Panasonic Avionics Corporation, which provides in-flight entertainment and connectivity (IFEC) systems across a large global fleet, partnered with AWS to develop an agentic AI system that accelerates diagnostics. The solution leverages Amazon Bedrock, Amazon SageMaker, and AWS Glue to improve operational efficiency and reduce manual effort in diagnosing system issues affecting millions of passengers. The system processes data through a multi-agent workflow, enabling faster detection and resolution of fleet-wide performance problems. This initiative aims to enhance diagnostic accuracy while allowing engineers to focus on strategic improvements rather than repetitive investigative tasks. Source: awsml
The collaboration between Panasonic Avionics and AWS addresses the challenges of analyzing vast amounts of operational data from thousands of unique IFEC configurations. Manual correlation of logs, metrics, and ticketing data across diverse fleet variants previously took hours, limiting the speed at which emerging patterns could be identified. By using a multi-agent architecture, the system enables real-time anomaly detection and automated diagnostic analysis. This approach reduces the mean time to detect (MTTD) and mean time to resolve (MTTR) issues, allowing for more proactive fleet health monitoring. The solution also incorporates a domain ontology to standardize terminology and create a unified view for cross-fleet diagnostics. Source: awsml
Panasonic Avionics’ solution processes operational data through five phases, starting with ingestion and normalization using AWS Glue and Amazon EMR. The data is stored in an Amazon S3 data lakehouse with Apache Iceberg, ensuring consistency across diverse configurations. The Trend Analyzer identifies anomalies by evaluating key performance indicators and service degradation metrics, while parallel diagnostic agents perform correlation analysis, system checks, and log pattern matching. A summarizer powered by a large language model integrates these findings into diagnostic reports with root cause analysis and recommended actions. This structured approach enables faster resolution of complex issues, improving both operational efficiency and engineering productivity. Source: awsml
