A Bottom-Following Preview Controller for Autonomous Underwater Vehicles

Abstract

The paper addresses the problem of bottom-following for autonomous underwater vehicles by taking into account the terrain characteristics ahead of the vehicle, as measured by two echo sounders. The methodology used to solve this problem amounts to posing it as a discrete time path following control problem where a conveniently defined error state space model of the plant is augmented with bathymetric (i.e. depth) preview data. A piecewise affine parameter-dependent model representation is used to accurately describe the AUV linearized error dynamics for a pre-defined set of operating regions. For each region, the synthesis problem is stated as a state feedback H2 control problem for affine parameter-dependent systems and solved using Linear Matrix Inequalities (LMIs). The resulting nonlinear controller is implemented within the scope of gain-scheduled control theory using the D-methodology. Simulation results obtained with a full nonlinear model of the INFANTE AUV are presented and discussed.