An Anomaly Detection Method for Spacecraft Using Relevance Vector Learning

Abstract

Abstract. This paper proposes a novel anomaly detection system for spacecrafts based on data mining techniques. It constructs a nonlinear probabilistic model w.r.t. behavior of a spacecraft by applying the relevance vector regression and autoregression to massive telemetry data, and then monitors the on-line telemetry data using the model and detects anomalies. A major advantage over conventional anomaly detection methods is that this approach requires little a priori knowledge on the system.

Granger Causality for Time-Series Anomaly Detection

“…For effectiveness, we compare our method with six state-of-art baseline methods, including PCA projection with T 2 statistic and Q statistic [8], dynamic PCA projection with T 2 statistic and Q statistic [8], relevance vector learning method [9], and modified distance measure [6]. For efficiency, we investigate five different lasso-type solvers as introduced in Section III for their suitability in our anomaly detection framework.…”

Finding Outliers in Satellite Patterns by Learning Pattern Identities

“…The propulsion subsystem henceforth provide a better second sight for validation. Eventually, despite several anomaly detection and curve fitting techniques exist such as (Li et al, 2010) and (Fujimaki et al, 2005) none of them actually plainly address the problematic.…”

“…However, due to the large quantity of information, only a small part of the data is being processed or used to perform anomaly prediction. A common accepted research concept for anomaly prediction as described in literature yields on using projections, based on probabilities, estimated on learned patterns from the past (Fujimaki et al, 2005) and data mining methods to enhance the conventional diagnosis approach (Li et al, 2010). Most of them conclude on the need to build a status vector.…”

An Anomaly Detection Method for Spacecraft Using ICA Technology

“…It is no doubt that the development of advanced anomaly detection and fault diagnosis technologies also plays an important role for this purpose. Actually, several attempts to develop such advanced anomaly detection and fault diagnosis methods by applying various information technologies and artificial intelligence have been made [3] [4]. Conventionally, anomaly detection methods based on prior expert knowledge and deductive reasoning process, such as expert systems and model-based reasoning, have been principally studied for this purpose [4], Though these knowledge-intensive approaches have been proved to be much better than the classical limit checking method, they are still costly and time-consuming to prepare the required knowledge or model.…”

“…Actually, several attempts to develop such advanced anomaly detection and fault diagnosis methods by applying various information technologies and artificial intelligence have been made [3] [4]. Conventionally, anomaly detection methods based on prior expert knowledge and deductive reasoning process, such as expert systems and model-based reasoning, have been principally studied for this purpose [4], Though these knowledge-intensive approaches have been proved to be much better than the classical limit checking method, they are still costly and time-consuming to prepare the required knowledge or model. On the other hand, in recent years, deductive reasoning techniques, such as data mining or machine learning technologies, have drawn much attention as alternative approach to the anomaly detection problems in various application fields [2].…”