基于函数链神经网络的传感器建模

A Sensor Modeling Based on Functional Link Neural Network

现代测控系统对传感器的精度和工作条件提出了很高的要求.为此,人们不得不采取一些中间补偿和修正措施,实现抗干扰、线性化,以提高传感器和系统的性能.讨论基于函数链神经网络(FLNN)的传感器建模新方法,其精度提高,结构简单、使用灵活、建模容易,易于实时硬件实现.两个算例说明网络的训练和非线性逼近方法,显示出网络的自适应能力、学习能力,基于FLNN的传感器模型可同时实现温度补偿和非线性校正.实际上,利用这种模型可以跟踪补偿环境改变引起的传感器特性的各种变化,在测控系统中具有良好的应用前景.

In the modern measurement and control system high demands are made on the accuracy and the working conditions of sensors. A new approach to sensor modeling based on a functional link neural network is discussed. Its structure offers less computational complexity. Two algorithm examples are given to illustrate the proposed method. The proposed inverse model approach compensates the effects of the nonlinearity and temperature. In fact , the flexibility of the technique automatically compensates any variation of the sensor response occurring due to change in environmental conditions. The approach for sensor modeling is found to be superior to the existing techniques. It has a potential future in the field of measurement and control.