Real-Time XML Data Processing in Aerospace Systems Using Event-Driven Microservices and Autoencoder-LSTM for Fault Detection

In the aerospace industry, the integration of engineering data with operational systems is essential for ensuring real-time decision-making, deep learning – based fault detection, and enhanced system reliability. Traditionally, this integration has been limited by the batch-oriented nature of data processing, rigid architectures, and latency in handling large volumes of XML-based sensor data generated during flight operations and maintenance. These limitations hinder timely fault detection and reduce the efficiency of condition-based maintenance strategies. To address these challenges, an event-driven microservices architecture is proposed for real-time XML data ingestion, transformation, and analysis in aerospace applications. This architecture decouples data producers and consumers, enabling scalable, asynchronous, and low-latency data flow between subsystems. Within this framework, an Autoencoder-Long Short-Term Memory (LSTM) model is implemented for fault detection, leveraging the model’s ability to learn normal patterns of system behavior and identify anomalies with high sensitivity. The system processes real-time XML data streams from aerospace components, transforming structured inputs into formats suitable for machine learning-based inference. Experiments were conducted using aircraft telemetry data containing operational logs with embedded faults. The Autoencoder-LSTM demonstrated high accuracy in detecting faults with minimal false positives, even under noisy conditions. Performance evaluation showed that the proposed system could handle continuous data streams efficiently, reducing latency by 40% compared to traditional monolithic systems. Additionally, fault detection precision exceeded 99%, indicating the suitability of the method for deployment in mission-critical aerospace environments. The combination of microservices architecture and deep learning-based anomaly detection bridges the gap between engineering data and real-time operations, paving the way for smarter, more responsive aerospace systems.