Adaptive control of a manipulator using artificial neural networks

In this paper, a hierarchical neurocontroller for manipulation of a robotic arm is presented. Specifically, two artificial neural network systems are considered. One neural system, at a higher hierarchical level, participates in the motion analysis, and the other, at the lower level, in the process of control emulation. At the higher level the neural system consists of four networks: a restricted coulomb energy network to delineate the robot arm workspace, two standard back propagation (BP) networks for coordinates transformation, and a fourth network which also uses BP and participates in the motion decision-making process by cooperating with other knowledge sources. The control emulation process is developed using a second neural system at a lower hierarchical level. This system provides the correct sequence of control actions. An example is presented to illustrate the capabilities of the developed architectures. The two neural systems working at different hierarchical levels provide the robot arm speed, adaptability, and computational efficiency which are desirable in robot schemes.