Ensure robocluster has a kill switch

[ottopasuuna]

If something goes wrong during the operation of the rover and we need to shut everything down,
a kill command from the master node should ensure every process on the system is stopped immediately.

[jpas]

Should this be a dead-man's switch?

[ottopasuuna]

Possibly, I think It would probably be more annoying than practical though. The issue isn't the operator not paying attention, but the rover not doing what the operator is telling it to do.