Designing of neural network-based SoSMC for autonomous underwater vehicle: integrating hybrid optimization approach

The control of an autonomous underwater vehicle (AUV) is regarded as a difficult challenge, owing to the nonlinear and uncertain dynamics of the AUV. In this work, Optimized neural network (NN) is integrated with the “second-order sliding mode control (SoSMC) approach” for control of yaw angle in AUV. More particularly, the positive gain of SoSMC is predicted by an optimized NN model, where the training is performed by a novel Sea Lion Distance-based FireFly algorithm via tuning the optimal weights. At last, the supremacy of the adopted model is validated under various measures. Accordingly, the RMSE values accomplished by the proposed model is 40.94%, 1.39%, 0.69%, 0.69% and 0.41% better than existing models like “GW-SMC, FF-SoSMC, SLnO-SoSMC, POA-SoSMC and GW-SoSMC”, respectively, for set point 1.