A Dry Run is the critical bridge between scientific theory and operational constraints. Ahead of the upcoming Human-in-the-Loop trials, the DIALOG project team gathered to test interoperability and robustness under real-time conditions.
Testing Operational Readiness
During the sessions in Salon-de-Provence, we migrated our three core services from isolated lab environments to the SIMADES platform (developed by SINTEF and hosted at ONERA).
The primary focus was ensuring the entire system’s operational readiness. We verified the multi-partner architecture’s ability to support real-time data exchange, confirming that the full technical pipeline is stable, integrated, and functioning exactly as expected for the upcoming trials.
Synchronizing Multimodal Streams
The “anatomy” of the Teamwork Assistant we are building relies on three pillars: Pilot requests, Environment, and Controller state.
During the run, a major focus was Data Exchange and Temporal Alignment. We worked to ensure that multimodal inputs, such as a spike in heart rate or a specific voice command, are perfectly timestamped across all logs. This synchronization is mandatory to accurately associate a specific operator state to his or her task goals. We are fine-tuning this approach to create an AI that truly adapts to the needs of the operator.
Verifying the Feedback Loop
Before testing how the AI supports the operator, the system needs a solid communication baseline. We used these sessions to verify the full loop, from initial data ingestion to displaying suggestions on the controller’s interface. With the connectivity challenges successfully addressed, we can now proceed to the next step, focusing entirely on human factors and operational strategy.
From Data to Insight
For an applied research project, the ultimate product is not just data. It is actionable knowledge that brings us closer to the operational solutions of tomorrow. By resolving integration issues early, we ensure that the Operational Validation in March generates clean and reproducible results. This preparation clears the way for what matters most: allowing us to gather human insights from expert controllers and truly understand the dynamics of Human-AI Teaming.
➡️ We are sharing live updates and further technical details with our professional network. Follow the progress of the DIALOG project on LinkedIn.
