Neural network and disturbance observer-based practical trajectory tracking of unsymmetric underactuated AUV with disturbance and input saturation

For the trajectory tracking of unsymmetric underactuated autonomous underwater vehicle (AUV), a neural network (NN) and disturbance observer-based strategy is proposed. Disturbance and input saturation are considered in the dynamics of AUV. Diffeomorphism transformation is employed to obtain an equivalent system to the original unsymmetric system. To deal with the underactuation, an improved approach angle is proposed and an additional control is designed to stabilise the velocity error in the underactuated sway motion. To deal with the external disturbance, an observer with guaranteed convergence is incorporated into the dynamics controller. To deal with the input constraint, adaptive neural networks are designed to identify the errors induced by input saturation. To avoid the calculation of time derivatives of virtual velocities, command filters are employed. Numerical simulation is performed to verify the effectiveness of the proposed control strategy. Under the proposed controller, both straight line and curve trajectories can be tracked well.