A Constrained Predictive Controller for AUV and Computational Optimization Using Laguerre Functions in Unknown Environments

Abstract

In this paper, a predictive controller approach is proposed for depth and steer control of an Autonomous Underwater Vehicle (AUV). The predictive controller is an advanced control technique that performs control in form of online at any sampling time. AUV control has a lot of complexity due to the coupled nonlinear dynamics, parametric uncertainty and external disturbances due to underwater conditions. In addition, the AUV in this paper has constraints on actuators, which make its control more complicated. One of the challenges against implementing of predictive controller is their computational burden and the time consuming control operations at each time step. In this research, the Laguerre orthogonal functions are used for the predictive controller design to optimize and educe computational burden in time interval. The designed controller has several advantages such as being online and optimized, high accuracy, implementation capability, interaction with the constraints and robustness to disturbances. In order to demonstrate the efficiency of the method, the proposed controller is simulated for the AUV and the calculation time of the controllers with and without the Laguerre functions is compared with each other. Using Laguerre functions, the simulation results and their implementation on the board show the favorable efficiency and effectiveness of the proposed controller. Additionally, we have compared the proposed method with the LQR method. The obtained results confirm the superiority of various predictive controller methods.