基于关联滤波算法的飞机任务段使用监控

Aircraft Usage Monitoring of Task Segments Based on Correlation Filtering Algorithm

针对飞机服役历程中复杂多变的使用环境和飞行任务,为了准确快速地获得其实际使用情况,提出了合成关联滤波算法。对于各实测数据任务段,建立对应的关联滤波器;并基于滤波结果和贝叶斯网络设计实现了一种飞机任务段推理决策方法。经实测数据验证表明,建立的任务段关联滤波器能够较好地识别出各飞行参数响应,采用任务段推理决策方法对测试样本的类别属性进行推断,结果的准确率达97.2%;可为飞机等大型机械结构的复杂使用情况监控提供新的技术途径,也可为其损伤/寿命监控提供必要的数据支持。

The combined correlation filtering algorithms were proposed for attaining the actual usage of aircraft accurately and quickly in complex and changing environments and missions. For each measured tasks segment, a corresponding correlation filter was established. Then an decision inference network for task segments was fulfilled based on filtering outcome and Bayes networks. By the validation of flight measured data, the results show that es- tablished correlation filters could quite identify the flight parameter responses, and a 97.2% rate of accuracy could be obtained via using the task decision inference network to infer class properties of task segments. The method presents a creative approach for complex usage monitoring of aircraft and other large-scale mechanical structures, but also provides the necessary data to support their damage and life monitoring.