Just having that coffee before heading to the office...
In the current design, the traverser table itself needs to do six things reliably.
It must move through the full travel in (let’s call it) X without deviation, hesitation, etc.
It must stay aligned in Y (track axis)
It must support the weight in Z.
It mustn’t Roll, Pitch or Yaw either, indeed, the guidance system must prevent Yaw, and both guide rail top and wheels should be flat and level, so roll and pitch are irrelevant.
The drive system, including the connection between the quite rigidly-guided carriage & the traverser table needs to do rather fewer things.
Travel through the full X range without deviation...
Not apply any loads in Y
Not apply any moments in Roll, Pitch or Yaw
Downwards Z loads may be acceptable. You could have a “pull-down” spring.
A single rigid post, fixed to the carriage of the drive unit, which engages in a socket (ideally a ball-joint) in the middle of the table, at guiderail top height will minimise or prevent the transmission or creation of moments at the table.
Note that a similar post pointing down from the table to meet the drive system will create a lever arm, and therefore will create a moment that will tend to lift the table off its rails.
To avoid applying any loads in Y, you can consider hinging the post, or making the socket in the traverser table into a slot.
The alternative approach is to rigidly fix the traverser table to the carriage of the drive system. In this case, the model traverser guide rails will be there for show, but will play no part in guiding or supporting the table. I’d take any wheels you can’t see off, and make any wheels you can see sufficiently flexible (sliding on their axles) so they can not provide any guidance at all - just go “along for the ride”.
You’ll then need to arrange a means of adjusting the supports for your drive system so that you can easily adjust the alignment in Y, Z and Yaw. Either screw thread & locknuts, or clamps with shims underneath.
Either approach should work fine. I suspect that the rigid approach might, eventually, prove the easier, and you’ll need it for the lift, so I would tend to go that way.
Have a great day y’all
Simon