Robustness Improvements from 8/28 Pilot

[amalnanavati]

Description

This PR implements various robustness improvements, spurred by the 8/28 pilot and subsequent testing. The big changes include:

1. Rotating the hospital table 90 degrees in both the seated and bedside planning scene, to better account for the user's spatial configuration.
2. Adding timeouts and reducing retries, intended to prevent adding the "in front of face wall" from taking unreasonably long.
3. Making the robot plan to n cm in front of the mouth in the direction of the fork as opposed to away from the wheelchair headrest.
4. Ensuring that the planning scene initialization takes into account all the currently-used robot configurations.
5. Destroying rates after use, to prevent overwhelming the ROS middleware with timers.
6. Have actions plan deeper, as we did in the Jan study.
7. Send acquisition report in feedback, not result, so we can register success even if part-but-not-all of the acquisition action succeeded (see feeding_web_interface#145).

Testing procedure

• Launch the code in real.
• Thoroughly stress-test it across multiple bites.

Before opening a pull request

• Format your code using black formatter python3 -m black .
• Run your code through pylint and address all warnings/errors. The only warnings that are acceptable to not address is TODOs that should be addressed in a future PR. From the top-level ada_feeding directory, run: pylint --recursive=y --rcfile=.pylintrc ..

Before Merging

• Squash & Merge