الفهرس 
1. INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 16 |  |  |
| | | from | 16 | | |
Abstract
DC motors with flexible shafts are widely used as actuating elements in a variety of industrial manipulators due to their advantages. Practically DC motor with flexible shaft suffers from uncertainties such as resistance and inductance change with time that result from temperature change of the motor, load change and uncertainty in the measurements. Besides uncertainties, DC motors have a nonlinear behavior due to coulomb friction and dead-zone, significantly influence the system operation when the rotation of the DC motor changes the direction. So, the DC motor with flexible shaft needs control methodologies to overcome the uncertainties and the nonlinearity, the conventional controllers such as Proportional-Integral-Derivative and fuzzy controllers have been used which benefits from the operator’s knowledge of the system. But the choice of the controller parameters is important to satisfy the system performance requirements. So, in this thesis we apply adaptive controllers using reinforcement learning which adapt the gains of the PID controller and the scaling factors of the input/output membership functions of the fuzzy controller. The main advantages of using reinforcement learning for controlling the systems are that it acquires experience of the system during the operation and does not need the system model. So it gives high performance for nonlinear and time varying parameters of the system.
The modified Elman neural network based on actor-critic learning method has been implemented. The proposed controllers have been designed and implemented practically using Arduino Due. Practical results show good and significant improvement in the performance of the proposed controllers to respond the system uncertainties and nonlinearities.