摘要
提出一种基于钟差预测的铯原子钟频率异常检测算法。将钟差预测不确定度的解析表达式作为理论依据,以预测误差作为检验统计量,对检验统计量做二元假设检验。在两种情况下,检验统计量都服从正态分布。通过约束虚警概率(PFA),查询正态分布的分布函数,得到对应的检测门限值。根据在异常情况下的检验统计量分布函数和检测门限值,推导给出了求解平均频率跳变幅度(Ya)对应的检测概率(PD)的表达式。仿真试验和实测数据试验都验证了理论分析的结论。在此基础上总结了提高检测概率的方法。
A cesium atomic clock frequency anomaly detection algorithm based on clock prediction is proposed.The theory is based on the analytical expression of clock prediction uncertainty.Prediction errors are used as the test statistics and a binary hypothesis test is adopted.The test statistics all follow standard normal distributions.By means of setting a false alarm probability(PFA)value and acquiring the distribution function of normal distributions,the detection threshold can be determined.According to the detection threshold and the distribution function of the test statistics when anomalies appear,the expression of the detection probabilities(PD)s with different mean frequency jump levels(Ya)s is derived.Simulations and experiments validate the theoretical analyses.Based on the analyses,the methods of improving the detection probability are summarized.
作者
伍贻威
WU Yiwei(Beijing Satellite Navigation Center, Beijing 100094, China)
出处
《测绘学报》
EI
CSCD
北大核心
2021年第1期52-60,共9页
Acta Geodaetica et Cartographica Sinica
关键词
原子钟
异常检测
钟差预测
检验统计量
检测门限
虚警概率
检测概率
atomic clock
anomaly detection
clock prediction
test statistics
detection threshold
false alarm probability
detection probability
