压力传感器温度漂移补偿的RBF网络模型

Temperature drift compensation for pressure sensor based on RBF network

压力传感器在实际应用中普遍存在着温度漂移现象,这降低了传感器的测量精度,因此需要采取适当的补偿措施对传感器的温度附加误差进行修正,从而提高测量的准确性。本文针对在压力传感器电路中采用温度硬件补偿措施成本较高且精度不高的情况,建立了RBF网络软件补偿模型。RBF网络具有良好的非线性映射能力和泛化能力,采用带遗忘因子的梯度下降算法进行RBF网络的参数调整,实验表明RBF算法学习速度快,精度高。对实验中采集的数据进行非线性补偿,取得了良好的效果,大大提高了压力传感器的性能和测量精度。

In practical application, temperature drift exists in pressure sensors, which decreases the measurement precision of the sensor. A proper compensation method must be adopted to adjust the additive error and improve the accuracy of measurement. Aiming at the problem that hardware compensation has the shortcomings of high cost and low precision,a software compensation model based-on RBF is suggested. RBF network has good nonlinear mapping and generalization abilities. A gradient descending algorithm with memory factor is applied to adjust the parameters of RBF network. Experimental results indicate that RBF algorithm has a fast learning rate and high precision. The performance and measurement accuracy of pressure sensors are improved greatly and satisfactory results are achieved.