At IARC 2016 we were once again able to show off the improvements we made to our vehicle, and demonstrated the steps we have taken towards making the impossible Mission 7 possible.
Leading up to the event MAAV went into overdrive, with most available team members working late into the night to finish tuning the vehicle. Every team member worked hard on the project, and our performance showed it.
The following was the IARC 2016 schedule:
Monday, August 1st - Travel
Tuseday, August 2st - Presentations
Wednesday, August 3st - Competition
Thursday, August 4st - Banquet, Aquarium visit
Friday, August 5st - Travel
Our trip began early at 7:00 am and took us all day. The following are selfies we took to pass the time:
After we arrived in Atlanta, we still had plenty of work to do. Ryan worked on finishing the Navigation pipeline, while the control subteam headed to the lowest level of the parking garage to do some late night PID tuning.
Our day began with IARC presentations. We first heard from the founder and judges, who gave us the history of the IARC, and gave an inspirational speech about why we were there. We were then introduced to the judges, who helped clarify some of the rules for the next day. Following that, teams began presenting their vehicles. Sajan Patel presented for MAAV, giving an excellent overview of both our vehicles and the algorithms that make it work properly.
Wednesday was the big day for our flight. We got three attempts at completing the mission each 10 minutes. Our goal for this year was not to complete the mission, but rather gather as much data as possible from the actual field, so we could make appropriate improvements.
Our first attempt was a mistake on our part. We began our flight and immediately went into autonomous mode. From there, we drifted to the side for a while, and then killed the vehicle to prevent it from going out of bounds. Because we exited autonimous mode, this marked the end of our attempt… The next time we let the judges know that we didn’t want to end our flight after killing, but rather keep going to keep gathering data.
The second time around we flew in manual mode, since this allowed us to get the most data from our flight. We also strapped a camera to the bottom of our vehicle, so we could get video from the flight to use for improving our vision algorithms. We flew once in a square around the field, then hovered above a few different roombas, tracking them to see what they would look like from the vehicle’s point of view. This is the video we took:
For our final attempt, we once again recorded video of our flight, however this time we flew much more aggressively, actually trying to land on the roombas. This would show us both how it would look like from the camera’s perspective, and also how easy it would be to trigger the roomba’s switch with our vehicle. As it turns out, triggering the roombas manually is much harder to do than one would expect. During this flight, we however discovered a critical issue. For a few seconds at a time, our vehicle would stop responding to controller input. We discovered that this was due to the receiver being wiggled loose due to shaking. However, before this we managed to lose control at a height of around 3 meters, causing quite the crash. But due to our design, the vehicle kept flying without problem, and without damage to any of the sensors. This likely helped us get the “Best System” award, since no other team had managed to keep flying after a crash of that magnitude.
This was the first day we were able to sleep in, since the IARC banquet didn’t start until 11:30. The banquet was held at the Georgia Aquarium. We listened to speeches from the organizers, and then had the awards ceremony. MAAV won the award for “Best Presentation” and “Best System”.
Finally, IARC ended with everyone being given access to the Georgia Aquarium, which we enjoyed quite a bit.
While many of us were sad that the trip was over, we already knew what to improve for next year. With two main awards under our belt, and an overall second place, we were satisfied with our performance. And if we keep up our hard work, we’re sure to impress next year as well.