基于旋转变分模态分解的IMU角速度去噪算法
IMU Angular Velocity Denoising Algorithm Based on Rotational Variational Mode Decomposition
覃舒娴 1覃晓兰 1刘运毅1
作者信息
摘要
惯性测量单元(IMU)的应用中,角速度的噪声误差积累对姿态解算性能有较大的影响.针对角速度中存在的噪声,提出一种融合了变分模态分解(VMD)和角速度旋转三维分解的去噪算法.首先通过坐标系旋转获得角速度在不同虚拟轴的输出,再利用 VMD 提取合适分量重构虚拟轴信号.VMD的非线性重构使得各个虚拟轴的残留误差相对独立,最终多个虚拟轴的反向旋转回到原始坐标系后通过独立信号的均值合并能有效消除IMU中角速度的噪声.基于EuRoC数据集的实验结果表明:该算法降噪效果显著,均方根误差降低70%～85%,且能有效平衡三轴误差.
Abstract
In the application of inertial measurement unit(IMU),the accumulation of angular velocity noise errors greatly af-fects the attitude calculation performance.Aiming at the noise existing in the angular velocity,a denoising algorithm was pro-posed,which combined variational mode decomposition(VMD)and angular velocity rotational three-dimensional decomposition.Firstly,the output of angular velocity in different virtual axes was obtained by rotation of coordinate system,and then the virtual axis signals were reconstructed by extracting appropriate components using VMD.The nonlinear reconstruction of VMD made the residual errors of each virtual axis relatively independent.Finally,after the reverse rotation of multiple virtual axes back to the original coordinate system,the mean value of independent signals can effectively eliminate the angular velocity noise in IMU.The experimental results based on EuRoC dataset show that the algorithm has a remarkable effect on noise reduction,root mean square error is reduced by 70%～85%,and the triaxial error can be effectively balanced.
关键词
变分模态分解/旋转重构/角速度去噪Key words
variational mode decomposition/rotational reconstruction/angular velocity denoising引用本文复制引用
基金项目
广西自然科学基金(2018GXNSFAA294121)
出版年
2024
NETL
NSTL
万方数据