Adaptive sliding mode control for an autonomous surface vehicle subject to wind, waves, and currents

Abstract

Autonomous surface vehicles operate on aquatic surfaces such as lakes, rivers, and oceans, with the potential to assist numerous activities in different industries. For instance, operations from disaster management, environmental monitoring, oil spill mitigation, and urban mobility may benefit from this autonomous technology. Fully-autonomous capabilities are required to enable operations in unstructured environments without human intervention. Then, control systems are essential for autonomous behavior development. Furthermore, trajectory tracking is one of the most challenging control problems, particularly for underactuated vehicles. Moreover, the presence of environmental phenomena such as wind, waves, and currents increases the complexity of the problem. In addition, most unmanned surface vehicles are small in size and are more susceptible to the effects of their unpredictable environment. Consequently, robust controllers able to counteract such disturbances are relevant and are an issue tackled by this research work. Thus, this thesis addresses two control strategies based on adaptive sliding mode for robust trajectory tracking of underactuated unmanned surface vehicles subject to external disturbances and uncertainties. Both strategies can follow predefined time-varying trajectories in the presence of perturbations. The so-called hand position point is applied in both controllers to avoid a singularity common in underactuated surface vehicle control. Both schemes are robust against bounded perturbations with unknown bounds and attenuate the chattering effect without overestimating the control gains. The first proposed approach is a cascaded control scheme based on an adaptive integral terminal sliding mode. The integral terminal sliding surface improves the transient behavior, and it can be tuned to approximate the maximum convergence time. The second proposed approach is a direct control scheme based on adaptive nonsingular terminal sliding mode. The direct control scheme reduces the number of tuning parameters and simplifies the implementation since only one control law is necessary. The proposed approaches achieve practical finite-time convergence of the tracking error variables via the terminal sliding surfaces. Furthermore, the practical finite-time stability of the proposed guidance and control laws are proven based on finite-time and Lyapunov stability theories. Numerical simulations, using mathematical models from wind, waves, and water currents disturbances, analyze the advantages and characteristics of the proposed algorithms, whereas field experiments demonstrate the control performance in a physical platform subject to environmental disturbances and an external payload. The payload generates uncertainty by modifying the mass, moment of inertia, and hydrodynamic coefficients. Finally, the proposed control schemes are compared with existent and current techniques.