I do not know of a way to do that, but here are some things to think about.
First, I don't know why it would have to be a different partition---i.e. I think you just want another instance of the SW. If you are concerned about a HW failure, then you'd want a different physical disk.
Second, I'd assume that this would involve code running outside of the process in question. The process would export diagnostic data which the watchdog program would use to determine if action was required. My question would be --What happens when the watchdog fails?
Hopefully, a real programmer will step in here pretty quick....
Meanwhile, I'd try googling on--eg--"redundant software", "fail-safe software", or something similar.