As the volume of telemetry data generated by satellites and other complex systems continues to grow, there is a pressing need for more efficient and accurate anomaly detection methods. Current techniques often rely on human analysis and preset criteria, presenting several challenges including the necessity for expert interpretation and continual updates to match the dynamic mission environment.
This paper critically examines the use of Deep Anomaly Detection (DAD) methods in addressing these challenges, evaluating their efficacy on real-world spacecraft telemetry data. It exposes multiple flaws in current DAD research, highlighting the tendency for performance results to be overestimated and suggesting that simpler methods can sometimes outperform more complex DAD algorithms.
By comparing established metrics for anomaly detection with newly proposed ones, this paper aims to improve the evaluation of DAD algorithms. It underscores the importance of using less accuracy-inflating metrics and offers a comprehensive comparison of DAD methods on popular benchmark datasets and real-life satellite telemetry data.
Among the DAD methods examined, the LSTM algorithm demonstrates considerable promise. However, the paper also reveals the potential limitations of this approach, particularly in complex systems that lack a single, clear predictive failure channel.
The paper concludes with a series of recommendations for future research, including the adoption of best practices, the need for high-quality, pre-split datasets, and the investigation of other prediction error methods. Through these insights, this paper contributes to the improved understanding and application of DAD methods, ultimately enhancing the reliability and effectiveness of anomaly detection in real-world scenarios.

As the profit margins of the airline industry are relatively low, it is of utmost importance to keep costs low in order for airlines to stay competitive. An important cost factor is maintenance costs, as it can take up around 10-20 % of the total direct operational costs. Currently, much development is taking place in developing condition-based maintenance (CBM) strategies. These strategies on the one hand leverage remaining useful life (RUL) predictions of components to enable better planning and lower repair costs while on the other hand less preventive maintenance tasks are required due to the increase of useful sensor data available. This paper develops insights in the potential benefits that CBM can have as a function of different prognostic performance levels. This is done by developing cost- benefit models which accept a wide range of parameters being able to simulate prognostic effectiveness on different aircraft fleets. Results are obtained by using real MRO and operator input data. The results show that CBM can be beneficial, given that the model has a sufficient specificity and the component supply chain scales accordingly.