基于BP神经网络的压力传感器误差补偿算法研究

摘要

A three layers' input-output model of diffused silicon pressure sensor is built using BP neural network, and then using the improved differential evolution algorithm to optimize the weights and thresholds of BP neural network in MATLAB simulation. Through the training, the compensated diffused silicon pressure sensor's output full-scale error can be achieved 0.035%. Theresults show that the BP neural network modeling based on the improved differential evolution algorithm is meaningful to improve the accuracy of the pressure sensor.%采用BP神经网络来建立扩散硅压力传感器的输出输入模型，其网络模型具有三层结构，采用改进型的差分进化算法来优化BP神经网络的权值和阀值，并在MATLAB中进行了仿真。经训练得到补偿后扩散硅压力传感器的输出满量程误差可达到0．035％，结果表明采用基于改进型差分进化算法的BP神经网络建模对提高智能差压传感器的测量准确度具有参考价值。