Application of neural networks to robotic arm control systems

Application of neural networks to robotic arm control systems

ABSTRACT

A robotic arm is a complicated beast. It encompasses controlling a substantially bulky, inherently unstable and difficult to control object placed in an unpredictable and complicated three-dimensional environment. Besides, the arm has more than two degrees of freedom making the mathematics extremely difficult to visualize and understand. The fact that a robotic arm is expected to perform fairly well in differing environments demands adjustability and flexibility from the controlling software. In such a case, the approach, which comes to the rescue, is the ‘neural network’ approach. This seminar aims at detailing the use of neural networks in the control of a robotic arm.