Ted Anderson ’22


Institute for Human & Machine Cognition, Pensacola, FL

This summer I set out to design a multi-page web application where users (researchers) could rapidly reconfigure human performance experiments at the Institute for Human & Machine Cognition (IHMC). My supervisor, Dr. Anil Raj, believed that in-person work interactions were essential for interns to understand IHMC’s technology, so at the end of May I drove 25 hours to Florida and began life in a humble rented room in the swamps just outside of Pensacola.

My IHMC shirt, a gift from Dr. Anil Raj.

The IHMC is in the process of redesigning the user interface of a proprietary software known as Adaptive Multi-agent Interface (AMI). AMI software acts as an information hub and computational clearing house for any number of separate research tools such as biometric sensors, augmented assist tools, and computational modules. Working in person allowed me to get a crash course on client server architecture as well as the biometrics involved with this use of AMI. Dr. Raj’s instincts were correct, as in retrospect I doubt I could have understood AMI without daily impromptu question and answer sessions.

While I was in the lab in Pensacola, I was able to observe AMI in action during one of Dr. Raj’s human performance studies on hypercapnia. Hypercapnia is when CO2 saturates one’s bloodstream, and can cause cognitive lapses, which even in very mild cases can be very dangerous for tasks with a high cognitive demand. Every week Marine fighter and helicopters pilots would come into the lab where engineers had built a mock F-16 cockpit. There, pilots would wear an oxygen mask and put on goggles, sometimes with a heads-up display, and enter a flight simulator and perform a standardized competence test. AMI allowed researchers to synchronize every biometric sensor or other research tools to the nanosecond, or to see the effects of hypercapnia on pilot performance in real time. Being in person helped me understand how researchers interacted with the AMI controller, the technical limits, and capabilities of each sensor it controls, and critically how information was organized and displayed for the researchers in the current version of AMI.

Critically, I learned how many layers you can add to a design process. For example, within two weeks I had a prototype of a multi-page application up and running, but its function lacked key features like time synchronization in the sensor controller and a log of past experiments. While I was learning JavaScript and trying to understand client server communication, I also had to balance the needs of good scientific practice in my design. This back-and-forth process was much easier in person because of my ready access to my supervisors while I was working.

I want to thank Ted McPherson ’67 for this incredible connection to IHMC, and the Class of 1974 and the ’68 Center for Career Exploration for their dedication and support—without them, this opportunity could not have been possible.