Citation
Mohammad, Karimadini
(2006)
Development of a Crisp Fuzzy-Like Controller Using Formula-Based and Vectorized Approaches.
Masters thesis, Universiti Putra Malaysia.
Abstract
Simplifying of implementation of linear state feedback fuzzy controllers is investigated
through the thesis. One of the most important problems in fuzzy controller design is the
number of fuzzy subsets (membership functions) for each fuzzy input/output variable.
The number of fuzzy subsets and consequently the number of fuzzy rules should be big
enough to achieve good approximation of control surface and have a smooth and robust
control. However as the number of rules increases, the memory space, and program
cycle time and total project cost will also increase dramatically.
The thesis proposes crisp-fuzzy like controller derived by two novel approaches. The
first one which is formula based crisp fuzzy-like controller proves that the monotonic
fuzzy controller is similar to nonlinear saturated controller and then represents several
different controller formulas. The second controller namely vectorized crisp fuzzy -like
controller maps the fuzzy variables in a vectorial space and derives formula that has the
structure similar to PID controllers. The proposed controllers are inspired from fuzzy logic where they can express the control law semantically but they are absolutely crips.Consenquently the needed memory space is minimizes since the rule table has been replace with the formula. On the other fuzzy controllers have high computational complexity while the new controllers are very simple to design, tune and implment.some new performance indexes also are porposed to evaluate the performance and stability of different controllers. Several well-known industrial models are used for simulation and a dimmer circuit to control the bulb temperature,has been used as a case study. Both simulation and experimental results show that the crips - fuzzy like controllers have the same or in some cases better performance and stability compare with the conventional fuzzy logic controllers, with extra merits of lower memory space and cycle time.
Download File
Additional Metadata
Actions (login required)
![View Item View Item](/style/images/action_view.png) |
View Item |