摘要
针对无人机采集飞行数据过程中存在数据野值,噪声以及丢失现象,论文分析了数据误差的表现种类及原因,提出了一种基于M估计稳健回归的算法,利用反复加权的最小二乘迭代法求解回归系数,实现对飞参数据的拟合,设置拟合数据与实测数据之间的阈值大小来识别野值。该方法大大改善了异常野值对回归系数的影响,克服了经典最小二乘对异常数据敏感的缺点。并在Matlab环境下对飞参记录的四边航行各通道数据进行了仿真,仿真结果验证了方法的有效性。
In view of the data outlier, noise and loss during the process of UAV data acquisition, this paper analyzes the types and causes of data error, and proposes an algorithm based on M-estimate robust regression, which uses repeated-weighted least squares iterations. The method solves the regression coefficients, realizes the fitting of the flight data, and sets the threshold value between the fitting data and the measured data to identify the outliers. This method greatly improves the influence of anomalous outliers on the regression coefficients and overcomes the shortcomings of classical least squares sensitivity to abnormal data. In the Matlab environment, the data of each channel recorded on the four sides of the flight parameters are simulated. The simulation results verify- the effectiveness of the method.
作者
王玉伟
高永
WANG Yuwei;GAO Yong(Naval Aeronautical University,Yantai 264001)
出处
《舰船电子工程》
2018年第11期38-41,共4页
Ship Electronic Engineering
关键词
数据预处理
经典最小二乘
稳健回归
data preprocessing
classical least squares
robust regression
