Cyberstreak said:
No, it (AP) struggled to stay on the GS until stick shaker & stall. Max thrust then apllied to no avail at the low altitude.
I'm not a professional pilot, but I am a professional software developer. What you say tells me the systems doesn't "talk" to each other.
AT thinks the aircraft has landed (or very close to landing) and closes throttles. AT struggles to stay on glide slope. Glide slope is a slope, hence after just a few seconds the "clever" system should notice there's something wrong. This is what we call a "sanity check".
10-15 seconds after the AT thinks it's on the ground, AP should yell: "Hey, I'm still trying to follow that glide slope!". One on them
must be wrong! What to do next?
There is always a chance of a faulty system. Either the AT is wrong, or the AP is wrong. As a software developer I would say the best way to deal with such situations is to assume the AT is wrong. That is, give the AP authorization to override AT and command a TOGA. If the aircraft is on the runway the responsibility lays in the hands to the pilots. If it's not on the ground it would just go around. And that should wake up the sleeping pilots.
I agree that the pilots should monitor and stay alert, but from a software point of view there are improvements to be made. To me it looks pretty simple; use all available sources of altitude info. If they do no agree simply do not allow Autoland. If you can't land manually go to alternative airport. Then you have to fix a faulty rad alt before you can Autoland. The money language would yield a quick fix: Too expensive to go to alternative.