A Totally Stable Adaptive Control for Path Tracking of Time-Varying Autonomous Underwater Vehicles

This paper deals with the problem of adaptive path tracking of autonomous underwater vehicles with time-varying dynamics. The controller design is based on a speedgradient adaptive law. A high-performance control behavior is aimed, so the full actuator dynamics is considered together with that of the vehicle. To this end, a state/disturbance observer is developed in the state feedback employing inverse dynamics. It is proved that the error paths can converge asymptotically to null when only the nonlinear static characteristic of the thrusters is involved in the design. When the actuator dynamics is considered too, only attractivity of the error paths to a residual set can be stated. The framework for this last proof relies on the concept of total stability. One main characteristic of our approach is that it can cope with a wide variety of bounded time-varying parameters with no limitations at all on their rates or a-priori knowledge. Copyright c ° 2007 IFAC