I would have a slightly different approach here. Instead of trying to make the thing smart with lots of sensors I would start from the "clippers on a long pole" or some other manually-actuated design and try to solve the things that make it fail in reality.
Some ideas:
Clippers on a unit that just sits stationary on the floor, actuation via a cable, so you can hold the actuator but the clippers don't wobble around. Some control to move the clippers but also some reliance on being able to position your foot. A foot rest that is designed to help you position your foot correctly (with, eg, individual toe-rest, etc). Would need to experiment a bit to work out what clipper or foot-rest movement is actually needed - the less the better, for simplicity and reliability, but given users have mobility problems and most people aren't super dexterous with their toes anyway, perhaps it does need several axes of controllable motion.
I imagine one problem is seeing what you're doing so add a camera in the unit on the floor, put a display in the control unit. Nice nail close-up to work with ;-)
If everything is motorized you can use a smartphone for the control unit, but motorization might not be the best approach and wouldn't be my initial approach. "Actuation via cable" could mean a literal physical cable like a bicycle brake cable.
Actual clippers are cheap so if it makes things easier to design you can use, eg, multiple clippers in the floor unit (different sizes or angles) to use for different nails. (I don't mean clip multiple nails at once)
For woodworking/working with machine tools, you can get lasers that fit to the tool and show on the piece exactly where the cut will be. These are very simple (they're just static things). Perhaps something similar could be included to help with positioning. Again, the human will be doing the positioning, just want to give them good controls and clear direct feedback to work with.