Fuzzy logic approach to deadbeat control

A fuzzy logic approach to the ripple-free, finite settling time output control (or deadbeat control) of simple plants is presented. The approach is implemented by a fuzzy learning system which 'learns' the appropriate sequence of manipulated values required for dead beat control of the plant. It is demonstrated how this learnt manipulated value sequence, together with the corresponding output value sequence, can be put to use in three distinct ways: firstly, for the open-loop control of the plant; secondly, for the parameter estimation of a conventional deadbeat controller; and, thirdly, for plant parameter estimation. A simulation study on a second-order plant mode and an application case study with a pneumatic process were used to test and illustrate the fuzzy logic approach. Deadbeat control for minimum, and for one sampling period greater than minimum, output settling times was implemented with the fuzzy learning system. A comparative performance study of the fuzzy estimated deadbeat controllers with other conventional controllers is presented. The result of the conducted studies indicate the effectiveness and flexibility of the fuzzy logic approach to deadbeat control.