Is this project an undergraduate, graduate, or faculty project?
Graduate
Project Type
group
Campus
Daytona Beach
Authors' Class Standing
Chi Bok Lee, Graduate Student Andy Jung, Graduate Student Yanbing Chen, Assistant Professor
Lead Presenter's Name
Chi Bok Lee
Lead Presenter's College
DB College of Aviation
Faculty Mentor Name
Yanbing Chen
Abstract
The Federal Aviation Administration’s (FAA) “Bottle to Throttle” rule prohibits pilots from consuming alcohol within 8 hours of a flight and mandates a blood alcohol content (BAC) below 0.04%. However, pilots may still experience impairments beyond this timeframe. Current alcohol testing methods, including random screenings and confirmation tests, are limited in scope and timeliness, leaving potential gaps in aviation safety. This study aims to develop an AI facial recognition technique using Human-Centered Computing (HCC) to evaluate pilots’ readiness to fly by identifying subtle alcohol-related impairments, regardless of BAC levels. The project involves three steps: 1) training an AI model on publicly available datasets of facial images under varying alcohol conditions; 2) testing the algorithm on 20 non-pilot participants through cognitive tasks before and after alcohol consumption, including memory, reasoning, and spatial planning assessments; and 3) fine-tuning the algorithm using data from 20 pilots performing flight simulator tasks under similar conditions. This innovative approach has significant implications: AI facial recognition enables rapid, non-invasive alcohol screening for all pilots, enhancing safety compared to traditional random tests. It can serve as a preliminary screening method before BAC confirmation, ensuring comprehensive monitoring. Furthermore, the AI adapts to individual variations, addressing the limitations of the uniform 0.04% BAC threshold. By leveraging AI and facial recognition, this technology offers a more effective solution for ensuring pilot readiness and aviation safety.
Did this research project receive funding support (Spark, SURF, Research Abroad, Student Internal Grants, Collaborative, Climbing, or Ignite Grants) from the Office of Undergraduate Research?
No
Pilots’ readiness to fly - alcohol detection assisted by AI Facial Recognition Technology
The Federal Aviation Administration’s (FAA) “Bottle to Throttle” rule prohibits pilots from consuming alcohol within 8 hours of a flight and mandates a blood alcohol content (BAC) below 0.04%. However, pilots may still experience impairments beyond this timeframe. Current alcohol testing methods, including random screenings and confirmation tests, are limited in scope and timeliness, leaving potential gaps in aviation safety. This study aims to develop an AI facial recognition technique using Human-Centered Computing (HCC) to evaluate pilots’ readiness to fly by identifying subtle alcohol-related impairments, regardless of BAC levels. The project involves three steps: 1) training an AI model on publicly available datasets of facial images under varying alcohol conditions; 2) testing the algorithm on 20 non-pilot participants through cognitive tasks before and after alcohol consumption, including memory, reasoning, and spatial planning assessments; and 3) fine-tuning the algorithm using data from 20 pilots performing flight simulator tasks under similar conditions. This innovative approach has significant implications: AI facial recognition enables rapid, non-invasive alcohol screening for all pilots, enhancing safety compared to traditional random tests. It can serve as a preliminary screening method before BAC confirmation, ensuring comprehensive monitoring. Furthermore, the AI adapts to individual variations, addressing the limitations of the uniform 0.04% BAC threshold. By leveraging AI and facial recognition, this technology offers a more effective solution for ensuring pilot readiness and aviation safety.