Posted on 2013/12/14
My research combines affective computing and soaring flight simulation.
Some details about my research: I’ll start with some background around soaring & sailplane/gliders:
Flying is my greatest passion. Although I am rated for land and sea powered aircraft, soaring flight (i.e. sailplanes) is my greatest passion. A sailplane (also referred to as a glider) is essentially an airplane without an engine that has been designed with a high glide ratio (i.e. glide as far as possible without losing altitude). Sailplanes are usually launched via aerotow (i.e. towed to an altitude of around 2000 feet) or winch launched (a giant motor and cable flinging you into the air). Modern sailplanes have a glide ration in the range of 45:1 to 55:1, meaning that for a 50 to 1 glide ratio, assuming still air (i.e. no lift or sink), you lose one foot of altitude for each 50 feet you travel forward.
Sailplane enthusiasts frequently fly in contests. The goal of the contests is to fly a course to designated waypoints as quickly as possible. This requires a lot of familiarity with optimal thermaling techniques (thermals are a body of rising air that allows a glider to gain altitude). The goal is to fly slowly in thermals to gain as much altitude as you can and then fly fast through still air or sink to the next way point. While in flight you are frequently working with a flight computer, which can be a bit of a distraction. Soaring has a lot of unique training aspects to it not found in other types of lying, such as the skills for aero towed launches and winch launches as well as thermaling, specialized landing techniques, and more.
So after that rather wordy description of soaring, what is it I want to research?
There really aren’t any great “off the shelf” sailplane simulators that provide a decent training experience for sailplanes (the only pseudo commercial option I am aware of was a prototype project that sold for a price in 6 figures). For the most part, sailplane pilots only train in real sailplanes, unlike powered planes which have relatively low cost FAA approved simulators available for basic VFR flight training and more advanced instrument flight training (IFR). My goal is twofold… Show the feasibility of a (relatively) low cost soaring simulator for contest pilots (and potentially new sailplane student pilots). However, I want more than just a “simple” run-of-the-mill simulator. I want to prove that you can build an adaptive simulator, and that is where “affective computing” comes in. I want the realism of the simulator to adapt to the affect(emotional state) and performance of the pilot (and perhaps their experience level). If the pilot is doing well, then the simulator should add realism and (i.e. harder to accomplish tasks). I am looking at measuring different things to measure pilot state, such as skin conductivity, force of grip on controls (like the joystick and rudder pedals), facial recognition, voice stress, heart rate, and oxygen saturation. I plan on prototyping ways to measure affect in both the simulator and in real world flights (in my own sailplane, and also getting other sailplane pilots to carry my gear on their sailplane flights, and also to use my simulator).
My hypothesis is that the adaptive simulator is feasible and will provide an improved training experience, and my extension, better pilot performance in the real world. I want to make soaring flight safer and more accessible to prospective pilots. I would also like to extend the research to hangliding based soaring later on. I myself was in a catastrophic hanglider accident that resulted in multiple surgeries and years of physical therapy before achieving an 80-90% recovery. More than one friend has been killed in aircraft or hanglider accidents, and I would like to do what I can to prevent further tragedies. I plan on continuing research in this area as post-doctoral work (my dissertation mentor thinks I could easily spend 5-10 years on some of the follow up research I have in mind).