摘要
针对复杂检测环境下传感器获得的特征信息具有不确定性和模糊性等问题,提出了利用熵权灰色关联算法获得基本可信度赋值函数(BPAF).根据基于证据理论的信息融合方法,设计了单传感器多量测周期时域融合和多传感器空域融合的二级证据融合算法,采用基于可信度的判决方法作为故障检测和识别依据.熵权方法解决了灰色关联算法中特征权重的选取问题,二级证据融合算法提高复杂环境下识别结果的准确率.仿真结果表明,这种方法比一般的故障识别算法具有更高的识别率、更强的鲁棒性和更广的适用性,是复杂环境下故障模式识别的一种正确可行的新方法.
Aiming at the uncertain and fuzzy problem of sensor feature information in complex testing environment,it provided entropy-weight gray correlation to get the Basic Probability Assignment Function（BPAF）.Based on the information fusion method of evidence theory,it designed a two-degree evidence fusion algorithm,including time fusion with multiple measurement periods and space fusion with multiple sensors,and uses a decision-making method based on the basic reliability as the support of fault detection and recognition.The selection of features weight in gray correlation was solved by entropy-weight method,and the correct rate in complex environment was improved by two-degree evidence fusion algorithm.The simulation result indicated that it had better recognition rate,stronger robustness and more widespread application than the common fault recognition algorithm.So,it is a valid and feasible method for fault pattern recognition in complex environment.
出处
《应用基础与工程科学学报》
EI
CSCD
2011年第2期314-322,共9页
Journal of Basic Science and Engineering
基金
国家自然科学基金(60802059)
中央高校基本科研业务费专项资金资助项目(HEUCF100800111)
关键词
信息熵
灰色关联
信息融合
证据理论
故障模式识别
information entropy
gray correlation
information fusion
evidence theory
fault pattern recognition
