具有输入饱和的航向离散非线性系统鲁棒后推设计

Robust Backstepping Design for the Ship Course Discrete-Time Nonlinear System with Input Saturation

本文研究了具有输入饱和的船舶航向离散非线性控制系统的鲁棒神经网络后推控制。首先,原系统通过变化得到一种新颖的能够预测变量船舶航向的离散非线性系统。然后利用后推技术进行该离散非线性系统的神经网络控制器设计。在控制器设计过程中,高阶神经网络用于逼近未知非线性方程。并且将输入饱和辅助系统引入存在输入饱和约束的船舶航向控制系统。和过去的控制方法相比,本文所提出的算法提高了航向离散非线性系统的鲁棒性。基于李雅普诺夫理论,闭环系统的所有的信号被证明半全局一致最终有界并且航向跟踪误差趋近于零。最后,运用“育鲲”轮仿真实例说明本文所提算法的有效性。

In this paper, a robust NN control scheme based on backstepping technique is proposed for the ship course discrete-time nonlinear systems with input saturation. Firstly, the original system is transformed into a novel ship course discrete-time nonlinear system which the state variable could be predicted. Then an adaptive NN controller design of the discrete-time nonlinear system is presented via backstepping. The HONNs are employed to approximate the unknown functions in the control design process. At the same time, an input saturation aided design system is introduced in the ship course controller in presence of input saturation constraint. Compared with previous research for discrete-time systems, the proposed algorithem improves the robustness of the systems. Based on Lyapunov Theory, the closed-loop systems are proven to be semi-globally uniformly ultimately bounded, and the tracking error converges to a small neighborhood of the origin. Finally, a simulation of vessel “yukun” is employed to illustrate the effectiveness of the proposed algorithem.