In this thesis, we discuss the modeling and control issues surrounding the development of a highly maneuverable autonomous underwater vehicle (AUV-HM1).
A numerical motion simulation system of an autonomous underwater vehicle is developed. In the equations of motions, the effects of trimming weight subsystem, deballast subsystem, control surfaces and main propulsion subsystem are included. The added mass terms are computed by the similar ellipsoids method, and the damping terms as well as the control surfaces effects are estimated by the data base “DATCOM”. The estimations of AUV-HM1 are compared with those obtained by Planar Motion Mechanism(PMM) testing system and Free Running Tests.
The control of AUVs has been a challenge to control engineers due to combined non-linear nature of both the vehicle itself and the environment in which they operate. The thesis presents an experimental research on the adaptive controller of an AUV test-bed in which the controller architecture is composed of multi-layered neural networks. The problem considered is that of designing a controller for an AUV to provide directional control. A Fiber Optic Gyro(FOG) is used to measure the yaw angle and yaw rate. Directional control performed by two thrusters in the horizontal plane. Weight adaptation of the neural network is achieved by minimize an objective function that is the weighted sum of tracking errors and control input rates. According to the experiments on various command trajectories, we show that when the learning process is kept active through the control operation, the neural network adapts to time-varying plant dynamics as well as disturbance upsets.