Hector Darmstadt - Heterogeneous Cooperating Teams of Robots

We uploaded some pictures from RoboCup, a video compilation of all missions, and a video of the mapping mission.

Our team was the first to do the mapping mission, which gives points for the best-in-class-autonomy award. The task is to map as much as possible of the arena and make the map as accurate as possible.

Over the whole week the mapping algoirthms didn't work properly, but yesterday evening and this morning we tuned the parameters for mapping and exploration behavior. Although we had some hardware problems, the Monstertruck found a way through roughly half of the arena and produced a good map.

The last semifinal mission is over. We could score one victim autonomously, our partner team from Iran scored 6 victims teleoperated. So overall we scored as a combined team 16 victims in the semifinals. We'll see later if that's enough for proceeding to the finals. In any case we'll have a mapping run tomorrow, to get scores for the best-in-class autonomy award.

The Monstertruck fell in love with a light and "found" it 4 times, but unfortunately didn't find any "real" victim. But our partners from team YRA, Iran, did a really good job and scored 6 victims by teleoperation.

Tomorrow will be the last semifinal-mission, after that about half of the remaining teams (currently 9) will proceed to the finals.

In mission 5, which was our first cooperative mission with team YRA from iran, we scored 3 victims in total. Our Monstertruck found 1 victim in the yellow arena, the teleoperated robot from YRA found 2 victims in the orange arena. At the end of the mission, the Monstertruck even tried to get to a red victim (and even recognized heat and a sticker inside the hole), but it was too far away for scoring.

The next mission starts at 15:25.

We are not qualified for the semifinals, because we did not find enough victims in the preliminaries. But we are still in the competition as we had the chance to build a team together with YRA from the Islamic Azad University of YAZD. They focus on mobility, therefore the robots have complimentary capabilities, which is a good basis for a cooperation.

Our first common mission starts at 11:40, we hope to be able to score more victims than in the preliminaries.

We scored one victim in the fourth mission. Afterwards, the robot again entered the orange arena (it seems to like that place) and didn't manage to get out autonomously. We tried to teleoperate it back to the starting point (to be allowed to restart the mission without loosing points), but we ran out of time.

Overall we scored two victims in the preliminaries. Tomorrow morning we'll see if that's enough for passing to the semi-finals.

The third mission of the preliminary round is finished. We improved the autonomous behavior and mapping, and the monstertruck drove around really nice. But unfortunately we had problems with our thermal camera (sometimes empty images), which led to the fact that we couldn't focus the victims over several frames. Therefore we didn't score points for that mission.

Today at 5pm is our 4th mission, in the meantime we try to repair the camera.

The first two preliminary runs are over. In the first run we were not lucky and got stuck in difficult terrain before finding a victim. In the second run we first found some spectators, unfortunately they didn't count as victims. But after that we also found a "real" victim, i.e. a baby doll with a heat source.

We continue working on software and hardware and hope to be able to improve our performance until our next run tomorrow.

This afternoon we tested the autonomous behavior in the quite difficult yellow arena. The Monstertruck could easily negotiate smaller obstacles, but it still had problems in some specific situations. We're working on the detection of these challenging arena parts and hope that we will find a solution until tomorrow.

Tomorrow the preliminary rounds start, our first run is at 10:30 am in arena A, the second run at 1:30 pm in arena B.