I’m currently working on detecting loss of control (LOC) on an overshoot of base to final. Using data models in X-Plane which depending on the aircraft model is known to be pretty accurate (there’s a certified commercial version suitable for flight training). Not practical to do this in an actual airplane for obvious reasons like crashing and fireballs.
Anyway, here’s a screenshot of data gathering. I’m collecting about 50 data points 10 times per second. I’m feeding this into a learning algorithm to teach it to predict the actions which may lead to an accelerated stall and/or spin.
Next step is to move the pitot/AOA probe to the other wing and see what difference this makes.
