I think in both cases you're entering a very saturated market. For computer interface software/hardware, we have
$1 for those with no vision,
$1 for people with some useable vision,
$1 for those who can't afford or don't want to pay for JAWS. On top of that, every modern mac produced in the last 5-10 years has screenreading and screen magnification software built-in (Voiceover and Zoom, respectively). I don't see what a new screenreader can offer, except perhaps if Microsoft ever finishes making Narrator useable enough that the third-party screenreaders aren't absolutely required on Windows.
For environmental navigation, we've got
$1,
$1, and
$1 (Android only). The common flaw with these and other technologies, as would be a flaw in your contribution, is current GPS technology isn't there yet. For example I can give it an address on my block, and it might get me to the vicinity, but it most of the time won't get me to the actual building. So I'd still need to know approximately where I'm going, particularly in relation to where the navigational aid drops me off. In particular, GPS systems love to direct people around to the back of my apartment building. I know that, and know my way around this area enough that I can account for that. But John Q. Blindy getting off the bus in this neighbourhood for the first time isn't going to have a clue.
It's the same problem self-driving cars have, and one of the reasons they're not expected to be out to market for some time--that's ignoring the self-driving cars that fail to see on-coming pedestrians. I know where I live. I can direct a human driver to where I live easily enough. But if I'm relying on a car's GPS system, it may drop me off somewhere entirely different from where I'm expecting to be dropped off, because that's where the GPS told it to stop. It could be around back of the building, like at my place. It could be two or three buildings up the street, like where I work. I'm not going to have the slightest clue until I get out of the vehicle and the building's not there, or it's the wrong one, or I spend 20 minutes figuring out where I am only to realize the back door was 2 feet away from me and I could be inside already.
There are apps that could potentially help with that problem (
$1 is one), but even that gets impractical after a certain point. Unfortunately that's probably the best we've got until we can successfully train algorithm A to accurately plot coordinates B and find building C.