After reading about the crowdsourced UAVforge.net competition here on DIYDrones, I conjured up the audacity to enter an Arducopter into the fray. Since then, the field has been reduced to 20 teams competing for 10 slots in a fly-off scheduled for May.
After some trial and error, I was able to assemble an Arducopter capable of completing the competition course. I passed the required live, directed demo indoors during a snowstorm last week.
Since it has a crowdsourced component, I need some votes (likes) on the UAVforge website to help make it to the fly-off. If you believe my entry has merit, I'd appreciate your vote.
Team Phase Analytic
http://www.uavforge.net/uavhtml/#
Please forgive the poor videos as I had no "file footage". It seems as though no one else I know is fascinated enough to stand and film endless loiter tests and PID tuning rituals...
If I am lucky enough to secure a slot in the fly-off, I'd like to invite at least 2 of the Arducopter Dev team members to the expenses paid trip to Georgia. It would really be their victory more than mine.
Any suggestion or criticisms are welcome. I am just a fellow hobbyist.
Thank you!
arashi
Team Phase Analytic
Comments
DARPA just released the cleared photos and videos of the event. They can be seen here.
http://www.uavforge.net/uavhtml/flyoff
@Ellison - Yes, we all suspected interference at the flight area. I tested my RC control range at the hotel and all was good. Nothing like driving around rural Georgia at 2 am and randomly stopping on the side of the road while wearing video goggles and a bunch of blinky lights to arouse suspicion. The funniest part is that I wasn't the only one.
We also thought jamming may be occurring from warlocks used in convoy training on base.
Day 3 - Thursday May 10, 2012 - First baseline attempt.
After the failures at the pre-trials, I got lucky enough to fly the first baseline attempt of the competition. Back at the hotel, I had a bunch of perfect flights and thought my troubles were behind me. I power cycled the quad at least 10 times trying to replicate the failure, but everything appeared rock solid. I knew that I hadn't done anything new, so I feared the problem would surface again at the most inopportune time.
I setup my 32' tower at the launch site and performed a quick hover to make sure all was well. I had planned to just fly to the city FPV and land on the grass near the gas station. I had the mission waypoints loaded as a backup.
I put on my googles, took off and started following the road to the city. All was easy and well, then it entered RC failsafe about 300 meters away and fell from about 200'. I still had perfect video and 100% telemetry. You have to remember that I am pumping out 7 watts of frequency hopping 433mhz power from an antenna on a 32' tower.
We quickly retrieved the quad and it had no damage, so I checked my RC system, and readied it for another attempt. That's when the gyro init issue surfaced again... It would just wobble on the ground.
My day was over.
Aeroquad took off and flew to about 100' above the launch. Their RC system RSSI started dropping and they almost lost control directly above the field. Since we were both using 433mhz systems, we suspected some type of interference. One of the DARPA staff had a radio with a spectrum analyzer and it showed nothing at ground level.
We then suspected the large portable cell tower that had been brought in for the event. We drove over to look at it and noticed that the uplink dish pointed over the launch area. DARPA called in a Verizon tech to turn it off to check. Aeroquad flew again above the field with the tower powered down and encountered no interference.
Interesting that it would fly fine in the hotel, but not in the field. Could it be radio interference, deliberate or incidental?
It appears that no one has completed the baseline mission.
Darpa has starting posting a day by day account of the event. http://www.uavforge.net/uavhtml/#
Here was one of my flights at the pre-trials. I was still having the intermittent problem, but thought the open sourced flashed ESC's were to blame. I had changed out the ESC's the night before leaving to the event and it flew perfectly at the hotel the night before this video. On the first flight attempt at the pre-trial, it just floundered on the ground. I suspected the ESC's again and quickly re-calibrated them at the field. I flew again and it took off fine, but had a yaw issue. I hadn't calibrated one of the ESC's correctly, causing the yaw issue, so I landed and calibrated them again. Then it had the gryo init issue and just floundered. It was a very frustrating day.
I went back to the hotel and it flew perfectly again.
Arashi, sorry to hear about your misfortune. I see the posts at the UAVforge website are filled with tales of disaster from some very reputable competitors. Your final model at the fly-off looked really sleek and quite hard to spot, which I suppose is a good thing in the context of the competition. Let's hope they hold it again.
I'll wait until the competition is finished to post detailed accounts, but I have nothing but praise to offer regarding all aspects of the event. I've read and heard a lot of negative comments about military involvement and conspiracy theories prior to the event. Heck, I had my own theories, but they were all dispelled in short order. I think my feelings are best captured by my post on the UAVForge blog:
What an amazing event!
I would like to thank all of the DARPA, SPAWAR, and military employees, volunteers, and observers, both civilian and military for making this happen. It was no small effort to invent, plan, and execute this first of its kind, logistically challenging event. At all times, every effort was made to help us succeed and address any of our concerns, both real and imagined. The caliber and dedication of the staff involved was inspiring.
My own success was hampered by an intermittent gyro initialization problem that wasn't solved until the 2nd day of the event, and a dodgy long range radio control system that plagued me throughout. I failed myself by not bringing all of the spare parts I had available to me. I'll never let that happen again.
The best part was meeting and befriending like minded people from around the world who's enthusiasm for UAV's matched or exceeded my own. The collaboration among teams to solve common, and individual problems was incredible. This was especially so, given it was a competitive event with a large purse. The selflessness with which teams helped each other was antithesis to personal interests and validated that elements of the event's structure truly attracted and provided an atmosphere in which "crowd sourcing" was possible.
Arashi, you're doing a lot of pioneering and advanced stuff! I would love to hear some detailed accounts of all your experience for the competition.
http://code.google.com/p/ardupilot-mega/issues/detail?id=431
Apparently is a known issue with the APM1, but was an unknown issue to me. I had always used the attopilot current sensor instead of the onboard voltage dividers. I took out the attopilot current sensor for weight and power draw concerns for the competition. I have to boot the APM before attaching the leads to the voltage divider. This was impractical because I used a mosfet switch to turn most everything off during the perch phase to save 1500mah plus in parasitic draws.
I was pretty happy when I got to the point of having a 35 min flight endurance with an additional 4 hours of live video during the perch stage with only a 2kg total take off weight.
Arashi: Many thanks for the great update. Robotics competitions are hard, and are rarely successfully completed in the first year, so you should feel good about the learning, at least. I'm not familiar with that voltage monitoring issue with the AttoPilot sensor (I always use the built-in sensor with APM1, which works fine). Can you post an issue on it in the Issue Tracker?
-
1
-
2
-
3
-
4
-
5
of 5 Next