I would be nervous about this also but I'm working on a lawnmower also. Small 12 Volt motors and not as beefy as wheelchair motors. My design is similar to the retail version. I like to try and build it myself...thats the fun.
Want a "sure to sell" idea? I'll give it to you anyhow...
Note the number of "cookie-cutter" homes (well, at least here in the States; not sure about other places) where the "front lawn" is a patch of grass the size of a postage stamp. Ok - that's an exaggeration, but not by much.
Anyhow - I've been in these kinds of subdivisions, and have watched the proud homeowner trot out a large mower, start it up, make three swipes, and call it done. Lots of work for little payout, plus the noise, and perhaps the pollution given that an electric seems to be rarely used, and the engine only runs for a minute, if that.
In short - if you could build a robotic mower somewhat the size of a Roomba, that could just be pulled off a charger, sat on the grass, turned on and let go - you could probably have a very marketable product.
Now - I've noted that I have a Friendly Robotics RL500 - this is not what I am talking about; that's a robotic mower designed to cut a much larger lawn. The mower itself is about the size of a regular lawnmower, and uses a couple spinning blades (and is anything but lightweight). It is much larger than what is needed for these small lawns that are fairly commonplace, it seems.
I think, though, that something the size of a Roomba that could handle such a micro-lawn would be a good selling product. Especially if it didn't require a buried signal perimeter wire or anything else to detect the edges. It just needs to be something that can taken out of the box, charged, plopped down and run. Making it somewhat "finger safe" would also be a nice thing, if possible.
Now - for you are any others building a robotic lawnmower - there's a few things to check out that might be worthwhile:
1. Take a look at past designs - particularly the Friendly Robotics RL500; it's a fairly simple, but robust design. Another design to look at is the "MowBot" from the late 1960s. Specifically, there's an Popular Science article in the January 1969 issue (see Google Books) about this mower. Really, it was a precursor of the RL500, and anticipated most of the technology.
2. Check out the radio controlled mower that was featured in Make Magazine not too long back; r/c mowers have a long history, too. Something to keep in mind, is that for large areas to cut, using a gas-powered cutting deck makes great sense, giving you longer run-times and concentrating the battery toward powering the wheels. You might also be able to rig something in a "hybrid" manner, using the engine to drive the blade as well as a generator to power the motors (maybe with a small battery in parallel). Don't discount an engine just because it is noisy and a bit messy.
3. What about the concept of a robotic reel mower...?
4. Rather than a perimeter wire (which needs power, can break, etc) - what about buried magnets sensed by a hall-effect sensor?
Just a few ideas...