Millions of passengers take to the sky every day and their entertainment is a key feature of modern-day air travel. From the first in-flight movie screened in 1921 at an exposition in Chicago, to single aisle coach style television systems in the late 1970s, today’s in-flight entertainment (IFE) systems offer each passenger with their own personal system. However, the need for a next generation IFE system with tailored multimedia recommendations and a personalized dashboard reflecting the needs of a passenger is the obvious step forward.

The collaborative research project CANARIA (Cloud-enabled Aircraft Network and ARtificial Intelligence-based data Analysis), funded by the German Aerospace Center (DLR) aims to create such an innovation IFE system based on a federated communication and edge-computing network architecture. CANARIA bring together state-of-the-art technologies in networked aircraft of the future, cloud computing, and cutting-edge methods for data processing using Artificial Intelligence (AI) and Machine Learning (ML).

The planned networking of hundreds of components in CANARIA, which includes several types of sensors, cabin systems such as IFE screens, portable passenger and crew devices, etc., poses new challenges in terms of cyber security, network management complexity and data management in terms of the volume of data made available. However, this also brings in new opportunities regarding the utilization of distributed network resources (storage, computing power) in the sense of a cloud platform that includes a new groundbreaking IFE system. Major innovations in core technological areas such as wireless connectivity of many cabin systems, among others for high-rate data transmission, including IFE, Software Defined Networking (SDN) and virtualized platforms are proposed to realize this secure and flexible CANARIA cloud platform.

The main contribution of Fraunhofer FOKUS in CANARIA is a cloud based federated recommendation system for existing IFE solutions. It captures the passengers’ interactions with the IFE, such as watching a movie to create personal and portable passenger profiles and stereotypes. While not sharing personal data, the usage of Federated Learning – a modern state-of-the-art approach for distributed machine learning – enables the collaborative learning of a global model. Furthermore, such a model can be shared along the hierarchy level of the federation such as airlines, airline alliances and content providers without disclosing secret business data. Portable profiles of each passenger would be used to pre-calculate desired content and to download personalized content to the airplane and either to their seat or to their personal device.

Thus, Fraunhofer FOKUS contributes centrally to the development of custom-fit AI/ML methods for AI-based aviation applications and lays a technical foundation for other CANARIA sub-systems. Furthermore, FOKUS contributes its expertise in the implementation of distributed data exchange and edge-based pre-processing system.

The system will use NEMI, a state-of-the-art container-based architecture, currently in development at Fraunhofer FOKUS. As exchanging data across distributed systems in a reliable, real-time, secure and privacy-aware manner can be challenging, with NEMI, our Network and Edge Data Management Interface, we address these issues by leveraging our more than 20 years of in-house expertise in network convergence, artificial intelligence, and data visualization. NEMI establishes an open data-sharing substrate by tightly integrating into edge and (5G/6G) communication environments to enable data sovereignty, extraction, exchange and analytics for AI driven use cases. Examples include future organic, self-optimizing 6G core networks, distributed real-time control-loops, federated data infrastructures, or GDPR compliant data firewalling.

The partners TriaGnoSys GmbH / Safran passenger Innovations, TU Braunschweig, TU Dresden, Cadami GmbH and Fraunhofer FOKUS work together in the CANARIA joint project.