Vision for auxiliary input

It may be useful to have a mode that drives us to a target using vision as an auxiliary input instead of something else like a gyro in order to control the heading of the robot. This would be most useful when we are mostly lined up, but need to ensure that we stay that way.

Start by locating the robot's heading-control entry point and the existing gyro or other auxiliary-input logic. Clarify how vision should provide heading correction, when it should be active, and what target behavior and tests would define the feature as complete.

Written by the indexing model from the issue text.