Anomaly Detection for Aviation Cyber-Physical System: Opportunities and Challenges

Anomaly detection in Aviation CPS is critical to ensuring safety and reliability. This systematic literature review explores the landscape of machine learning techniques used for anomaly detection in Aviation CPS, analyzing studies published between 2014 and 2024. The review identifies a strong preference for unsupervised learning methods, driven by the challenges of acquiring labeled data in aviation contexts. Additionally, it highlights the emerging trend of hybrid models that combine supervised and unsupervised techniques, offering improved detection accuracy and robustness. However, the review also reveals significant obstacles, such as the limited availability of publicly accessible datasets, which hampers research progress and the ability to benchmark models. Moreover, while accuracy is the most commonly reported performance metric, the need for a broader evaluation framework that includes precision, recall, and other metrics is emphasized. The findings suggest several directions for future research, including developing standardized datasets, optimizing hybrid models, and integrating explainable AI (XAI) to enhance model interpretability. This review contributes to the field by synthesizing current knowledge and providing insights that could guide the development of more effective and reliable anomaly detection systems for Aviation CPS.