> > > ... just move the axle forward about half way and place the castor
> > at the rear. Of course doing this negates the ability to perform
> > turns in the robots own diameter
> > It seems to me that this is quite a disadvantage, since the algorithm
> > for getting out of confined spaces has a lot less freedom. Not that it
> > is impossible of course, but it does mean you need a lot more collision
> > sensors and a clever backtracking routine.
> Nothing particularly clever is needed.
Well it seems clever to me. You now have to make sure you don't
back up in to anything (since you're going to have to be able to go
backwards---a circular robot never needs to), and that your corners
(or off-centre curves in this case) don't hit anything as you turn.
If they do, you have to be able to jiggle yourself round so nothing snags.
This isn't trivial in a small space---I've tried to do it! The easiest
way should be just to retrace your steps, but after you've hit something
you usually aren't quite where you thought you were and every movement
seems to hit somewhere else!
Granted, Mowbot isn't likely to run in to many very confined spaces,
so maybe that argument is a bit flimsy. But I still like the idea
of a robot that can turn inside its own outline.
> The prime spot for just about everything seems to be along the axle
> line. Both cutter and grass sensors would benefit from being there, but
> I can't yet see how we would be able to have them coexist peacefully.
If the aim is to make the cutter ride at a constant average height above
the ground, and the cutter is parallel with the drive wheel axle, then
the optimum place for it somewhere between the axle and half way to the
support wheel, depending on the width of the cutter. The grass height
sensor(s) needs to be just in front of the cutter.
What about my variable-height cutter idea?
Robin.
-- Copyright (C) 1996 R.M.O'Leary <robin nospam at acm.org> All rights reserved. For licence to copy, see http://dragon.swansea.linux.org.uk/mowbot/