基于EMD-TFPF的仿生偏振光罗盘去噪方法

为了提高仿生偏振光罗盘定向精度,降低罗盘在进行航向角测量时存在的高斯白噪声,基于经验模态分解(EMD)和时频峰值滤波(TFPF),设计了一种用于仿生偏振光罗盘的EMD-TFPF联合去噪方法.在去噪过程中,首先将含噪声的航向角信号分解为不同的模态,对不同模态采用不同窗长进行时频峰值滤波,进而改善了单一窗长去噪对有用信号造成的衰减,有效提高了去噪算法的自适应能力.经过机载实验验证,采用该方法去噪后的仿生偏振光罗盘可以实现定向精度0.3259°,比原始信号精度提升了18.4％.

Bionic Polarized Light Compass Denoising Method Based on EMD-TFPF

In order to improve the orientation accuracy of the biomimetic polarized light compass and reduce the Gaussian white noise in the heading angle measurement of the compass, based on the empirical mode decomposition(EMD) and time-frequency peak filter(TFPF), this paper designs a bionic polarized light compass EMD-TFPF joint denoising method. In the denoising process, the noisy heading angle signal is first decomposed into different modes, and different window lengths are used for time-frequency peak filtering for different modes, and then the attenuation of useful signals caused by single window length denoising is reduced, which effectively improves the adaptive ability of the denoising algorithm. The results of airborne experiments show that the bionic polarized light compass denoised by this method can achieve an orientation accuracy of 0.3259° and an accuracy increase of 18.4%.