Modeling and Control of an Unmanned Underwater Vehicle

The Unmanned Underwater Vehicle (UUV) designed at the Mechanical Engineering Department of the University of Canterbury is in an early stage of development. With the design of the AUV completed, the primary focus now is to design control software. The control software has to be able to stabilize the vehicle at a desired position and let the vehicle follow a desired trajectory within a reasonable error. The AUV is able to operate in six degrees of freedom and the dynamics of an AUV are nonlinear and subjected to a variety of disturbances. A kinematic and dynamic model of the AUV is derived for the six degrees of freedom operating range. The degrees of freedom are decoupled, where only the surge, heave and yaw degrees of freedom will be controlled. A system identification approach is proposed to estimate unknown mass/inertia and damping parameters, treating the surge, heave and yaw degrees of freedom separately. Unfortunately due to sensor malfunctioning the unknown mass/inertia and damping terms could not be estimated, a parameter selection method is used instead. The parameter selection is based on parameter estimation results of software programs and a parameter comparison with other similar shaped AUVs. A feedback linearizing control technique is chosen to design control laws for control in the surge, heave and yaw degrees of freedom. The controller performed well under parameter perturbation and noise contamination on the feedback position and velocity signals. An under-actuated control problem arises in x-y plane, since one is only able to control two degrees of freedom while the AUV has three degrees of freedom. To steer the AUV smoothly, a path planning model is derived, which describes a path from A to B with the use of polar coordinates. It works as expected, except for planning a straight line trajectory due to singularity reasons. v The under-actuated control problem is again investigated, assuming sway motion of the AUV is not negligible, so Coriolis and centripetal forces need to be considered. A state feedback control method is explained, which is able to follow a reference trajectory in the x-y plane. The only disadvantage is that the yaw degree of freedom of the reference trajectory needs to be persistently exciting, which makes following a straight line impossible.