LOADING...
Text Size
NASA Chat: Cheating the Weather to Improve On-Time Arrivals
September 9, 2010

[image-62]

[image-82]

[image-98]

We've all heard this announcement on an airplane before: "Uh, folks, we've got some weather around the airport right now. So I expect we'll be circling probably for the next half hour or so." Heather Arneson, a NASA Aeronautics Scholarship recipient and doctoral student at the University of Illinois, spent her summer at NASA working on ways to better control flows of aircraft into airspace impacted by weather.

"Weather conditions in recent years have caused approximately 65 percent of delays," says Arneson. "When weather is present near in an airspace, the number of flight allowed to occupy that airspace becomes lower than normal. Current methods of scheduling and routing flights that might fly through airspace predicted to experience bad weather can excessively delay flights. I'm seeing how mathematical optimization and modeling techniques can help us use that space more efficiently and reduce delays caused by weather."

Heather's solution can react in real-time as the weather and capacity situation changes. At NASA's Ames Research Center in California this summer, she simulated her new traffic strategy to see how well it worked.

On Thursday, September 9, at 3:00 p.m. EDT, Heather answered your questions about applying for the aeronautics scholarship, or even just about being an engineering student in today's competitive environment.

More About Heather Arneson
Heather is working on her doctorate in aerospace engineering from the University of Illinois at Urbana-Champaign. She has a masters in that field already and a bachelors degree in mechanical and aerospace engineering from Cornell University. Originally from Rhode Island, Heather also worked for several years as a member of NASA's Mars Exploration Rover Panoramic Camera Team, using her engineering skills to help the team acquire images of the Martian landscape.

More Information
View a Simulation of Air Traffic Using NASA's FACET → 

Smart Skies FlyBy Math Air Traffic Control Learning Experience → 

NASA Aeronautics Scholarship Program → 

Chat Transcript
(Moderator) Jason: Welcome to the "Cheating the Weather to Improve On-Time Arrivals" Chatroom. We're chatting with NASA intern Heather Arneson about her ideas for getting aircraft through areas of bad weather. This is a moderated chat. To submit your own question, please type it in the box at the bottom of the window and click the 'Ask' button on the right side of the box. Thanks for your patience as we answer your questions.

Kheervani: why are airplanes delayed during bad weather? May be the commercial airplanes have a reason but why is it so in case of a spacecraft?

Heather: I only deal with commercial flights. When there's a thunderstorm in a certain region of the airspace, it's unsafe. So the aircraft has to be routed around the storm. That means the number of aircraft that can fly through a certain region is lower than usual. That means aircraft have to be delayed either in the air or on the ground by giving departure delays.

(Moderator) Jason: We're working on answering the first few questions. To ask your own, please type it in the box at the bottom of the window and click the 'Ask' button on the right side of the box. Thanks for your patience as we answer your questions.

Kheervani: could you explain the significance of your diagram

Heather: Are you talking about the black and white diagram toward the bottom of the page?

Kheervani: yes

Heather: That's a diagram illustrating how our control algorithm works. The squares represent regions of airspace and the arrows represent how aircraft fly between those regions. The looping circles are our control input that delays the outflow of each section. The circle labeled "A" is a departure airport; the circle labeled "B" is the landing airport. This is a very simple case that we use to demonstrate our control theory. In reality, if you were actually looking at actual regions of airspace above the United States, it'd be a lot more complicated.

(Moderator) Jason: Heather is a NASA Aeronautics Scholarship recipient. If you're interested, the application period for Fall 2011 just opened. Go to http://nasa.asee.org.

Pharmaspace: Are there criteria which permit a plane to cross bad weather zones ?

Heather: The pilots and air traffic controllers get weather information; how high the clouds are, level of moisture in the clouds. From that the pilots will decide whether it's safe to fly through that type of weather. There's a gray region where some pilots will fly through and some won't, but there is generally a cut-off where pilots definitely won't fly through due to safety.

Bean: Predicting the weather has always been difficult; how much of the modelling is done based on the predicted forecast, probabilities of occurrences, etc? How far in the future of weather does it take into account?

Heather: Very good question. That is the major difficulty in scheduling flights going through a bad weather region; the accuracy of the weather prediction. Predictions are really only reliable for up to one to two hours. When flights are in the air for 4 or 5 hours or so crossing the country, you have to let aircraft take off before you really know what the weather will be like in a region for when the aircraft will get there. There're a lot of different methods that deal with the uncertainty of weather. The specific control algorithm I was working on uses no "look-ahead" time and will react in real-time to changes in weather. When you're reacting in real-time, the only thing you can do is add airborne delays to flights already in the air, which is expensive and difficult to implement. But we used that to try to match capacity constraints better when weather actually materializes. I used a two-part method: one method of control developed by researchers at NASA Ames with a longer look-ahead time but an inaccurate weather forecast. But then I'd run my control over that which would use the actual constraints in real-time and control the aircraft flow to meet those constraints.

(Moderator) Jason: We're working to answer your great questions. Keep them coming! To submit your own question, please type it in the box at the bottom of the window and click the 'Ask' button on the right side of the box. Thanks for your patience as we answer your questions.

Abi: Can you explain me about air traffic; I hear it is aeroplanes trying to land at same time. Is it true?

Heather: The work I'm doing is dealing with regions of airspace that have lower capacity than usual. We're trying to use that airspace more efficiently but we're not increasing capacity beyond the clear-weather-day traffic levels. We're not necessarily increasing traffic through a region of airspace beyond normal. There's a general effort to modify the entire air traffic control system, but, yes, that would have to look at overall capacity of airspace and airports to make sure all the timing works out well. That's a larger picture than what I was looking at.

(Moderator) Jason: Heather is a NASA Aeronautics Scholarship recipient. If you're interested, the application period for Fall 2011 just opened. Go to http://nasa.asee.org.

Charla: What type of model did you use and which weather variables did you include in the model?

Heather: I used what we call an "aggregate flow model," which models groups of aircraft flying through the airspace. As far as the capacity constraints due to bad weather, what I'm using is a measure of the reduced capacity. Eventually we'd like to use real weather predictions to generate those lower capacity values, but for my work this summer we used this approach to test my algorithm. We're not using actual weather forecasts yet.

simransohi01: how does it feel like working in NASA

Heather: It was a lot of fun working at NASA. There are a lot of people working on real problems, especially in the air traffic management area. You really see the direct applications of our work to those problems.

Mj: What have you enjoyed most about your internship at NASA Ames?

Heather: Having a chance to interact with the other researchers at Ames. On this particular project for this summer, I was working mostly on my own. I was trying to implement an algorithm that my PhD advisor and I had developed. I was using a software tool called FACET, which was developed at NASA Ames, to apply my algorithm to a realistic simulation of air traffic in the United States. It was nice to be at Ames and to hear what researchers are doing, what the newest research is in these different areas.

Mj: Do your simulations use real flight data from the past? Or how do you represent real flight conditions?

Heather: One of the great things about working at NASA Ames is I could get real flight data. That was the specific goal of my work this summer - to use real flight data in terms of departure times, routes, etc. I wanted to get a realistic traffic demand to use my algorithm with. I did choose data from a specific time period but that was because I wanted to be able to compare it with results another researcher reported.

(Moderator) Jason: We're working to answer your great questions. Keep them coming! To submit your own question, please type it in the box at the bottom of the window and click the 'Ask' button on the right side of the box. Thanks for your patience as we answer your questions.

Jochen: Hi Heather. On what exactly were you working on at the Mars rover project? Thank you.

Heather: I was a member of the team managing the color cameras on the rovers. Before the rovers launched, I helped with camera calibration for the Pancam and all the other cameras. Once the rovers landed, I wrote sequences to take images with the Pancams. A Pancam is the panoramic camera - the color camera on the rovers.

Kheervani: cant a weather predicting technology be built inflight which can help the pilots know the weather in the forthcoming region and thus providing them instantaneous solution. What i mean is cant there be a computer based solution to the problems rather than manually think how to handle the situation?

Heather: We are trying to come up with those computer solutions - to get the information about the weather and use it to plan routes, to schedule delays. We're working to come up with algorithms that will be used by a computer to generate the solutions. But first humans have to develop those algorithms so that computers can solve the problem. As to where the computers are, one thing that might happen is there'd be better computers on board planes, but first we have to come up with the solutions themselves that computers can use.

(Moderator) Jason: Do you have a question you've been waiting to ask? Go for it! To submit your own question, please type it in the box at the bottom of the window and click the 'Ask' button on the right side of the box. Thanks for your patience as we answer your questions.

Jason: What's the simulation on the computer screen in your photo?

Heather: That's a screen shot from FACET, which is the software that simulates air traffic in the U.S. Below that, the blue lines, is another screen grab from FACET that shows air traffic coming in to Chicago O'Hare Airport. The problem I looked at this summer was scheduling flights going into O'Hare. Over a 3-hour period there are roughly 200 aircraft coming in from the east. In current operations, delays are added as departure delays or airborne delays. We're trying to improve the throughput over current operations. We want to use the airspace to its fullest capacity, in real-time, rather than have excess delay.

Charla: As your study progresses, will the real time weather data only be used to generate the reduces capacity, or will aspects (temperature, wind speed, cloud density, precipiation intensity, etc) be included in your algorithm?

Heather: My algorithm is really just going to take in as an input the reduced capacity data. It's a really difficult problem to try to translate weather predictions into airspace capacity predictions. I'm not going to be getting into that field of research. That is a current topic of research going on at Ames and other research labs.

Jason: How well did the algorithm work?? Were you able to increase the throughput?

Heather: I focused on the software challenges of getting my algorithm working with FACET. I did find that my algorithm did help to match more closely the constraints that were input. For the future and what I'll be doing over the next several months is using more realistic constraints and applying my algorithm with different capacity constraints to evaluate how well my algorithm performs. As an intern, for 10 weeks, I was basically setting myself up to more rigorously test the algorithm and compare it with others, which I'll now be doing. It's all a process that takes time!

Kheervani: what should be one good in to become an aeronautic engineer?

Heather: In high school I was very interested in math, science and physics. I did well and enjoyed them. That led me to wanting to apply to engineering departments for an undergraduate degree. I got a bachelor's in mechanical engineering; I used a lot of math and physics. Similarly, for graduate work in aerospace engineering, with a focus on control theory, again math was very important.

(Moderator) Jason: We've got time for just a few more questions...

Mj: Air traffic control seems like a very hard problem, are there any simpler systems that you use to test or develop your ideas?

Heather: In a way, yes. I started with a simpler problem like the one in the diagram that we discussed earlier to test the algorithm. That's what I was working on up until this past summer. Then I worked on the more complicated problem of working with real flight data and the actual airspace system. Yes, air traffic management is a complex issue, but creating solutions for it will have a huge benefit for everyone.

(Moderator) Jason: Thanks Heather for the great answers to everyone's questions. We appreciate you taking time out of your day to sit down with us. Heather is a NASA Aeronautics Scholarship recipient. If you're interested, the application period for Fall 2011 just opened. Go to http://nasa.asee.org. Our chat is over! Thanks for participating. A transcript will be available within 48 hours.
 

Image Token: 
[image-47]
Heather Arneson.
Heather Arneson.
Image Credit: 
NASA / Dominic Hart
Image Token: 
[image-62]
A screen capture from one of Heather's simulations.
A screen capture from one of Heather's simulations.
Image Credit: 
Heather Arneson
Image Token: 
[image-82]
Heather created this diagram to communicate her new solution for air traffic flow.
Heather created this diagram to communicate her new solution for air traffic flow.
Image Credit: 
Heather Arneson
Image Token: 
[image-98]
Page Last Updated: July 28th, 2013
Page Editor: Lillian Gipson