The optimal way to solve this is to invest more money in screen-readers. Unlike robotic wheelchairs climbing stairs, software can be copied for free once it's developed. If the government wants to step in, they can provide purchase vouchers to the disabled and let them drive the direction of screen-reading innovation.
Also, that's the best way of putting of the problem for the future to solve it. We have really good ways of dealing with it /right now/, as long as people pay attention to it early in the process of building/creating/designing something.
In a perfect world, we'd do both things at the same time. When people thing about accessibility now they think about people with mental or physical problems, but what about when our generation gets old? The only reason this is not a very widespread problem is that there are very few people with accessibility problems, this will not be the case 10-15 years from now.
Incidentally, the robotic wheelchair that can climb stairs (the iBot) is now discontinued [1]. Based on the bevy of comments, users are sad to see it go. Also, I believe Deka (and now J&J) have extensive patents in this space. They should release into the public domain so that others can build on the technology.
