Application of soft computing techniques to control of a flexible mechanical system

Robotic manipulators play an important role in industry. Nowadays most company needs more productivity with less energy consumption. For reasons like these flexible link manipulators in other words flexible link systems are used more widely in several applications. Although flexible systems have several advantages, they have some disadvantages. The most vital disadvantage is residual vibration at the end of motion. This vibration is undesirable in industry. So different type of control schemes have been applied to eliminate or reduce this residual vibration. In the content of this thesis four different control techniques are applied to the single link flexible manipulator. Among these controllers PID controller (see chapter 4) and command shaping controller (see chapter 5) are classical type of controllers and they are designed for comparison of classical controller with soft computing types. The control techniques are proportional-integral-derivative control, command shaping control, fuzzy logic control and neural network control. These four controllers are applied separately to the system and results are discussed. Proposed controllers are designed by considering by considering quick response and small overshoot. Key Words: flexible manipulators, PID control, command shaping control, fuzzy logic control, neural network control