Robosub 2016 – Day 6: The Validation


Today was a big day. There was a lot of sails to set (see what we did there) but we managed to pull through.

We fixed all the problems we had yesterday (malfunction in timeouts and weird behaviors) during the night and the morning.

We had 3 tests. The first one was spent getting hydrophones data. We realize we did not have reliable settings for 35 kHz. We were however able to locate the pigner at 30 kHz in the other ring. Alexi spent the day improving the firmware and adding debug steps.

The next state was for mapping. There was a lot of things to do and we were not as ready as we should have. As soon as we started, the submarine was going way off its course. We were able to do some mapping and acquire some vision data, but no real achievement there. We did not have any software issue. It turned out it was an auto-complete error (putting the Yaw as Y target tends to send the submarine pretty far…). The vision alignement and the sonar mapping worked as expected.

For the last test, we had a fully prepared mission, reviewed by the entire team. We had fail-case, redundancy and predefined behavior. We however had a very weird behavior, where we exited the mission way too early. We spend the rest of the test trying individual states and combinations. Good news: we were able to locate the pinger with the new firmware! It’s very encouraging!

We decided to implement a form of simulator tonight. Using ROS and Python (@EricArsenau), we’ll mimic our devices data and we’ll step through the state machine, testing any cases we want. As we speak, 3/4 of the simulator is done. We’ll spend the night playing with the code going through the possibilities.

Today we did not acheived as much as we wanted to, but the team performed really well. Yesterday, everybody worked on patches and upgrade to the software. Also, since the beginning of the week, we have fixed every bug and we never lost any test time with redundant problem. The submarine is very stable, we only need more test time and less sunburn 🙂.

P.S. We had a soldering that broked on the frame. It was fixed before we got out of our ThinkTank, thanks to our mechanical pirates, Charles, PierLuc and Antoine!