We remapped the joystick inputs and it now works as expected. We tested all 6 axes, including roll and pitch.
Like Will, we noticed Foxglove becoming unresponsive if multiple windows are open. We noticed that the Foxglove bridge node was reporting that the send buffer limit has been reached which is the most likely cause of this issue. As a potential fix, we can increase the buffer limit with the send_buffer_limit argument. Instead of fg-ws, try using:
roslaunch --screen foxglove_bridge foxglove_bridge.launch port:=8765 send_buffer_limit:=100000000We depth held without yaw drift for 2 minutes. Immediately after turning the robot on, it is important to hold the robot very steady until controls are enabled.
We were able to move in a square and in a straight line towards the pole. However, we are still having difficulty putting these two subtasks together.
To keep the prequal consistent, put the pole at the 6 ft sign and start the robot facing the pole on its left side. The current task planning code should move the robot forward, turn right at the gate, then move backwards to the starting position.
It is difficult to balance the pole since it only has 2 feet. The pole needs 4 feet to be stable.
Humidity started at 70% and reached high 80s during the pool test.
When we took the capsule off the first time, we noticed a large amount of water around the old DVL port.
When we took the capsule off the second time, we noticed some water around the DAI camera ports.
See the following video of our best prequal attempt yet: