Envisioning the construction of navigation constellations for the next generation navigation systems combining with high,medium,and low orbits,with the sharp increase in the number and types of remote sensing parameters,traditional health eval-uation methods face the problem of insufficient coverage of past expert knowledge and fault mechanism reserves.To address this,a health evaluation method for navigation satellite payload subsystems based on Local Outlier Factor Detection and Bayesian Network Structure Learning is proposed.Through the collected data before and after actual fault occurrences in a satellite system,Experiments is designed and validated the ability of the Local Outlier Factor Detection method to accurately output the health status at a coarse-grained level.The efficiency and accuracy of the Bayesian Network Structure Learning were analyzed and compared using three sco-ring functions.Experimental results indicate that with scoring functions of BDeuScore,K2Score,and BicScore,the health assessment accuracy of the learned models on the system is 87.4％,80.5％,and 85.2％,respectively.it summarizes the shortages of the Local Outlier Factor Detection and Bayesian Network Structure Learning methods,and provides a new direction and idea for the health as-sessment of navigation satellite subsystems.
维普
万方数据