If the loop is large, you don’t have to move too fast. Fairground rides spin fast enough to exceed 1g with little kids in them, and those rides are usually not wider than 10m, often less. Speed here is distance around the circumference, so twice the radius, half the turning speed.
I don’t have the numbers in front of me, but generally it’s considered that the issue is more that if you’re standing on a spinning cylinder, you might have weird balance issues due to speed differences between head and feet. Also, you have to remember you don’t have gravity, so as you intuit, entry and exit are trickier.
Two broad ideas for entry/exit: do it at the center where the physical position is stable, then “spin up” some how, or alternately get up to external speed and lock on to something that you then elevator through the surface.
Take a really long cable, put a suitable weighted asteroid on the other end, then spin the whole thing up. As long as the cable holds, you have a decent amount of centrifugal force without needing a particularly large habitat.
The usual design for these things have very low surface speeds, so you're not likely to go off into the deep even if it fails, as you're going to be pretty close to your prior combined orbit even without thrust, and the Δv needed to get back will be tiny.
But also, the forces are never more than the apparent gravity times the mass, so when someone does finally end up experiencing this, it will be part of a larger pattern of bad maintenance and/or engineering that likely has other things going wrong at the same time.
Neal Stephenson suggested having two habitats with hooks so they’d spin against each-other. Then you could have fun and climb up to see your neighbors.
Was assuming they do a kind of ferris wheel with center and 3 or 4 spokes to start, maybe 200m diameter, and gradually fill out the rim with modular units maintaining balance, and center of mass at central dock ?
