CAMBRIDGE, Mass. (AFNS) —
This week, the Department of Air Force-Massachusetts Institute of Technology’s Artificial Intelligence Accelerator will launch the CogPilot Data Challenge 2.0. The challenge invites participants to explore AI-based solutions linking a pilot’s physiology, such as heart rate, eye tracking, and muscle activity, to their behavior and performance in flight tasks. variable difficulty. The AI Accelerator welcomes all participants from the Department of Defense, academia and industry, and aims to accelerate innovation by engaging the broader AI community to meet the challenging technology needs of the DoD.
Since its creation in 2019, the AIA has published more than 10 challenges resulting from its various research projects.
“These challenges have become some of our greatest successes at the Artificial Intelligence Accelerator,” said Major Kyle McAlpin, Project Liaison for Performance Prediction and Tuning. “They have surfaced and nurtured organic machine learning talents in air and space forces, built vibrant communities of interdisciplinary researchers and operators pushing the state of the art, invited the public to join the resolution of some of our toughest problems, and given back to the machine learning community by funding and publishing large machine learning-ready public datasets. The machine learning community has seen time and time again that fundamental advances in ML start with strong concurrency over large, unique, and public datasets. »
The CogPilot Data Challenge 2.0, hosted by AIA’s Performance Prediction and Optimization Research Team, consists of two tasks. First, participants are challenged to develop a model that predicts the level of difficulty of an airplane landing performed in virtual reality based solely on the physiology of the pilot. For the second task, participants predict how well the pilot performed each approach and landing task using pilot physiology alone.
To collect the dataset, the team used an immersive virtual reality simulator to record many types of physiological measurements as the pilots made approaches and landings on a runway at four different difficulty levels. The scenario used a simulation T-6A Texas II piloted fixed-wing trainer with a stick, throttle and rudders. The CogPilot Data Challenge 2.0 is a follow-up to a previous challenge that had over 180 participants. CogPilot 2.0 provides additional data from 20 pilot participants as well as new rewards.
The Performance Prediction and Optimization research team investigates how quantitative performance measures and physiological monitoring can provide a more individualized and objective assessment of pilot training compared to current subjective and crude measures. In partnership with pilot training units, the team uses AI to integrate various physiological assessments into a single measure of cognitive workload. Cognitive workload measurement can be used to speed up debriefings by targeting parts of a flight or simulation that induced high cognitive workload. It can also be used by instructor pilots to adapt their in-flight instruction.
Registration for the CogPilot Data Challenge 2.0 will be open through January 2023 and challenge submissions are due February 15, 2023. Team results will be released March 7, 2023. To learn more and register, visit here.