提高测量可靠性的多传感器数据融合有偏估计方法  

Improving Measurement Reliability with Biased Estimation for Multi-sensor Data Fusion

作　　者：岳元龙[1] 左信[1] 罗雄麟[1] 

机构地区：[1]中国石油大学(北京)自动化系,北京102249

出　　处：《自动化学报》2014年第9期1843-1852,共10页Acta Automatica Sinica

基　　金：国家重点基础研究发展计划(973计划)(2012CB720500);国家自然科学基金(21006127;61104218);中国石油大学(北京)科研基金资助项目(YJRC-2013-12)资助~~

摘　　要：为了提高测量数据可靠性,多传感器数据融合在过程控制领域得到了广泛应用.本文基于有偏估计能够减小最小二乘无偏估计方差的思想,提出采用多传感器有偏估计数据融合改善测量数据可靠性的方法.首先,基于岭估计提出了有偏测量过程,并给出了测量数据可靠性定量表示方法,同时证明了有偏测量可靠度优于无偏测量可靠度.其次,提出了多传感器有偏估计数据融合方法,证明了现有集中式与分布式无偏估计数据融合之间的等价性.最后,证明了多传感器有偏估计数据融合收敛于无偏估计数据融合.实例应用验证了方法的有效性.The multi-sensor data fusion method is widely adopted in the process control community to improve the reliability of measured data. Based on the idea that biased estimation can reduce the variance of least square estimation, a multi-sensor biased estimation data fusion method is proposed to improve the measurement reliability. Firstly, biased measurement is presented based on ridge estimation, and a quantitative description method evaluating the reliability of the measurement data is proposed. The biased measurement reliability superior to the unbiased one is also verified. Secondly, a multi-sensor biased estimation data fusion scheme, which synthesizes the merits of the biased measurement and multi-sensor data fusion, is proposed, and the conclusion that is proven to decentralized unbiased estimation data fusion is equal to the centralized one is proved. Thirdly, the proposed fusion method is proven to converge to the unbiased estimation data fusion. Finally, a laboratory scale experiment illustrates that the presented method is effective.

关 键 词：测量 可靠性 数据融合 有偏估计 岭估计 

分 类 号：TP202[自动化与计算机技术—检测技术与自动化装置] TP212[自动化与计算机技术—控制科学与工程]