论文部分内容阅读
神经网络具有同时逼近某一函数及其高阶导数的功能，这一结果为神经网络在非线性系统中的应用提供了可行的工具．本文提出了一种利用网络近似功能的非线性系统的近似线性化方法，无论系统是否满足可积条件，神经网络都可实现其对合条件的近似积分，从而构造满足系统近似线性化的反馈控制．对球－杆系统的仿真结果显示了这种方法的有效性．
Neural network has the function of approximating a certain function and its higher-order derivatives at the same time. This result provides a feasible tool for the application of neural network in nonlinear system. In this paper, an approximate linearization method of nonlinear system using network approximation function is proposed. The neural network can achieve approximate integration of the conditional approximation regardless of whether the system satisfies the integrable condition or not, so as to construct a feedback control that satisfies the approximate linearization of the system . The simulation results of the ball-rod system show the effectiveness of this method.