一种湿度传感器温度补偿的融合算法

摘要

针对白动气象站上湿度传感器在实际应用过程中易受温度影响的问题,提出采用RBF神经网络与最小二乘相结合的融合算法实现湿度传感器的温度补偿.该方法将湿度传感器在温度影响下的特性曲线分为两个非线性段和一个线性段,并且自适应的确定线性段和非线性段,在线性段利用最小二乘方法拟合出直线方程,在非线性段利用RBF神经网络补偿温度产生的影响.仿真结果表明,这种方法简单易行,与一般的BP神经网络和最小二乘多项式方法相比,具有拟合训练速度快,补偿精度高的特点,可以有效用于湿度传感器的温度补偿,提高传感器的测量精度和可靠性.%According to the humidity sensors on the automatic weather station influenced easily by temperature in the actual application, RBF neural network and least squares combining fusion algorithm is proposed to realize compensation of the humidity sensor. The characteristic curve that the humidity sensor is under the influence of the temperature is divided into two parts,i. e. a non-linear part and a linear part,and in the adaptive determination of the linear segments and non-linear segments,the least squares method is used to fitting a straight line equation in linear segments, then RBF neural network is used to compensate the impact of temperature in non-linear segments. Simulation results show that compared with BP neural network and least squares poly, the method is easily implement, the speed of fitting training is faster,and makes the temperatures compensation for high accuracy. The temperature compensation of the humidity sensor can be effectively used to improve sensor measurement accuracy and reliability.