基于观测器的自适应神经网络Backstepping输出反馈控制

Observer–based, Adaptive Neural Networks Backstepping Output Feedback Control

针对一类单输入单输出状态不可测非线性系统,提出一种自适应神经网络bakstepping输出反馈控制方法。首先,用神经网络逼近非线性函数,然后设计神经网络自适应观测器估计系统的状态。其次,在backstepping设计框架下,设计了自适应输出反馈控制器。最终,证明了所提出的自适应神经网络控制方法能够保证系统所有信号有界的同时,跟踪误差趋近于原点的一个小邻域内。仿真结果进一步验证了所提出方法的有效性。

An adaptive neural networks output feedback control approach was developed for a class of SISO nonlinear systems in unmeasured states. Using neural networks to approximate the unknown nonlinear functions was done first, and then an adaptive neural networks observer was introduced for state estimation. Under the framework of the backstepping design, adaptive output feedback control was constructed recursively. It was proved that the proposed adaptive neural networks control approach guaranteed the semi-global boundedness property for all the signals and at the same time, steered the tracking error to a small neighbordhood of the origin. Simulation results verified the effectiveness of the proposed approach to still further extent.