乏信息多传感器压力数据自助模糊融合估计  

Pressure multi-sensor data fusion and estimation of poor information based on bootstrap-fuzzy method

作　　者：王中宇[1] 王倩[1] 付继华[2] 

机构地区：[1]北京航空航天大学仪器科学与光电工程学院,北京100191 [2]中国地震局地壳应力研究所,北京100085

出　　处：《北京航空航天大学学报》2013年第11期1426-1430,共5页Journal of Beijing University of Aeronautics and Astronautics

基　　金：技术基础科研资助项目(J132012C001);国家自然科学基金资助项目(50908215)

摘　　要：乏信息多传感器压力测量数据的融合估计是压力测量研究的重要问题,不同于经典的统计学方法,结合自助法和模糊数学的相关算法,提出一种实现乏信息多传感器压力测量数据融合估计的自助模糊数学模型.对具有乏信息特征的多个压力传感器的测量数据进行自助抽样;利用最大熵算法构建出不同时刻与位置的多个压力传感器测量数据的自助分布;用自助分布进行加权均值计算,提取相应特征值,得到自助融合序列;通过模糊隶属函数得到所测压力值的真值与区间估计.实例计算表明:在乏信息条件下,算法精度可达87%;在大样本条件下,测量数据在置信水平99.7%下,融合估计可靠性可达95%,验证了乏信息自助模糊融合估计算法的有效性.Pressure multi-sensor data fusion and estimation of poor information is a common problem in the field of pressure measurement. Small measurement data obtained from multi-sensor for the data fusion and estimation makes the data processing much difficult. Different from the statistical methods, a novel model based on bootstrap-fuzzy method for pressure muhi-sensor data fusion and estimation of poor information was presented. The pressure multi-sensor measurement data was processed by the bootstrap sampling. The boot- strap fusion sequence was derived from the bootstrap distribution. The true value and the interval of the pres- sure multi-sensor measurement data were estimated. Experimental results show the model has high accuracy and the data fusion sequence is in a good agreement with the original measurement data. The validity of the proposed method is examined.

关 键 词：乏信息 多压力传感器 自助模糊 数据融合 

分 类 号：C37[社会学]