故障预测是雷达健康管理系统的一项重要功能。在对雷达实施故障预测时,应先对性能参数的可预测性进行度量分析。首先,介绍了雷达性能监测传感器网络以及相适用的线性建模预测方法;然后,基于相空间重构理论和混沌特性量,提出一种雷达性能参数可预测性度量方法,并为确定有效时间提前量提供科学依据;最后,结合雷达信噪比参数监测序列进行实验分析,验证了所提出方法的有效性。

Fault prediction is a critical part in radar prognostics and health management. Before employing characteristic parameter monitoring methods to predict radar忆s fault, the predictability of the performance observation sequence should be measured. Firstly, the sensor network used to observe radar忆s performance was introduced as well as the suitable linear modeling and predicting algorithm. Then a method was proposed to measure the predictability of radar忆s performance parameters based on phase space reconstruction theory and chaotic characteristic parameters, which provided a scientific basis to find a valid time advance. Finally, the experimental analysis carried out to radar signal-to-noise ratio parameter observation sequence validated the proposed method.

TN957.51

“十三五”国防基础科研计划项目