2. Introduction to the Principles of Feedback
As we have seen in Chapter 1, examples of dynamic systems with automatic controllers abound: advanced process controllers are operating in virtually every industrial domain; micro-controllers pervade an immense array of household and entertainment electronics; thermostats regulate temperatures in domestic- to industrial-sized ovens, and autopilots control aircraft.
Designing any one of these systems requires the close cooperation of experts from various disciplines.
To particularize the principal goal of control engineering within this team effort, it is helpful to distinguish between a system's tangible realization and its behavior. The aircraft's physical realization, for example, includes fuselage, wings, and ailerons. Its behavior, on the other hand, refers to the aircraft's dynamic response to a change in throttle, aileron, or flap position.
To control such a system automatically, one needs to interface the system to a controller, which will also have a physical realization and behavior. Depending on the application, the controller could be realized in a chip, analogue electronics, a PLC, or a computer. There also needs to be a channel by which the controller and system can interact via sensors and actuators: sensors to report the state of the system, actuators as a means for the controller to act on the system.
With this process and control infrastructure in place, the key remaining question pertains to the controller behavior. In the aircraft application, for example, if the controller (here called an autopilot) detects a deviation in speed, height or heading via the sensors, just how should it command throttle and ailerons to get back on target?
This is the control engineer's key concern, or, stated in general terms, the fundamental goal of control engineering is to find technically, environmentally, and economically feasible ways of acting on systems to control their outputs to desired values, thus ensuring a desired level of performance. As discussed earlier, finding a good solution to this question frequently requires an involvement in process design, actuator and sensor selection, mathematical analysis, and modeling.
The control engineer's perspective on the aircraft navigation example described above includes a cyclical dependency: autopilot commands affect the aircraft, whose changed speed, height, and heading in turn affect the further actions of the autopilot.
Such a cyclically dependent interaction between system behaviors is called feedback.
Feedback phenomena exist both in nature and technology. The periodic population growth and reduction in the famous predator-and-prey interactions are an example of feedback occurring in nature. The high-pitch whistling sound occurring as a result of interaction between microphones and loudspeakers in a concert hall is a technical example of feedback.
In both of these cases, neither of the two interacting systems can be clearly designated as controller or process--they are simply two systems interacting in feedback. Nevertheless, the feedback interaction is seen to have a profound impact on the behavior of the engaged systems.
This behavior-altering effect of feedback is a key mechanism that control engineers exploit deliberately to achieve the objective of acting on a system to ensure that the desired performance specifications are achieved.