摘要
针对单传感器测量数据处理问题,提出了一种基于最大熵的数据处理方法。首先利用最大熵方法的离散形式估计样本数据的概率分布,再根据概率分布计算样本的测量不确定度,从而确定出一个有效数据区间来进行粗大误差的判别,最后基于信息熵对有效数据进行融合,根据每个样本的不确定性来分配融合权系数,并以实例将该方法与其他方法进行了比较。可以较好地避免主观因素的影响,充分考虑了各测量数据的不确定性因素,提高测量的可靠性和精确度。
To single-sensor data processes,a data processes method based on maximum entropy was presented.First the sample probability distribution was estimated adopting maximum discrete entropy method,then the sample measure-ment uncertainty,which was calculated on the probability distribution,was used to establish an effective data range for elimination gross errors,at last the effective data were fused based on information entropy and the fusing coefficient was set according to each sample＇s uncertainty.An example compared the method with other methods.The method can avoid subjective factors＇ effect,takes the measurement data＇s uncertainty into account and improves reliability and accuracy during measurement.
出处
《电子测量与仪器学报》
CSCD
2012年第12期1096-1099,共4页
Journal of Electronic Measurement and Instrumentation
关键词
信息熵
最大熵
粗大误差
信息融合
不确定度
information entropy
maximum entropy
gross error
information fusion
uncertainty
