Abstract
It is difficult for existing deep learning-based satellite on-orbit anomaly detection methods to define the residual-based detection threshold and identify false anomalies. To solve the above problems, this paper proposes both a detection threshold determination and dynamic correction method and a causality-based false anomaly identification and pruning method. We use the GRU (Gated Recurrent Unit) to model and predict the telemetry parameters to obtain the residual vector; determine and dynamically correct the threshold according to the prescribed false positive rate; propose an improved multivariate transfer entropy method to identify the causal relationships between the telemetry parameters; and, based on the causality, determine whether the detected parameter anomalies are false. Experiments show that the precision, recall, and F1-score of the method proposed in this paper are superior to the current typical method, and the false positive rate is significantly reduced, demonstrating the effectiveness of the proposed method.
Highlights
D OMESTIC and foreign aerospace practices show that regardless of how strict the measures that are taken in the development stage are, on-orbit satellite failures are unavoidable
The fixed threshold method is currently the most commonly used satellite anomaly detection method in engineering [37], but due to the large number of telemetry parameters and the influences of unknown factors such as the space environment, it is difficult to set an appropriate threshold for each parameter, and it is difficult to guarantee the rationality of the detection threshold [38]
In order to judge whether the difference in the DTE is significant, this paper proposes a mean statistical test method to judge whether the directional transfer entropy between two parameters is significant so as to judge whether there is a causal relationship between the two parameters and the direction of the causal relationship
Summary
D OMESTIC and foreign aerospace practices show that regardless of how strict the measures that are taken in the development stage are, on-orbit satellite failures are unavoidable. To solve the above problems of data-driven methods, this paper proposes a satellite on-orbit anomaly detection method based on a dynamic threshold and causality pruning. 3) This paper proposes an anomaly pruning method based on the causal relationships between telemetry parameters. 4) Combining the above methods, this paper proposes a satellite on-orbit anomaly detection method based on a dynamic threshold and causality pruning. In order to solve the above problems, this paper draws inspiration from multivariate transfer entropy theory proposed by Lizie et al [21] and proposes an improved multivariate transfer entropy method to identify the causal relationships between high-dimensional telemetry parameters
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
