Detecting Anomalies in Space using Multivariate Convolutional LSTM with Mixtures of Probabilistic PCA

Tariq, Shahroz; Lee, Sang Yup; Shin, Youjin; Lee, Myeong Shin; Jung, Okchul; Chung, Dae-Won; Woo, Simon S.

doi:10.1145/3292500.3330776

Cited by 73 publications

(29 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, most of the values remain constant for some period, e.g., for a few minutes or much longer. Therefore, we compress several timestamps (rows) into a single row, as analogous to [37], and achieve efficiency in terms of the data size (from 29.62 GB to 9.52 GB).…”

Section: Data Interpolation and Normalizationmentioning

confidence: 99%

“…This can lead to a high false positive rate due to the presence of minor glitches, which are not actual anomalies, but trivial outliers [41]. To address this issue, we have proposed a Convolutional LSTM (ConvLSTM) [40] and deployed the ConvLSTM-based anomaly detection module in 2019 [37]. In that model, we first trained the model with ConvLSTM and then found the error score between the real value and predicted value by the model.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Itad

Shin

Lee

Tariq

et al. 2020

Proceedings of the 29th ACM International Conference on Information &Amp; Knowledge Management

Self Cite

View full text Add to dashboard Cite

Reducing false positives while detecting anomalies is of growing importance for various industrial applications and mission-critical infrastructures, including satellite systems. Undesired false positives can be costly for such systems, bringing the operation to a halt for human experts to determine if the anomalies are true anomalies that need to be mitigated. Although rule-based or machine learning-based anomaly detection approaches have been studied, a tensor-based decomposition method has not been extensively explored. In this work, we introduce an Integrative Tensor-based Anomaly Detection (ITAD) framework to detect anomalies in a satellite system with the goal of minimizing false positives. We construct 3rd-order tensors with telemetry data collected from the Korea MultiPurpose Satellite-2 (KOMPSAT-2) and calculate the anomaly score using one of the component matrices obtained by applying CANDECOMP/PARAFAC decomposition to detect anomalies. Our result shows that our tensor-based approach outperforms existing methods, achieving higher accuracy and lower false positive rates. And we successfully deployed our anomaly detection system in real KOMPSAT-2 mission operation. CCS CONCEPTS • Computing methodologies → Factorization methods; Anomaly detection; • Applied computing → Aerospace.

show abstract

Section: Data Interpolation and Normalizationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Itad

Shin

Lee

Tariq

et al. 2020

Proceedings of the 29th ACM International Conference on Information &Amp; Knowledge Management

Self Cite

View full text Add to dashboard Cite

show abstract

“…e current common method used to detect faults in satellites is to compare telemetry parameters with preset thresholds directly [8,9]. is fault detection method is suitable for detecting abrupt and large faults.…”

Section: Introductionmentioning

confidence: 99%

A Method for Detecting Incipient Faults in Satellites Based on Dynamic Linear Discriminant Analysis

Zhang

Yang

et al. 2021

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

Timely detection and treatment of possible incipient faults in satellites will effectively reduce the damage and harm they could cause. Although much work has been done concerning fault detection problems, the related questions about satellite incipient faults are little addressed. In this paper, a new satellite incipient fault detection method was proposed by combining the ideas of deviation in unsupervised fault detection methods and classification in supervised fault detection methods. First, the proposed method uses dynamic linear discriminant analysis (LDA) to find an optimal projection vector that separates the in-orbit data from the normal historical data as much as possible. Second, under the assumption that the parameters obey a multidimensional Gaussian distribution, it applies the normal historical data and the optimal projection vector to build a normal model. Finally, it employs the noncentral F-distribution to test whether a fault has occurred. The proposed method was validated using a numerical simulation case and a real satellite fault case. The results show that the method proposed in this paper is more effective at detecting incipient faults than traditional methods.

show abstract

“…Hundman et al [15] proposed the use of LSTM and nonparametric thresholds for spacecraft anomaly detection and proposed an anomaly pruning method. [37,32] also used these meth-ods for anomaly pruning. However, the pruning methods were judged only from the perspective of the parameters themselves and did not consider the correlation between the telemetry parameters.…”

Section: Introductionmentioning

confidence: 99%

Satellite On-Orbit Anomaly Detection Method Based on a Dynamic Threshold and Causality Pruning

Chen

Jin

2021

IEEE Access

View full text Add to dashboard Cite

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 causalitybased 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.INDEX TERMS Dynamic threshold, causality pruning, anomaly detection, GRU, on-orbit satellite.

show abstract

Detecting Anomalies in Space using Multivariate Convolutional LSTM with Mixtures of Probabilistic PCA

Cited by 73 publications

References 25 publications

Itad

Itad

A Method for Detecting Incipient Faults in Satellites Based on Dynamic Linear Discriminant Analysis

Satellite On-Orbit Anomaly Detection Method Based on a Dynamic Threshold and Causality Pruning

Contact Info

Product

Resources

About