Comment by Jason Larocque on June 22, 2010 at 6:45am

Hi to all,

I'm from Sherbrooke University, team VAMUdeS, and I want to bring some presisions.

Sherbrooke University had 2 teams at the competition. Team VAMUdeS and team Elise.

Team VAMUdeS

Team VAMUdeS uses a modified Paparazzi TWOG autopilote. The airframe is a Multiples Mentor. Team VAMUdeS had won the 1st place for Security. More details on www.VAMUdeS.com

Team Élise

Team Élise uses a Micropilote AP. The airframe is a quad tilt rotor. Only 2 motor are used in horizontal flight. More details on http://pages.usherbrooke.ca/elise/


@ Jo: can you please modify the info in your poste using my message.

Thanks

Comment by Mike Bakula on June 22, 2010 at 7:17am

I was with Embry-Riddle SOAR. We flew the ArduPilot Mega, with a V2 sensor board patched in for XYZ and airspeed. The aircraft was my original EasyStar with the Eflite series-six motor and 8 square inch rudder. We were flying both a Webbie using an arduino pro-mini as a timer, and an Axis 207MW with an outboard amplifier for real-time visuals. These were attached under the wings with 2-inch Velcro.

We had a number of challenges, from a last-minute code problem (I had a hard time getting below the rc-mixer to get a fixed pose for flight-termination) to a hardware failure on the flight line (the battery voltage-sense lead came unplugged and touched the sensor board; that wiped out both the board and a set of thermopiles). After an overnight time-out, we replaced the failed hardware and got to fly the next morning.

We got a good auto-takeoff (no hands on the RC controls!), and made it to the first waypoint. However, flying straight away put us at a bad aspect from the Xbee's antenna, and we basically lost the command and telemetry link from takeoff. The plane climbed out to the waypoint, but as it started its turn (which would have restored telemetry), the watchdog timer timed out, and the flight was terminated. (A 30 second timer for link loss is mandated by the rules, to prevent fly-aways.) We got imagery while airborne, but no targets.

This might sound like a disappointing performance, but in fact it was a big step forward for SOAR. In previous years we had suffered failures, and didn't have any spare parts to effect repairs -- they were just too expensive. This year we got past them and got into the air. We also were able to make changes on the fly when the system didn't do what we expected -- previously this would have been cellphone calls to a vendor, and maybe no solution at all.

Despite the short flight, we learned a lot, and validated the Ardupilot as a safe system for use at AUVSI competitions. There was a lot of interest, even by some of the judges, in DIYDrones and Ardupilot. The major system change we'll be making is to improve the command/telemetry link. I think we'll probably go away from the Part 15 Xbees to a Part 93 Amateur Radio solution, based on WiFi.

Comment by Mecha on June 24, 2010 at 4:47am

Yep, we used Attopilot for the second year in a row!, I am very pleased with the results. We had a crash Sunday night around 8:30 pm, 3 days before the competition started. Although the IMU was kicked around in the mangled fuselage from the impact (109kph from 125m high, according to the log file) no damage was sustained.

During the crash our wifi repeater broke (plane side) and I could not get a replacement working in time. The only problem we had with Atto was with a software change that was needed in order to get the old (GooPs with GE) moving map software working. The change altered the log data order which we used for target queuing, rendering our Matlab scripts useless. I notice this behavior only 30 minutes or so before our flight.

The first flight after the crash was at the competition and Atto performance very well. We took 12 minutes from take-off to landing and 24 minutes to manually sort and identify the targets. We placed 8th overall out off 22 teams, not bad at all.

NCSU did help us out at the flight line with two students, Sam and Jhon, thank you guys! During the flight Cindy (Team Captain) was controlling/monitoring Atto, Sam was doing the tracking antenna job and Jhon was idle waiting for images, I was hanging out talking to the judges, liaison job. After landing, Cindy, Sam and John moved on to analyze the targets. I helped by writing down whatever target description they provided, and on Cindy’s command handed the score sheet to the judges.

We also had two behind the scene team members, Chris McNair and Dean Goedde who both work tirelessly to get us the moving map going. Thank you both for an outstanding product and an out of this world tech support!

I look forward to next year’s event!

Comment by Frank Bolduc on June 24, 2010 at 12:30pm

Hi, First of all I want to say congratulations to all the team who participated to the event! I was with team VAMUdeS from the University of Sherbrooke and this competition has been a great experience for us. We look forward to next year's event as well!

@ Joe: Some precisions for the results, The 2010 Safety Award has been attributed to team VAMUdeS, University of Sherbrooke.

Comment by Bob Waldron on June 24, 2010 at 2:37pm

Joe -- I'm involved with the Experimental Aircraft Association activities for electric aircraft, http://www.airventure.org/news/2010/100311_electric.html. We'd love to have representatives from the AUVSI SUAS competition. Would you be able to send me email addresses for the organizations that had electric powered aircraft, or connect me with someone who could help me invite the electric aircraft developers? Thanks! -- Bob (bwaldron@eaaelectricaircraft.com)

Comment by Malcolm Gibson on September 8, 2010 at 7:54pm

Hi,

I'm part of an emerging Aerial Robotics Club at my university and we are looking to start doing some work with integrating a camera system into one of our club planes. Ideally, we would like to also use computer vision to be able to recognize some ground objects. It seems like this competition has the competitors do a similar thing by identifying targets. It would be a tremendous help to know some of the software these teams used for their computer vision. Does anyone know what software is common (or best) for doing computer vision tasks similar to those required by this competition? It would be a tremendous help since I don't have any experience! Thank you and I look forward to becoming part of the diy drones community.

All the Best,
Malcolm

Comment by Mike Bakula on September 8, 2010 at 8:26pm

We use LabVIEW IMAQ imaging, but the school has a site license with National Instruments (NI). If you can't get NI support (you should ask, they're very helpful), look into the open-source OpenCV libraries for your basic functions. Any good machine vision book can get you started with how to use them to isolate and identify simple shapes. Location is done mostly by trig, using your GPS position and aircraft pose (Euler angles) as the basis.

Comment

• ‹ Previous
• 1
• 2
• Next ›
• Page