Quadcopter Engine Failure Detection and Recovery Methods

Hello. I am very new to this community, and apologize if a similar discussion exists elsewhere.

I am working on a thesis relating to quadcopter engine failure detection and recovery methods. Essentially, in the event of motor failure, on-board algorithms will quickly detect that something is wrong, idenitfy the failed rotor, and change the flight control method. Then, the control algorithm will "safely" and autonomously land the vehicle.

The algorithms will be tested in a Simulink-based virtual test environment/simulation I have developed and then (time-permitting) on an inexpensive quadrotor flight system.

Before I get too far into the thesis, I wanted to verify that the need for such a solution exists. So, I have a few questions:

-Are there similar detection and control algorithms currently in use in the hobby community/MAV industry?

-Qualitatively speaking, how difficult is it to manually "safely" (causing minimal damage) land a quadcopter with a failed engine?

-How frequently do engine failures occur?

-What are factors affecting engine failures (BLDC motor quality, flight hrs, current draw, ESC failure, etc)?

-Anything else that I should consider?

Thanks in advance for any suggestions or feedback.

The autopilot would need to recognize it has a problem and take an action. I'm not ready to disclose the action yet but it's simple. I know the autopilot can figure that out if we want it to. the sooner the this action occurs the more likely the craft will not crash. a human couldn't recognize the issue and solve the problem using my approach but the autopilot could. it would limp to the ground on the three remaining engines using my trick.

timmy

ps. i do have a question though - maybe you know the answer. it isn't my solution but it's related. if the QR were to lose an engine, roll over 360 degrees and become upright again, would the autopilot be able to resume controlled flight if the engine had restarted and was again available? my impression from what i have seen and read is that it could...

Replies

Marc January 12, 2014 at 8:32am

So there wouldn't be any way to reverse thrust on a prop would there? If so, the solution is extremely simple for a quad and will probably even give you ~70% flight capacity as well. I haven't done hobbies like this in a while...
Dennis Okeefe September 24, 2013 at 4:56pm

The autopilot would need to recognize it has a problem and take an action. I'm not ready to disclose the action yet but it's simple. I know the autopilot can figure that out if we want it to. the sooner the this action occurs the more likely the craft will not crash. a human couldn't recognize the issue and solve the problem using my approach but the autopilot could. it would limp to the ground on the three remaining engines using my trick.

timmy

ps. i do have a question though - maybe you know the answer. it isn't my solution but it's related. if the QR were to lose an engine, roll over 360 degrees and become upright again, would the autopilot be able to resume controlled flight if the engine had restarted and was again available? my impression from what i have seen and read is that it could...
Dennis Okeefe September 24, 2013 at 1:15pm

it doesn't have to crash. i know how to keep it flying. is the general consensus that it will crash if one of it's four engines fails? not so.

timmy
Eric Cox August 23, 2013 at 10:07am

*bump*

This reply was deleted.

Discussions

Quadcopter Engine Failure Detection and Recovery Methods

Replies

Activity