NASA | Fuzzy Logic Models for Real-Time Simulations

NASA | Fuzzy Logic Models for Real-Time Simulations


We had the seedling fund, you know,
which we competed against three hundred and some other proposals and
were selected and teamed with the National Test Oilot School in California to
try and demonstrate the kind of a large global type nonlinear
aerodynamic modeling capability which we think the fuzzy logic algorithms can do. We picked a little jet that
can fly an unrestricted envelope and it was instrumented already… and got some data
with that and then try it out the fuzzy logic approach and it looked like the modeling
was very successful. And what the fund from the
Center sort of sponsored was to see if we
could go backwards from that. Now take the model, the mathematical model and
put it in a simulation and see if we can make it fly like the airplane. We developed those fuzzy logic models
from the flight test data and then went over to the differential maneuvering
simulator and tried to develop the equations of motion. You know, we didn’t spend a lot of time
and a lot of money to do it, but we got kinf of our quick look results
and ya, it does look airplane like. You know, you can fly
and spin…things like that. And there’s still some things to
work out but it certainly looks like it has promise to the at this point. What makes a special is the the
nonlinear modeling and we develop the flight test maneuvers as well
as the algorithms to analyze those, so it it makes flight testing much much
more efficient if this kind of a technique works…much cheaper. You can get
a high fidelity model over a large range of flight conditions very quickly which is very unusual for
standard flight test methods. There’s a lot of applications. One in flying,
if you want to fly an airplane and it has some problem, has a
failure of some sort, not only can it tell it has a problem but it can tell what
the problem is. You know, how it’s changed its dynamic capabilities. If you’re
a piloted airplane you can use that to warn the pilot and give him new cues of how
to fly the airplane. If it’s a UAV or an autonomous airplane, you know, it can
reconfigure in-flight itself in the best way it can to try to
survive the mission, that kind of thing. Additionally this kind of a
techniques helpin the development of an airplane, such like
wind tunnel testing. Next session, we have a computer that’s mounted in
the back seat so that we can get real-time maneuver guidance and also
be developing the models while we fly.

One thought on “NASA | Fuzzy Logic Models for Real-Time Simulations

  • March 3, 2019 at 9:03 am
    Permalink

    Link to the paper?

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *