Single phase of polyphase is one question and perhaps one
of the more important - polyphase (N >=2) could really help
with how the VIN gets slugged.
Current mode control is easier to get current sharing from,
but I wonders to meself whether you really care about the
sharing other than ensuring none of them goes "over the
top" at the big end while some other one is loafing. I'd tend
to think that "just barely good enough" inductor series
resistance and FET on resistance ought to enforce some
degree of equitable sharing at higher currents and equal
pulse widths.
Now, for the single phase case, why not just a plurality
of drivers and FET-pairs all run off a single PWM source?
Presuming equality of drive and FET negative local
feedback vs temp, maybe "good enough" comes for free?
As in simply paralleling the FETs from a common gate,
common drain, which seems to be non-controversial?