一类非线性系统的自适应神经网络控制

针对一类具有非仿射函数和下三角结构的、受干扰未知的非线性系统，提出一种新的自适应神经网络控制方法．它是严格反馈不确定系统和纯反馈系统的更一般化表达．在Backstepping设计思想基础上，证明了闭环信号的半全局最终一致有界性，并很好地处理了控制方向和控制奇异问题．通过仿真验证了该方法的有效性．

Adaptive neural network control for a class of nonlinear systems

To a class of unknown perturbed nonlinear systems an adaptive neural network control scheme is presented. The systems with disturbances and non-affine unknown functions have low triangular structure that generalizes both strict-feedback uncertain systems and pure-feedback ones. Based on the idea of backstepping, the Semi-global uniformly ultimately boundedness of all the signals in the closed-loop is proved. The problems of control directions and control singularity are dealt with well. The effectiveness of proposed scheme is showed by a proper nonlinear system.