Welcome back to our pool testing blogs! This blog will serve as updates for both tests.
02/07/2026
Acoustics
On Saturday, we focused on collecting long-overdue acoustics data for the acoustics team. We originally had some issues with the Logic software and our data collection scripts. Due to time constraints, we decided to use Logic 1 to collect the data instead of Logic 2, preferring some lower-quality data over no data. During data collection, we observed clear pings on all channels when the hydrophone was placed into the water. This data will help drive future decisions on our acoustics codebase, and we hope to integrate this vital sensor into our task planning soon.
02/08/2026
Sonar
We started Sunday’s pool test with sonar task planning. First, we tested our task to align the robot at an angle from the wall. This had mixed results, which continued into our testing when we tried to align straight against the wall. We observed that, depending on whether the robot was oriented clockwise or counterclockwise to the wall, it would either correctly align with the wall or align the other direction (180 degrees offset). We believe that this is mainly due to a geometric calculation error in our code, which should be able to be resolved with some additional angle offsets. We plan to finalize this testing at a future pool test.
Pre-Qualifying
We then set out the gate and buoy in the water to complete a pre-qualifying run. After some HSV re-tuning, we were able to consistently move forward through the gate and towards the buoy as expected. However, we quickly ran into issues since the pre-qualify code was intended for two consecutive lanes of the pool, while we only had one lane rented for the session. After some minor adjustments, we concluded that it would be too difficult to accurately maneuver around a buoy in one lane and decided to attempt this at a future date.
Yaw to CV Object
We finally worked to re-test our Yaw to CV Object algorithm. However, syntax errors quickly blocked our progress when the object was in frame. We suspect this is due to message type and attribute reference errors, but we were unable to quickly diagnose and come up with a fix during the limited time we had at the pool. Additionally, when the object was not in the frame, the robot would surface for no apparent reason. Looking into the code yielded no reason for this phenomenon either. We decided to shelve this and revisit the syntax this week, and hope to test again at the following pool test.
DVL Logger
Throughout our time with Oogway at Duke Robotics, we have noticed minor inconsistencies with the data reported by our DVL. Since the competition last year, we have been looking into fixes, whether configuration-wise or code-wise. Recently, our team suspected that data was not being sent from the DVL at a consistent rate, so we have worked to collect timestamp data from the DVL during this pool test. We hope to analyze this data this week and continue diagnosing to look for a fix.
Post-Pool Test Thoughts
While we made progress on some of the tasks that needed CS testing, pool test time continues to be a limited and needed commodity. This weekend has taught us that we need to do as much testing as possible outside of the pool to rid our code of as many syntax and logistical issues as possible – before heading to the pool.
Back at the lab, we once again observed some water pooling in the battery capsule, as well as minimal amounts of droplets in the capsules, and some water pooled around the electrical stack pillars. We have turned over the relevant photo evidence to mechanical, and hope they can come to a solution soon. In addition, we worked with non-functional temperature, humidity, and voltage sensors during the past two pool tests. We hope electrical can debug this issue soon, so we can continually monitor the robot’s health while in and out of the water.
Unfortunately, I will be out-of-state during the next pool test this upcoming Saturday. However, look forward to an update from other members as we continue to push forward with our testing. That is all for this week!