基于小波分析的MEMS加速度计去噪优化算法

针对惯导系统中MEMS加速度计输出漂移误差较大的问题,本文在对噪声分析建模的基础上,提出了一种小波阈值去噪优化算法.利用Allan方差分析加速度计输出噪声特性,构造噪声模型,并设计了一种基于多尺度阈值函数的去噪方法.该方法有效克服了传统软、硬阈值函数的局限性,且在各尺度上选取不同的调节系数优化滤波性能,算法适应性强.仿真结果表明,该方法在信噪比、均方根误差和波形相似度方面均有明显改善,较之传统阈值去噪算法信噪比提高了5 dB,并能在一定程度上提高系统导航精度,100 m范围内相对误差降低2.69％.

An optimized MEMS accelerometer de-noising algorithm based on wavelet analysis

In view of the large drift noise of MEMS accelerometer in the inertial navigation system,a wavelet thresh-old de-noising optimization algorithm is proposed based on the analysis and modeling of noise. The noise model is constructed using Allan variance to analyze noise characteristics,and a de-noising method based on the multi-scale threshold function is presented finally. This method effectively overcomes the limitations of soft and hard threshold function,in addition,the performance of filtering is optimized by selecting different adjustment coefficients at each scale. The simulation results show that better SNR,RMSE,and similarity could be obtained:the SNR is improved by 5 dB compared with the traditional methods and the navigation relative error within 100 meters is reduced by 2. 69％,which gets higher precision to some extent.