Inspiration
Aviation accidents are often caused by human error, fatigue, or distraction. We wanted to create a system that helps pilots stay safe in real time, enhancing cockpit situational awareness and reducing risk.
What it does
Pilot-Safety monitors pilots using cockpit cameras and flight data, detects fatigue, inattention, or risky behavior, calculates a Pilot Safety Index (PSI), and triggers instant alerts to prevent incidents.
How we built it
We combined lightweight AI models for behavior and gesture detection, computer vision for facial and posture analysis, and real-time alert systems. Everything runs on-device for privacy and low latency.
Challenges we ran into
Processing video and sensor data in real time without lag. Detecting subtle signs of fatigue or distraction accurately. Balancing model accuracy with lightweight deployment on edge devices.
Accomplishments that we're proud of
Developed a real-time pilot monitoring system with a working Pilot Safety Index. Integrated dual-camera monitoring for multiple behavioral cues. Achieved offline operation without relying on cloud connectivity.
What we learned
Real-time AI requires careful optimization for speed and accuracy. Human behavioral monitoring has nuances that require multi-modal analysis. Edge deployment challenges teach the importance of efficient model design.
What's next for Flight Safety
Integrate more cockpit sensors for richer behavioral data. Expand to automated corrective suggestions or guidance. Test and validate in real flight scenarios for enhanced reliability.
Log in or sign up for Devpost to join the conversation.