Satellite Telemetry Anomaly Morphology Extraction Based on Normal Mode Gap Mining

It is difficult for conventional data mining methods to find out non-frequent and long-sequence satellite telemetry anomaly morphology, and the mining results are not compact and lack of time continuity. In this paper, an anomaly morphology extraction algorithm NPGM-AMEA for telemetry data based on normal pattern gap mining is proposed. Firstly, the satellite telemetry parameters are discretized by KT-Means algorithm, and the original telemetry parameters are converted into feature strings. Secondly, the feature pattern mining is carried out to obtain the feature pattern of telemetry normal sequence, that is, the frequent closure of feature string;Then, the gap between adjacent normal feature modes is located to obtain the location of the abnormal pattern; Finally, the abnormal feature string between the normal mode gaps are extracted and the abnormal curve shape of telemetry state is obtained by matching the abnormal feature string with the time series data. The correctness and reliability of the method are verified by the simulation experiment of abnormal shape mining of the telemetry parameters of a certain type of satellite.