Event-triggered adaptive neural asymptotic tracking control of autonomous underwater vehicles

This paper investigates the depth tracking control of autonomous underwater vehicles (AUVs) with unknown model dynamics and environmental disturbances, where the RBF neural networks (NNs) serve for approximation. To comply with the requirement of high-precision operation, the integral-bounded functions are constructed in the derived control laws with the minimum learning parameters (MLPs), such that the asymptotic tracking performance is attained in the nonlinear model. Besides, the “algebraic loop” problem is solved according to the intrinsic property of neural basis functions. To reduce the communication traffic, the event-triggered control (ETC) is designed in the controller-to-actuator channel, where a variable-threshold triggering condition is fabricated with the exemption from the “Zeno” behavior. All the tracking errors are proved to be asymptotic stable by using the Barbalat's lemma. The simulation corroborates the superiority of the proposed scheme.