June 30, 2017
Search and rescue missions are complicated and employ expensive vehicles like helicopters, enlisting the help of hundreds volunteers or even a team of highly skilled dogs to sniff out a hiker lost in the woods. But now, these missions may become less risky and complicated with the help of unmanned aerial vehicles (UAVs), or drones.
Students at The University of Texas at Austin have been involved with fixed wing drone (airplane) development at the competition level since 2009. Each summer, members of the UT Unmanned Aerial Vehicle team compete against teams from universities across the world in the Association for Unmanned Vehicle Systems International Student Unmanned Aerial Systems (AUVSI UAS) competition at Webster Field in Maryland.
This year, competitors were challenged to design and build a UAV that could detect signs of a lost hiker, then deliver a water bottle to the hiker from the aircraft. Out of the 54 competing teams, UT placed 16th overall and walked away with $1,550 in prize money.
Chief engineer and newly graduated aerospace engineering senior James Bell said that despite flying the airplane for the first time only two weeks before the competition, the aircraft’s takeoff, landing, and flight were seamless thanks to a freshman autopilot operator.
“We did autonomous everything,” Bell said. “We had a perfect autonomous takeoff, perfect autonomous flight, we hit all the waypoints we needed to, and we completed an autonomous landing.”
The landing was a high point of the mission because the team didn’t actually practice landing the aircraft on the ground before. Instead, they practiced by simulating landings in the air.
“It’s a bit risky to practice landing the plane on the ground close to competition because if you do it wrong, you can destroy your aircraft, so we just waited until we got to the competition and did it there,” Bell said. “We were fairly confident it would work, but it’s still exciting that it did.”
The team placed 20th for their mission demonstration which was the bulk of their score and also performed well in other areas of the competition. Teams are graded over three mission components: the mission demonstration which makes up the bulk of a team’s score; a technical paper which describes their engineering decisions in creating the UAV and counts for 20%; and a flight readiness review video, a recorded explanation of their work, is another 20%.
The biggest challenges came from the telemetry and image processing of the UAV. Targets of different shapes and colors are scattered across the ground of Webster Field spelling out “LOST HIKER.” Teams must be capable of reading each target’s location, letter and shape from their UAV and communicate the data back to contest judges.
UT’s team developed a UAV different from previous years so that it could identify targets autonomously. A video camera placed inside the aircraft identifies targets on the ground by their shape, color, letter and location.
“Impressively enough, one of our groups used neural networking to find images automatically,” Bell said. “Instead of a person looking at an image, the program looks at the image. That allowed us to, actually, for the first time ever, find targets automatically.”
Even though Bell said the UAV automatically identified four out of nine targets, its shortcoming came in communicating the information back to the ground. After some quick thinking, Bell and his team were able to send the data to the judges, but a formatting error kept them from receiving points.
“I feel like we did well enough and this was the first time we’ve ever had a good video feed for competition. So it’s really disappointing we didn’t get it but at the same time, it’s all good because the camera was good enough to see it and his code was good enough to recognize it. Next year, that’s looking up.”
The team’s persistence and success didn’t stop there. UT’s technical paper received 5th place out of 54 and their flight readiness review, a video explanation of their work, was 10th despite being unedited.
Bell said even though he plans to retire his spot on the team and pursue graduate school, he is confident the remaining members will do well in future competitions. The team recently began planning training courses on manufacturing and programing for newcomers as well as encouraging old members to stay and pursue leadership roles.
“We have a lot of really promising team members who are incoming sophomores and some leadership we’ve maintained over the years—a lot of good juniors and even seniors staying,” Bell said. “Because of our efforts to build a consistent team and keep members coming back, I’d say we’re going to have 10 to 15 repeat members next year which is far more than we’ve ever had before.”
