remote administration and power failure

Hi all,

Suppose I have a Linux box without UPS (uninterruptible power supply) at
a remote site, and suppose power is lost momentarily. When the machine
reboots, there is a chance that fsck cannot repair all the partitions
that were not cleanly unmounted, and asks the user to do it manually.
However, since this is a remote machine, I cannot be present at the
console to run fsck. How do I set things up so that the machine is more
robust to situations like this? I want to maximize the chance that I
can ssh into the machine even after a power failure. My experience so
far with manual fsck is that all that is required is a simple

<enter root password>
fsck -y /dev/hda1

or some such thing; how do I automate this?

Thanks,
Hsiu-Khuern.