Embedded Space Stations

As humanity moves out into the solar system, practically everyone (except for hard-core space-station chauvinists) anticipates thats people will eventually settle on the moon, the moons of Jupiter, large asteroids, and other medium-sized space bodies.

However, all of these objects have a pretty big disadvantage for human life in common: low gravity. The gravity of the Moon (and Io), is around 1/6th that of earth, most of the big outer planet moons have lower gravities (around 1/7th of Earth), and things get even worse when you consider medium-sized moons , Pluto, and various big TNO's.

We all know that micro gravity has pretty serious and unhealthy effects on humans, which we cannot completely alleviate with our exercise regimes and special diets (for astronauts). What about these worlds? Is 1/7th Earth's gravity really sufficient to make a healthy human? Even if it is, does it not seem likely that bodies born and acclimatized to low gravities might not be sturdy enough to allow them to visit higher gravity worlds like Earth (or even Mars and Mercury)?

People propose artificial rotating space stations for colonies floating free in space. Oddly enough, however, I haven't seen anything similar proposed for actual planetary/satellite/object surfaces. After all, centrifuges still work, and produce outwards accelerations, on earth. So why not make rotating surface colonies, or 'embedded space stations'?

There would have to be some differences of course. Since you would have gravitational acceleration as well as centrifugal acceleration, you would have to curve the floors, so that the vector sum of these two accelerations is always perpendicular to the surface (I've calculated that this curvature is that of a parabolic cone). Thus, the actual, net acceleration would vary depending on where one is standing. If the embedded colony is wide and not very deep, however, this effect would be minimal. This would also minimize any difference in the direction of acceleration between a person's head and feet. For a low gravity moon, the floors near the outer rim of an embedded space station would be nearly vertical, if you wanted an earthlike acceleration.

Then, of course, you would also have to support the entire rotating colony on the surface of whatever object it is placed on. However, people handle rotating objects and magnetic levitation all the time, so this should not be too much trouble. For smaller colonies, you might even be able to use wheels. It would be best if the colony was sunk into the surface.

Finally, the colony would be connected to the outside surface via a sort of 'docking hub', in the center of the colony. Connecting to non-rotating sections, people could then enter tunnels extending under (or bridges extending over) the rotating parts of the colony, to get to the surface.

Edit: Just for an idea, imagine a 1 km wide colony on the moon. To get earthlike gravity at the rim, it would have to rotate at 1.33 RPM. At the rim, the angle of the floor to the horizontal would be around 80.5 degrees (Note: math corrected after I re-checked it. Whoops!). Thus, as one walked downwards, you'd move closer to the center, and the angle of the floor would become less steep, and the net 'gravity' would lessen. If the floor went all the way down, you'd eventually reach the center, with a floor horizontal to the lunar surface, with normal lunar gravity.

In reality, of course, it would be impractical to make the colony that deep, so before long, you'd reach the bottom wall, and you'd have to take some sort of elevator to reach floors closer to the center, with less gravity and less steeply tilted floors. At the center, you floor could be a parabolic dish, with the non-rotating docking hub in its center.