MEMS陀螺零偏温度补偿方法

Bias Temperature Compensation Method of MEMS Gyroscope

MEMS陀螺零偏是惯性导航系统的主要误差源之一,而外界环境温度变化对MEMS陀螺零偏具有重要影响。针对上述情况,提出基于小波阈值去噪与RBF神经网络相结合的零偏温度补偿方法。通过设计好的实验方案采集与温度相关的MEMS陀螺输出数据,并采用不同的温度补偿方法进行零偏温度补偿。实验结果表明,与原始输出、多项式拟合及RBF神经网络相比,基于去噪RBF神经网络的零偏温度补偿方法精度更高,适应性更强,有效地提高了MEMS陀螺输出精度,进而提高惯性导航系统精度。

The bias of the MEMS(Micro-Electro-Mechanical System) gyroscope is one of the main error sources of the inertial navigation system(INS), and the change of ambient temperature has an important influence on the bias of the MEMS gyroscope. In view of the above situation, a bias temperature compensation method based on wavelet threshold denoising and radial basis function(RBF) neural network was proposed. The temperature-related output data of the MEMS gyroscope was collected through the designed experimental scheme, and the bias temperature compensation of the obtained data was carried out by using different temperature compensation methods. The experimental results show that, compared with the original output, the polynomial fitting and the RBF neural network, the bias temperature compensation method based on wavelet threshold denoising and RBF neural network has higher accuracy and stronger adaptability, and effectively improves the output accuracy of the MEMS gyroscope, thereby improving the accuracy of the inertial navigation system.