls specialist is the quintessential systems engineer who must be able to converse with process engineers, specialists in dynamics and structures, and with computer systems programmers as well. As well, he must not only be able to converse but must understand the essentials of each of these fields to the extent that they influence the feedback performance of the final system.
Given this situation, it is small wonder that a great variety of textbooks have been written from which control can be taught. At the advanced level the author's task is relatively simple because an advanced point of view naturally leads to a narrow focus. However, at the introductory level, the pedagogical challenge is severe. One begins with the appreciation that the control engineer needs to understand the basic scientific principles underlying the physical devices being controlled and the environment in which the processes might operate. There are also usually many physical alternatives for actuators and sensors that must be selected to effect the design. The essentials of this understanding is to be able to write specifications that represent acceptable performance and are practical and achievable with the time and money available.
If one chooses to cover only the principles of feedback control as such, one must still consider at least the problems of modelling, dynamic response analysis, and controller design. The models may be linear or non-linear, lumped or distributed, continuous or discrete, constant or varying, single or multiple input and output and the signals may be random or deterministic. Analysis tools include transform techniques and matrix differential equations of state. If the controller is to be implemented by digital computer, to these must be added sampled-data models and "real-time" computer architecture and software. The approaches available for control design are also diverse. For about twenty years from 1960 to 1980 control theory was divided between "classical" concepts of Nyquist, Bode, and Evans and "modern" concepts typified by the work of Kalman, Bellman, and Bryson. The central concepts in the classical theory are sensitivity to model errors, stability margins, and disturbance rejection in SISO feedback loops. The tools of classical design are the Laplace transform (frequency response) and complex analysis. In the modern theory, optimal control of matrix differential equation models is central and key concepts are controllability, observability, the Kalman filter and the LQG problem. Limitations found * Digital Control System Design by Gene H. Hostetter.