The management of NOTAMs on the flight deck remains a challenge for flight crews, who must process large volumes of dense textual information to identify the notices relevant to their mission. Despite ongoing modernisation efforts, the NOTAM system still relies on a legacy format with limited support for automated filtering and prioritisation, which increases pilot workload and motivates the use of Natural Language Processing (NLP) techniques for information management. This paper presents therefore an NLP framework based on a shared pre-trained distilled BERT (Bidirectional Encoder Representations from Transformers) encoder with two task-specific prediction heads, designed to support NOTAM analysis at different stages of the flight. The first task classifies each NOTAM into one of 20 operational categories to support pre-flight briefing. The second task estimates the operational severity of a NOTAM under three emergency diversion scenarios to support in-flight decision-making within cockpit decision-support systems. To address the lack of annotated data and the imbalance across categories, a dataset of 3000 NOTAMs was built through a human-supervised Large Language Model (LLM) generation workflow seeded with real messages. All samples are annotated for both tasks according to written annotation rules. The proposed framework is based on a DistilBERT encoder which receives fine-tuned through Low-Rank Adapters (LoRA), achieves a macro F1-score of 0.967 on classification and R²; values between 0.82 and 0.87 across the three severity scenarios.

In the event of an earthquake, information on damage and the need for assistance is often unclear, as familiar communication and transport routes may be interrupted. DLR designed a technical solution to assess situations in the aftermath of disasters as fast and as automatically as possible by using aerial remote sensing. Current information from the damaged area is provided by a remote-controlled drone, DLR’s Dornier 228-101, which is equipped with an optical aerial camera system. The system tailored to disaster operations is able to record, process, and transmit aerial images to the ground in real time. Satellite data can be used additionally as a reference source for area-wide information. This setup was assessed for earthquake response in a realistic but secured trial environment to identify opportunities for further development. Complemented by AI image analysis, this solution shows benefits for information procurement during emergency response. Besides, 3D mapping derived from this campaign has been applied in parallel during a full-scale exercise to test new ways of cooperation between on-site response and remote-control centres during international assistance operations. Both attempts of applying a new solution in a realistic test scenario are presented and discussed in this chapter from a technology provider’s perspective.

The management of NOTAMs on the flight deck remains a challenge for flight crews, who must process large volumes of dense textual information to identify the notices relevant to their mission. Despite ongoing modernisation efforts, the NOTAM system still relies on a legacy format with limited support for automated filtering and prioritisation, which increases pilot workload and motivates the use of Natural Language Processing techniques for information management. This paper presents an NLP framework based on a shared pre-trained BERT encoder with two task-specific prediction heads, designed to support NOTAM analysis at different stages of the flight. The first task classifies each NOTAM into one of 20 operational categories to support pre-flight briefing. The second task estimates the operational severity of a NOTAM under three emergency diversion scenarios to support in-flight decision-making within cockpit decision-support systems. To address the lack of annotated data and the imbalance across categories, a dataset of 3,000 NOTAMs was built by combining expert annotation with synthetic samples generated under human supervision. The proposed framework, based on a DistilBERT encoder fine-tuned with LoRA, achieves a macro F1-score of 0.97 on classification and R² values at or above 0.80 across the three severity scenarios.