AutoPilot Glider(no motor) Released at 5000 Feet 2 miles from landing location...

There are several pretty good discussions on the v2.67 thread that are relevant. One part discusses airspeed control and that the code isn't really flexible enough now to handle gliders. My observation here is that there's a flag in defines.h that covers what attitude.pde does for pitch control, but instead of fully testing this I just changed attitude.pde to ONLY do airspeed control and set gains accordingly. I've progressed past my current UAV testing to set gains for my next flight. In other words, I think it's hard wired right now not to do airspeed control. But, look at the code and the 2.67 release thread on the forum. The other big elephant is also discussed in the 2.67 thread related to temperature compensation for the accelerometers. Personally I don't see several degrees of misalignment as an issue since we're closed loop, but I'm more worried about temperature compensation for the gyros and/or how the DCM aiding reacts to observed drift rates of 30 deg/sec. I saw drift rates grow to 30 degrees/sec in freezer testing, but luckily attitudes were still good. Who knows what happens when dynamics are introduced. I have also noted that current methods in AP_aprs.dcm are simplified. Not to get into too much detail here, but there may be issues if the plane is suspended from the tail. Don't quite know yet. What I do know is that there is an assumption that the body attitude is co-incident with the velocity vector for GPS aiding; and, this is done when the ground speed goes over a fairly small threshold. In a HAB flight velocity vector and body attitude are not aligned while under the balloon. This part is fairly easily fixed in SW. My current code only uses GPS aiding (in AP_APRS.cpp) after the drop occurs.

I have information and post flight root cause and corrective action + flight data posted on my blog for NTNS-2. My progress to NTNS-3 is going fairly well and I'm doing HIL with Flight Gear to verify recent changes. Almost forgot. If you're using the UBLOX LEA-6 GPS v2.67 has a bug that "sometimes" initializes such that HAB altitudes are not supported. Best to wait until this is fixed before attempting a flight.

Regards,

Larry G

Replies

Larry Grater November 30, 2012 at 4:44pm

There are several pretty good discussions on the v2.67 thread that are relevant. One part discusses airspeed control and that the code isn't really flexible enough now to handle gliders. My observation here is that there's a flag in defines.h that covers what attitude.pde does for pitch control, but instead of fully testing this I just changed attitude.pde to ONLY do airspeed control and set gains accordingly. I've progressed past my current UAV testing to set gains for my next flight. In other words, I think it's hard wired right now not to do airspeed control. But, look at the code and the 2.67 release thread on the forum. The other big elephant is also discussed in the 2.67 thread related to temperature compensation for the accelerometers. Personally I don't see several degrees of misalignment as an issue since we're closed loop, but I'm more worried about temperature compensation for the gyros and/or how the DCM aiding reacts to observed drift rates of 30 deg/sec. I saw drift rates grow to 30 degrees/sec in freezer testing, but luckily attitudes were still good. Who knows what happens when dynamics are introduced. I have also noted that current methods in AP_aprs.dcm are simplified. Not to get into too much detail here, but there may be issues if the plane is suspended from the tail. Don't quite know yet. What I do know is that there is an assumption that the body attitude is co-incident with the velocity vector for GPS aiding; and, this is done when the ground speed goes over a fairly small threshold. In a HAB flight velocity vector and body attitude are not aligned while under the balloon. This part is fairly easily fixed in SW. My current code only uses GPS aiding (in AP_APRS.cpp) after the drop occurs.

I have information and post flight root cause and corrective action + flight data posted on my blog for NTNS-2. My progress to NTNS-3 is going fairly well and I'm doing HIL with Flight Gear to verify recent changes. Almost forgot. If you're using the UBLOX LEA-6 GPS v2.67 has a bug that "sometimes" initializes such that HAB altitudes are not supported. Best to wait until this is fixed before attempting a flight.

Regards,

Larry G
Darren Orange November 30, 2012 at 2:43pm

The real goal is to drop the glider from 100,000 feet from a range of almost 100 miles to return to launch site.

The reason for this is to test it in a vacuum and confirm it can mange the low pressure and strong wind sheers.
Bradley J Carr November 30, 2012 at 9:28am

I have also wondered this. How the APM would react if its not getting to were it needs to be. But by controlling the decent rate of the airplane and controlling the airspeed. I wonder what would actually happen. I have a airplane that is chomping at the bit to try this. I just have a bad feeling about what would happen. I don't mind crashing a couple planes, but I would like to keep my electronics.
Tim McPherson November 30, 2012 at 9:02am

Hi,

With that floater and the 2 to 1 glide ratio, I'd say that the APM will fly slow and straight to the destination. A not so windy day will help.

This reply was deleted.

Discussions

AutoPilot Glider(no motor) Released at 5000 Feet 2 miles from landing location...

Replies

Activity