When roll acceleration is zero, the remaining acceleration (i.e., due to gravity) felt by a passenger on the wing edge will be the same as a passenger in the middle of the plane.

Not if the pilot did it right. The net force on you should always be directly toward the floor.

Even when the plane is tipped over at 15* (or whatever their max bank angle happens to be)? The floor is no longer directly underneath you.

The plane turns while banking which results in a force vector mostly straight down towards the floor. If the plane did not turn while banking what you wrote is true.

The ground is whatever you accelerate towards.

Balanced out by the centripetal force of the plane turning!

Underneath is a relative term.

It is called a coordinated turn and it is what one normally aspires to fly.

https://www.boldmethod.com/learn-to-fly/aerodynamics/the-aer...

A coordinated turn with 60 degrees bank should be achievable by any airworthy airplane (it requires pulling 2g), and the maximum bank angle for a steady-state coordinated turn depends on how many g you can pull before you run out of control authority, cause an accelerated stall, or reach the airframe's structural limit.

Next time you fly, get a beverage of your choice and observe how it behaves. Here's an example, courtesy of the great test pilot Bob Hoover:

https://www.youtube.com/watch?v=V9pvG_ZSnCc

If the turn is coordinated you can't tell the difference. I've been in small GA planes turning at 35-40 degrees and it doesn't feel like you're sliding at all, you're pulled back into your seat and "down" (relative to you) into the flooring.

?

I wonder if this could be solved by a rotating inner cabin that always maintain perpendicular to force of gravity when the plane banks?