A neural network based efficient leader-follower formation control approach for multiple autonomous underwater vehicles

This manuscript proposes an efficient and novel method for the leader-follower formation control of multiple autonomous underwater vehicles. The goal is to efficiently allow the follower to follow the desired formation that the leader has presented. It is noteworthy how the model-dependent and model-free control schemes are combined to achieve this goal. The dynamical model of the system always contains inherent uncertainties, so the model-based control technique is incapable to manage these systems. Due to uncertainties and external disturbances, it is impossible to have complete knowledge of the system's dynamic model in real-world applications. Therefore, whatever the partial information about the dynamic model is available has been used in the design of the controller. The approximation ability of the neural network is used to enhance the inefficiency of the model-based control strategy. The unknown dynamics of the system is estimated via a radial basis function neural network. An adaptive compensator is introduced to the controller's part to counteract the effects of neural network reconstruction errors as well as those of external disturbances. After that, a Lyapunov function makes use of the online adaptive laws for the parameter vector as well as for the weights of neural network which ensures that the system is stable. As a result, by Barbalat's lemma, formation errors not only stay inside the desired levels but also converge to a small neighbourhood of zero asymptotically. At long last, contrastive MATLAB simulations are conducted to affirm the capability and superiority of the suggested technique.