This summer, the main goal is to test Crush and make sure that static power, PID, recent updates to Sonar, and the introduction of CV function as intended.
06/06/2026
This was the second weekend of pool tests for the summer, and it began in spectacular fashion with us spending all of Saturday in the Foundry. Due to the bent fiber-optic cable from last week, likely a result of opening Crush’s capsule, we still could not SSH into Crush and get it ready for the pool test. Luckily, Niko came to the rescue and installed the new fiber-optic cable early on Saturday, and after being a little worried about some cut wires, the capsule was connected, and we were able to connect into Crush. Unfortunately, when testing the thrusters, we found that the Arduino, found in the battery’s capsule (not the Jetson’s), was loose and would keep disconnecting, killing any chances of getting Crush into the pool this weekend.
06/07/2026
The focus on Sunday was getting Oogway into the water, establishing a routine with the people at Brodie Gym, and hopefully getting some solid acoustics data. Going into it, Saagar and I discussed concerns about the pool at Brodie being a saltwater pool and it warping acoustics data. We agreed that we should first try to replicate the same frequency and distances that we tested at Wilson to ensure that our original models ported well to saltwater data before proceeding with more frequencies. I also considered the possibility of applying a transformation based on the salinity of the pool, so I measured the salt concentration in Brodie’s pool, likely incorrectly, as I only obtained 1.3PPT, while Wilson’s freshwater pool got around 3PPT.
Due to some confusion from the staff, being a little late, and taking extra time to set up all the power cables in a new setting, we had around 45min-1hr to get some acoustics data. Unfortunately, because I didn’t open the Logic 2 script saved in Oogway when SSH’d into it, I could not run the recorder.py script to collect the acoustics data. To make matters worse, Jared noticed Oogway’s humidity reached around 88%, so we pulled it out of the pool promptly and saw that the temperature and humidity sensor were soaked. Nothing too bad, but it did mean that the pool test was over for the day.
Reflections and Future Goals/Plans
Being new to running pool tests, I made countless errors. Whether it was running the recorder.py script incorrectly, not realizing that running the launch file in the execute package could not be run with controls enabled, without task_planning’s task_runner being spun up, or taking the bot out of the water way too late, there is definitely room for improvement. Hopefully, having established a rhythm, we can get some acoustics data next weekend, and get Crush back in the water and begin static power tuning and PID control. I’m definitely still hopeful we can make this a productive summer. Stay tuned for next weekend!