Just my opinion, but you are actually cleaning a filter over the sensor and
not the sensor itself.  Part of the problem is that the sensor is so far
down in there it would be difficult to just go in there with a polishing
cloth.

There's something called a low-pass filter over the actual sensor --
and that's what we clean. Since it's so close to the actual sensor,
"cleaning marks" would be a bad idea (I've got a Minolta Autocord that
has cleaning marks all over the front element, but they don't have a
particularly serious effect on the lens since they're a bit farther
away from the film).

Most of the dust can be blown off -- and that which can't often comes
off with a fine brush.

In article
<44f13219$0$10796$5a62ac22@per-qv1-newsreader-01.iinet.net.au>, Gerrit
't Hart <gthart@sad.au> writes

Most, if not all, do put high quality glass over the sensor. The various
specialized cleaning tools are ways of reaching it, and of reaching your
money.

The sensor already has a high quality glass cover over it.  That's where
the dust is sticking!  And it's a pain to clean easily.

But, maybe you mean another glass piece that is larger and more easilly
accessed?  This would be further from the sensor plane proper, so:

Every glass element adds "defect" to the light path, increases flare,
decreases contrast and resolution.  (eg: IS and VR lenses resolve less
than their non-IS/VR counterparts).

Further, a glass element at the location you describe would pass light
rays that are at non-right angles to it (except dead center) and so
would make a minor change in the focal plane as well as creating
chromatic aberration towards the edges/corners of the image.  The focal
plane is easilly fixed through focusing, but the chromatic aberration
could not be for all focal lengths.

The idea you propose is not new, not stupid, not dumb, and it does have
merit, but compared to the drawbacks, just doesn't carry the day.

Cheers,
Alan