Improved composite neural learning control for marine unmanned vehicles with the actuator gain constraints

This paper presents an improved composite neural learning control algorithm for marine unmanned vehicles with the actuator gain constraints. The developed prediction error is constructed via the serial-parallel estimation model (SPEM) to improve the compensation effect of the neural networks (NNs) approximation. In the proposed scheme, system uncertainties are dealt with by employing NNs. The adaptive law based on the improved composite neural learning, which is designed to stabilize the related effect caused by the external time-varying disturbance and the actuator gain constraints. The improved composite neural learning control scheme achieves the semiglobal uniformly ultimately boundedness (SGUUB) of all the signals in the closed-loop system. Finally, a comparative experiment is illustrated to verify the superiority of the proposed algorithm.