Design of an intelligent control system for remotely operated vehicles

The application of a neural network controller is described. Three learning algorithms for online implementation of the controller are discussed. These control schemes do not require any information about the system dynamics except an upper bound of the inertia terms. Selection of the number of layers in the neural network, the number of neurons in the hidden layer, initial weights for the network, and the critic coefficient was done based on the results of preliminary tests. The performances of the three learning algorithms were compared. The effectiveness of the neural net controller in handling time-varying parameters and random noise was tested by a case study on a remotely operated vehicle (ROV) system for robotic underwater operations. The results of the comparisons and the testing are presented in detail