'Man rating' is a term that gets thrown around so much that it doesn't really have a strict definition any more. It traces its origin back to the start of the space program, when boosters originally designed for launching warheads were retasked to launch people.<br /><br />People are a lot more fragile than thermonuclear weapons, so the first aspect of man rating is ensuring that the entire flight profile is benign enough to deliver the astronauts into orbit more-or-less intact. This includes the acoustic and vibration environments.<br /><br />The aspect which gets more attention is ensuring reliability. An ICBM might be allowed to have a failure rate as high as 1 in 10 during boost phase, since the concept of MAD meant that <b>huge</b> salvos of missiles would be flying. The Shuttle, OTOH, is considered risky (and derided as a 'death trap') at 1 in 438, and the astronaut office has a stated requirement of 1 in 1000 for the CEV booster. In order to achieve this high a reliability rating manned boosters undergo <b>much</b> more scrutiny to ensure that its failure modes are understood, and to design against them.<br /><br />Using the Shuttle as an example, every system is given a criticallity rating - meaning the number of steps that component is away from loss of the the vehicle. The APUs (as a whole) would be a Crit 3R - it requires 3 failures and there is redundancy, the SRB ignitors would be Crit 1 - 1 failure with no redundancy. Man rating for the Shuttle meant eliminating every single possible Crit 1 system, and providing redundancy when it couldn't be eliminated (for example, all the pyros have multiple paths for the ignition signal to pass through, with current supplied from two power busses).<br /><br />(Okay, I probably just totally screwed up the Crit rating system.)<br /><br />All this takes <b>way</b> longer and is more expensive when applied to an existing non-'man rated' design than it does during the original design phase.