Mehdi Yaghoti - Department of Mechanical Engineering, Karaj Branch, Islamic Azad University, Karaj, Iran.
Abolfath Nikranjbar - Department of Mechanical Engineering, Karaj Branch, Islamic Azad University, Karaj, Iran.

The disturbed environment of the underwater turns out the design of appropriate controller for Autonomous Underwater Vehicle (AUV) more challenging and usually an accurate, intelligent and adjustable hybrid controller is highly demanded. In this paper, the compound Model Reference Adaptive Controller (MRAC) along with Artificial Neural Network (ANN) compensator of the AUV in x-y plane is illustrated. 2 Input 2 Output (2I2O) nonlinear dynamic system is linearized with inverse dynamic method and the MRAC controller is implemented. Stability of system demonstrated by Lyapunov theory. To increase the robustness of the MRAC controller in such agitated environment, a two layer ANN compensator benefiting online backpropagation learning algorithm to tune the weights and biases is augmented to the control system. The results of the separate simulations of the MRAC controlled system with and without ANN compensator in Matlab Simulink program clearly shows the performance of ANN compensated control method versus its non-ANN compensated counterpart in increasing the robustness and more accurate trajectory tracking performance of the control system subjected to the disturbances.

Autonomous Underwater Vehicle (AUV), Model Reference Adaptive Control (MRAC), Artificial Neural Network (ANN) compensator, Backpropagation algorithm, Lyapunov theory, Inverse dynamic linearization