基于RBF神经网络模型和SVM模型的压力传感器温度补偿方法

RBF and SVM Based Temperature Compensation Algorithms for Pressure Sensor

压阻式压力传感器由于受环境温度影响会产生较大的温度漂移,测量精度明显降低,不利于其他依赖于压力数据的测控环节的工作。简要分析了当前常用的压力传感器温度补偿方法,然后采用RBF神经网络模型和支持向量机模型对压力传感器因温度变化所产生的非线性进行补偿。结果显示:传感器的温度漂移分别降低到0.6%F.S和0.5%F.S.,大大提高了压力传感器的性能和测量精度。最后,通过补偿结果的对比分析,讨论了两种算法的优劣性。

Piezoresistive pressure sensors have a temperature drift due to the influence of ambient temperature so that the measurement precision decrease obviously and it is conducive to the work of other monitoring aspects de- pending on the pressure data. The common pressure sensor temperature compensation methods is analyzed briefly, then RBF neural network model and support vector machine （SVM） model were used to compensate for the nonlin- ear due to temperature changes, the results showed that： the temperature drift of the sensor decreased to 0. 6% F.S. and 0.5 % F. S. , greatly improving the performance and measurement accuracy of the pressure sensor. At last, advantages and disadvantages of the two algorithms are discussed by the compensation results.