Applying Radial Basis Function Neural Network to Data Fusion for Temperature Compensation

In order to decrease the impact of environmental temperature on pressure transducer measurements with temperature compensation, a new method of data fusion based on radial basis function (RBF) neural network was proposed, at the same period, a practical test was carried out with the environmental temperature ranging from 10 to 60 degC and the pressure as 15 kPa. The results of the investigation showed that the relational curves between output voltage of the transducer and environmental temperature was horizontal after compensation, and the convergency of RBF neural network was faster than BP neural network, in addition, the maximum difference of the output voltage before compensation was 9.48 mv while it was 0.03 mv after compensation. The results of the present work implied that the objective of temperature compensation has been achieved essentially, furthermore, RBF neural network was better than BP neural network while used on temperature compensation to pressure transducers and the influence of temperature variation could be greatly reduced