After 5 weeks out of service, we were ecstatic to get back into the pool. Thankfully, it was not 6-7 weeks — that would have caused major timeline delays.
Goals
Our goals for this week’s pool test was to calibrate Oogway’s Gyro after returning from California, collect sonar data, and demo the robot.
Setup and Logistics
Today’s setup went much smoother than last time. We mostly had all that we needed, sans joystick for the demo (granted, that was not on the checklist). We had to adjust buoyancy on Crush after a few mechanical changes, but that was quick. We then got both robots in the water quickly.
The only thing that was missing that we wish we had was a torpedo banner for sonar data collection. Unfortunately, we did not yet rebuild the prop since our return from LA, so we were unable to scan the torpedo banner. We used the buoy as a rather imperfect substitute.
Oogway
Upon putting Oogway in the water, we noticed bubbles. This is of course concerning; the two causes for this would be either new 3D prints, or water leaking into the capsule. While the former certainly contributed to the bubbles, unfortunately that was not the only cause. Our humidity sensor readings kept rising at a constant pace. Once it had risen 15% without showing any slowdown in pace, we took it out of the water and did notice some water pooled inside the capsule. Since we last tested Oogway, the battery cable was redone, so mechanical will take another look at that. Regardless, Oogway was done in the water for the day.
We then took the chance to re-tune the zero-bias on Oogway’s gyro given the relatively significant geographical difference and thus magnetic field difference between Duke and LA. Using our previous NC values as a baseline, we were able to get it within 0.00001 rad/sec, which is more than sufficient for our purposes, so Oogway should be fully ready to start tasks the next time we get in the water.
We fortunately did not notice any of the previous power issues plaguing Oogway, and thanks to Electrical’s hard work, the peripherals Arduino appears much stabler (with correct voltage readings!). We did not notice any issues with the thrusters Arduino, so we hope that the magic voodoo from the peripherals Arduino has fixed the thrusters Arduino, but time will tell.
Crush
Crush had a successful pool test. After getting in the water, we slightly re-tuned Crush’s buoyancy then connected the joysticks to demo Crush to our parents for Duke’s family weekend. Thank you to all the parents that came! Then, of course some members wanted to drive the robot with the joystick, of which we were not going to say no 🙂
Post-demo, Crush was handed over to the sonar team to collect data with the wall and buoy, of which the data looked quite good. However, the bigger challenge will be with the much less solid tarp.
Concluding Thoughts
This pool test went much better than last time. We had a few small hiccups, but actually got some substantial work done. Moving forward, it’ll return to just Oogway in the water. Crush will be going out of service for a switch to eight thrusters and adding some marker droppers. Going forward, we plan to ensure our cleaned-up comp code works well (this pool test we ran the vanilla comp code), then move on to task completion with an emphasis on search algorithms.
Thank you to Ahaan and Wynton for swimming!