We added Duke Robotics Club labels to the chairs, table, and monitor. We swapped out the HDMI cable for our pool monitor since we were having problems with the old one.

We recorded a 10 minute (~10,000 frames) underwater video of the pool from the front DAI camera. This will be used as null images in our CV model. In the middle of the pool test, the DAI auto-connect script connected to the downwards DAI camera and published to the front camera topic. Multiple camera support is addressed in cv-refactoring so this should be fixed soon.

We spent significant time tuning static power global. We found that using only static power to hold depth at -1 m is very difficult. The robot will not stabilize at this depth unless there is a large amount of static power (>1) which causes us to significantly overshoot our Z setpoint when PID is enabled. We ended up increasing static power to -0.25 to -0.50. The robot will still slowly drift upwards at depth -1; however, increasing static power beyond this will interfere with PID too much.

We also double checked yaw after yesterday’s changes. We made the robot depth hold at -1. For 3 minutes, the yaw slowly drifted up to +30 degrees then stabilized back to about +10. This is significantly better than before where the robot would yaw continuously; however, there is still some drift for a few minutes after starting motion.launch.

We attempted to move the robot in a square. Although Will increased the task planning tolerances yesterday, the robot was still having problems moving on to the 2nd task after stabilizing at the 1st setpoint. We should add some logging messages to task planning to determine why the robot is not continuing to the next task.

The thruster Arduino disconnected 2-3 times during the pool test and we had to restart motion.launch.

We took out the robot 3 times during the pool test since humidity reached 70-80%. Each time we took the robot out, we dried the exterior, took the capsule off, and observed minor leaks inside the capsule. Each time we put the robot back in the water, the humidity increased faster than before. Because of this, we ended the pool test early at 3:00 PM just in case.

Temperature increased up to 90 F.


  • Task planning tolerances too high
  • Minor yaw drift for a few minutes after starting motion.launch
  • Rare Arduino disconnect issues
  • Minor robot leaks