[mowbot] A Lurker Says, "Hello".

Robert Bergstrom (robert nospam at us.net)
Sun, 22 Sep 1996 20:26:17 -0400


Having watched the discussion I thought I'd introduce myself. My name
is Robert Bergstrom. I am a software consultant in Virginia, USA with
an Electrical Engineering background (Control Systems). I've been
toying with robotics for about five years now. Occassionally entering a
contest here and there.

Dave Everett wrote:
> I agree that safety would be a prime prerequisite for a commercially
> available autonomous lawnmower, but I don't think it is of any
> importance at this stage, what we need to do is prove the concept
> before we concern ourselves with the realities of consumer safety.

I agree. I would caution the group about focusing too much on the safety
and security aspects of the project. It seems to me that navigation and
sensors will be the critical components of the system. The method used
to acutally cut the grass needs to be efficient, but the specific method
can be tailored to the environment. One you have a functional system, I
believe steps could be taken to "commercialize" it.

Personally, I would prefer a robot which could navigate over one which
simply does a random walk. It is more dificult and the lessons learned
would of course carry over into other tasks. I envision a robot which
has the same proportions as conventional lawn mower, untethered, and
electrically powered. Though more complex than a wheeled robot, a
legged robot has interesting posibilities -- maybe the next model.

As far as behaviors, the first one is to navigate back to its base for
recharging. This is where I'd put any solar cells. Given that the
robot uses differential steering, one simple heuristic for covering an
area is the iterative flood fill from computer graphics. This gives
nice long lines and can account for areas with holes. This does assume
the robot knows the area a in advance.



P.S. If there are any readers in the Northern Virginia area, as the
project progresses, maybe we could get together.