神经网络在传感器非线性参数识别中的应用

The Application of Neural Network in Nonlinear Parameter Identification of Sensors

传感器的动态特性模型一般可以表示为一个线性环节与一个非线性环节的级联形式。利用神经网络的自适应性、任意非线性函数逼近的优点,以多项式函数代替常用的S-函数,运用改进的BP算法可以辨识出传感器动态特性模型的非线性参数。仿真结果表明了这一方法的可行性。

The dynamic model of sensor is generally denoted as the connection of one linear system and one nonlinear system. Making use of some advantages for self-adaptation of neural network (NN) and optional nonlinear function approximation, normal S-function is replaced with P-function, nonlinear parameter of the dynamic model of sensor is identified with the improved BP-Algorithm. The simulated result shows that this method is available.