The Application Of Distance Sensor Nonlinear Distortion Correction Based On Artificial Neural Network

The tradition method of Sensor nonlinear distortion correction, such as hardware compensation or checking table, curve fitting etc, is complex and needed too much sample with low accuracy. Sensor nonlinear distortion correction based on artificial neural network (ANN) need fewer sample and simply worked. When one part of sensor is changed, the network is needed only to study again. The ANN method is one of important theory and technique to realize intelligence instrument.This dissertation introduces tradition method of sensor briefly for distortion correction. Through theory analyses it is proved that new method of sensor nonlinear distortion correction based on artificial neural network is useful. Here the BP algorithm is used. The paper describes in detail the calculation process and characteristics of the BP network algorithm, and analyses the need to pay attention to some issues of the application in the process of modeling, such as hidden nodes choice, the choice of permission error and so on.This dissertation selects the Hore sensor CS3503, a distance testing sensor, as an example, with nonlinear relationship between output and input of the sensor. A lot of distance and voltage date are collected. A MATLAB program is designed to train the BP network and realize sensor's nonlinear distortion correction finally. The experimental data show that designed network achieves good results to correct sensor nonlinear distortion. The method can also be used for other nonlinear distortion corrections, with better promotion prospects.