Fact sheet - section 6.1

[fselius]

6.1. Details of how the team is addressing the Flight Termination requirements (transmitter failure, loss of data link, loss of GPS, lock-up or failure of autopilot and lockup or failure of GCS, as well as loss of motor power).

When you finish each sub-section mark a V in the checkbox.

• [x] Transmitter failure
• [x] Loss of data link - Point out that the drone doesn't need a data link to perform the missions.
• [x] Loss of GPS - Simply lands instead of RTL?
• [x] Lock up or failure of autopilot - What does Lock up mean? Can't find it on the FRI.
• [x] Failure of GCS - Similar to loss of data link, drone doesn't need a GCS to operate
• [x] Loss of motor power - Does the Pixhawk know how to identify and handle it?

[fselius]

@serbuh @shadows89 Post questions on forums (diydrones, ardupilot.net, efly) and see if the Pixhawk knows how to handle motor failure (For example shutdown all other motors) - ASAP.

[serbuh]

We are using the FrSky X8R receiver. And in case of a failure it configured to use the “No Signal” method – the receiver stops sending signals to the flight controller, and the autopilot activates the failsafe mode = RTL. Did they ask for RTL? How long (In seconds) before the receiver would stop sending signals? Check the rules and make sure it’s what they’ve asked for.

Quote from rules: "The MAV should automatically return home (to the takeoff location) after loss of primary communications link signal for more than 30 seconds. If no return home or RTL is installed the flight termination should consist of a gentle descent until touch- down with a subsequent shut-down of power to the motor(s)."

We thought about the transmitter as a primary communication link. And according to this rule: The flight termination system should be capable of activation by a manual safety override. If we are loosing the link with the remote control that means that we aren't able to make a human override in case of quad's insanity and that means we need to RTL in that case..

Correct me if I didn't understand it right. Aaand.. I didn't find an option to react after 30 seconds. Just an immediate reaction to a loss of RC link.

Usefull link: http://copter.ardupilot.com/wiki/configuration/throttle-failsafe/

[noamyogev84]

Got first comment at gitter/ardupulot : they say that there's no special specification on how the quad reacts after 1/couple motor losts. What happens in such a scenario depends on the quads load, human pilot expertise...etc

[fselius]

@serbuh Do we have a way to set a delay before the Pixhawk activates the failsafe? If something blocks the RC for three seconds we wouldn't want it to abort the mission and land.

[roeemz]

@noamyogev84 is there no indication incase of motor failure? Do we need to manually look for odd situations to figure out something is wrong with the engines? When a failure occurs does it continue to fly "normally" and tries to stabilize itself?

[shadows89]

I think we need to land if we lose RC becasue if that happens we lose the abillity to teminate the flight in case of an emergency, no?

[fselius]

Lockup/loss of motor power/autopilot failure - Human operator shuts off the drone.

Regarding losing RC, we'll do whatever they asked for in the rules. If we lose RC for 3 seconds because someone moved and blocked Line of Sight we don't want to terminate. If we lose RC for 15 seconds we can terminate the flight.

[roeemz]

@fselius, their requirement is 30 seconds.

[noamyogev84]

According to DIYdrones guys and in other groups, in case of motor failure there is no 'auto stabilize' on behalf of the apm. In this situation, there will be an immediate lost of stabilization (maybe spinning or doing all kinds of unpredictable moves) the way to handle this is to take over the drone with the transmitter, and try to land it or fly it with a slight fix to the yaw/pitch values in order to "overcome" the error. if the quad still has enough thrust after motor failure, then it won't crash.

[serbuh]

Yes, absolutely. Take a look at this link a @6:45 http://www.ted.com/talks/raffaello_d_andrea_the_astounding_athletic_power_of_quadcopters?language=en They have a strong processing power outside the quad. They have position sensors on the quad and a good number of cameras in the room. I understood that only with inner sensors you can not reach the stability. The sensors are too slow for such case. So to summarize. I don't see it is fiscally possible. So we need to take manual control and land (read as "crash") in order not to allow for the quad to fly into the audience and make some mincemeat :)