一种鲁棒的多旋翼姿态解算方法  

A Robust Algorithm for Multi-rotor UAV Attitude Measurement

作　　者：李忠威 敬忠良[1] 董鹏[1] LI Zhong-wei;JING Zhong-liang;DONG Peng(College of Aeronautics and Astronautics,Shanghai Jiao Tong University,Shanghai 201100,China)

机构地区：[1]上海交通大学航空航天学院,上海201100

出　　处：《计算机仿真》2020年第5期30-34,共5页Computer Simulation

基　　金：国家自然科学基金资助项目(61803260,61673262);上海市“科技创新行动计划”基础研究领域项目(16JC1401100)。

摘　　要：为解决多旋翼无人机姿态解算由于量测野值存在导致传统算法性能下降问题,充分考虑野值所导致的非高斯噪声分布特性,针对多旋翼无人机姿态解算模型,将量测噪声协方差矩阵近似为逆威沙特(Inverse-Wishart IW)分布,提出一种鲁棒的多旋翼姿态解算方法(Robust Attitude Measurement algorithm,RAMA)。基于加速度计、陀螺仪、磁力计相关数据,针对噪声存在野值的场景进行仿真验证。结果表明,当噪声存在野值时,提出的鲁棒姿态解算方法(RAMA)具有良好的鲁棒性,并且精度优于传统扩展卡尔曼滤波(Extended Kalman Filtering algorithm)EKF姿态解算方法。In order to solve the performance degradation problem of traditional algorithms for attitude determination of multi-rotor UAV due to the presence of measurement outliers, this paper proposes a robust attitude measurement algorithm(RAMA) method for the multi-rotor UAV attitude system, where the measurement noise covariance matrix was approximated to Inverse-Wishart IW distribution and the non-Gaussian noise distribution caused by outliers was fully considered. Based on the data of accelerometer, gyroscope and magnetometer, the simulation results show that the robust attitude measurement method(RAMA) proposed in this paper has good robustness, and the accuracy is better than the traditional Extended Kalman filter(EKF) when the noise has outliers.

关 键 词：姿态解算 非高斯噪声 多旋翼无人机 逆威沙特分布 

分 类 号：V301.6[航空宇航科学技术]