摘要
基于信息融合中的数据关联技术,在卡尔曼滤波基础上,结合相互作用多传感器多模型的概率数据互联算法,建立故障监测报警和现场传感器量测数据关联二者之间的关系,建立更具一般性的分布式传感器系统基础上的多传感器多模型,改进概率数据关联方法,以用于故障监测报警中相互作用的算法.并结合一个时变系统中空间位置传感器的故障诊断问题为例,运用多传感器多模型相互作用的数据关联方法进行仿真分析,研究此类故障判据的数据关联问题和数据关联算法的改进,研究表明相互作用多模型的概率数据互联改进方法不仅与有限维数的特定测量阈值相对应,而且直接针对故障模式,能够体现出动态模型的优点,可以与系统诊断知识相融合,为故障诊断的单步的、多步的、长期的预测预报提供依据.
Based on the data association technique of information fusion and Kalman filter, and linking the probability data association algorithm of interaction multi-model, the relations between fault detection/ alarming and measuring data by local sensors are set data association, the multi-sensor and multi-model system is set up. The study indicates that the metrical threshold in limit dimensions, but also embody the advantage of dynamic model and lay forecast by merging the knowledge of systematic detection and alarming is also discussed. up. By improving the algorithm for the probability method for the more common distributing sensor improved method not only corresponds to the specific directs to the fault model. At the same time, it can the foundation for the single-step, multi-step and long diagnosis. The application of the method in the fault
出处
《江苏大学学报（自然科学版）》
EI
CAS
北大核心
2005年第5期444-447,共4页
Journal of Jiangsu University：Natural Science Edition
基金
江苏省科技厅科研基金资助项目(00kjt510007)
关键词
故障诊断
概率数据关联算法
多传感器多模型
卡尔曼滤波器
fault diagnosis
probability data association algorithm
multi-sensor and multi-model
Kalman filter
