الفهرس Only 14 pages are availabe for public view
 |  | from | 252 |  |  |
| | | from | 252 | | |
Abstract
Conventional model based control system have been found to be Inefficient in handling some systems. Such systems are either too complex or too vague to be described by accurate mathematical models Some difficult issues related to these systems art inertia. momentum, 1rictioi, damping, oscillations. nonlinearities, and complex structure. dc Robot control is one of the industrial applications that has many such problems, which calls for more flexible strategies as adaptive fuzzy controller. An adaptive controller is a controller that can modify its behavior in Response to changes in the dynamics of the process. An adaptive fuzzy system is a fuzzy logic system equipped with a training algorithm, where the fuzzy logic system ix constructed from a collection of fuzzy rules, and the training algorithm adjusts the parameters of fuzzy system based on numerical input-output pans. The adaptive fuzzy control problem is then to find a method of adjusting the fuzzy rule-base parameters for the controller when the characteristics of hic process and its environment arc unknown or changing. In this thesis, some training algorithms for adaptive fuzzy system ore presented (the hack-propagation training algorithm. The orthogonal least squares training algorithm, the lookup-table training algorithm and the clustering-based training algorithm), so that the plant would track a desired reference trajectory. Two of the proposed adaptive fuzzy systems are successfully used as a controller for the robot arm trajectory trucking control problem They are used as function approximators for nonlinear and uncertain robot regression matrix.