Air traffic control is a high-stakes environment where even minor lapses in attention or judgment can lead to serious consequences. A global leader in air traffic services was looking for an advanced AI-based solution to assess and monitor the mental and physical fitness of their air traffic controllers. Their primary goal was to detect and mitigate risks related to fatigue, distraction, drowsiness, and substance impairment—factors that could compromise operational safety.
To address these sensitive and complex goals, the client required a solution that would be both technologically advanced and strictly compliant with data privacy and operator consent requirements.
DEEPEYES' AI-Powered Emotion and Gesture Recognition
DEEPEYES developed a comprehensive monitoring tool that uses facial micro-expressions, gestures, and other visual cues to assess operator readiness.
Key Features Developed in Step 1:
- Electronic Profile of the Operator (EPO): A composite model using video data, physiological reactions, and psychological test results to track operator fitness.
- Fatigue and Distraction Detection: Algorithms to detect current levels of fatigue and distraction, including predictive modeling for “time to fatigue threshold.”
- Engagement Monitoring: Contextual analysis of workload and attention levels based on gaze distribution and task focus.
- Operator Identification: Reliable recognition of individuals for assessing time spent at the console.
To address these sensitive and complex goals, the client required a solution that would be both technologically advanced and strictly compliant with data privacy and operator consent requirements.
Key Features Developed in Step 1:
- Electronic Profile of the Operator (EPO): A composite model using video data, physiological reactions, and psychological test results to track operator fitness.
- Fatigue and Distraction Detection: Algorithms to detect current levels of fatigue and distraction, including predictive modeling for “time to fatigue threshold.”
- Engagement Monitoring: Contextual analysis of workload and attention levels based on gaze distribution and task focus.
To address these sensitive and complex goals, the client required a solution that would be both technologically advanced and strictly compliant with data privacy and operator consent requirements.
Data Privacy and Security by Design:
- All operators gave informed consent to participate and share personal data.
- Video data remains in full client possession; no footage is transmitted outside their infrastructure.
- No personal data is ever processed via the cloud—all AI operations run locally using DEEPEYES’ edge-based technology.
- The system was developed and tested in a privacy-secure environment, fully aligned with data protection best practices.
Over a six-months initial phase, DEEPEYES deployed a dedicated team of scientists and developers to build the foundational AI algorithms. Working closely with the client team and conducting fieldwork at the client facility, DEEPEYES successfully tested early versions of the AI recognitions under quasi-real-life conditions.
Results: Building a Roadmap to a Market-Ready Solution
The collaboration led to a successful proof-of-concept, with the system demonstrating the ability to detect the nine essential behavioral and physiological markers identified by the client.
Immediate Benefits:
- A functioning prototype capable of recognizing key signs of mental and physical strain.
- An adaptable module that integrates seamlessly into existing training and operational environments.
- Real-time analysis without dependency on external servers or high cost infrastructure.
- Full compliance with privacy standards, ensuring operator trust and ethical use of data.
With the foundational architecture in place, both teams are aligned to begin Phase 2, focusing on refining the AI system, increasing accuracy through medical and psychological consulting, and moving toward full market readiness.
The partnership sets a new standard in operational safety through proactive monitoring of human factors. As the system evolves, it will enable supervisors to make data-informed decisions in real time, enhancing both safety and performance in air traffic control centers.
The goal is to deploy this solution across multiple client facilities and potentially extend it to digital control towers, international airports, and other high-responsibility environments where human vigilance is critical.
