基于注意力机制与1D-CNN-Bi-GRU的MEMS陀螺仪降噪方法研究OA
为降低MEMS陀螺仪输出信号噪声,提出一种基于注意力机制和 1D-CNN-Bi-GRU的MEMS陀螺仪误差补偿方法.首先通过 1D-CNN网络进行陀螺时序数据特征提取,然后采用Bi-GRU网络进行特征分析,最后利用注意力机制进行数据权重分配.实验结果表明,与对照方法相比,该文方法噪声平均值减少 57.03%和 47.8%,噪声方差减少 81.07%和 73.01%,表明该文方法具有一定优势,具有较好的误差补偿效果.
In order to reduce the output signal noise of MEMS gyroscope,an error compensation method of MEMS gyroscope based on attention mechanism and 1D-CNN-Bi-GRU is proposed.Firstly,the feature of gyro time series data is extracted by 1D-CNN network,then the feature is analyzed by Bi-GRU network,and finally the data weight is distributed by attention mechanism.The experimental results show that,compared with the control method,the average noise of this method is reduced by 57.03%and 47.8%,and the noise variance is reduced by 81.07%and 73.01%,indicating that this method has certain advantages and has a better error compensation effect.
王宁;武雷;强敏利;王凯;郭英超
陕西华燕航空仪表有限公司,西安 710100
计算机与自动化
MEMS陀螺仪降噪1D-CNNBi-GRU注意力机制
MEMS gyroscopenoise reduction1D-CNNBi-GRUattention mechanism
《科技创新与应用》 2024 (026)
37-41 / 5
评论