Re: [mowbot] mowbot itenary

robin nospam at acm.org
Tue, 8 Oct 1996 13:46:50 +0100

DON AUTRY <Don.L.Autry nospam at msfc.nasa.gov> wrote:
> My yard:
> 105'w x 400' which yields right at 1 acre. 3/4 of it mown, 1/4 wood
> Very complicated. fencing diagram below - not to scale
3/4 acre ~= 3000m^2. That's well within Mowbot's proposed target area.

[diagram cut]
What are the boundaries? Are they Mowbot-friendly?

> I separated it from Fred's yard when he ate the AC unit.
What is this? Is it something Mowbot has to worry about?

> Additionally across the front and next to the street is a 4' deep
> ditch that runs across the front of the property. The incline is
> steep as the county really fouled it up with their backhoes 2 years
> ago.
This sounds as if it might be beyond Mowbot's proposed abilities.
We haven't got a maximum slope yet, but 45 degrees has been mentioned.
The back-of-the-envelope calculations I posted about small DC motors
suggest that this would be a difficult requirement to meet.

> Mowbot Shape:
> The pointy front will move pinecones and small sticks out of
> the way. ... Actually a cow catcher type
> attachment will work as well, but solid less likely to tangle.
I can't imagine that a pointy front will help much here. If it is so close to
the ground that it will sweep such small debris as sticks out of the way,
surely it runs the risk of getting embedded in slightly bumpy ground.
But if it is high enough off the floor to avoid this problem, it won't
be doing its job as a sweeper. Maybe a skirt of brushes would work.
But if the cutter is adequately protected from curious fingers anyway,
this shouldn't be necessary as it won't be able to interact with
such large obstacles.

> The front can be tapered from tip to top to facilitate
> getting under low branches etc.
What if you need to go backwards? A smooth curve all round seems to be
the best shape to avoid snagging.

> Size:
> a little over 1 foot wide to give a a 12" or so cutting width.
> Size will actually be determined by motor strength and costs etc.
This sounds about right. We can probably afford to go a bit bigger than
this to get the benefits of bigger motors and batteries. Perhaps 0.5m
(20") diameter gives us a bit more room to play around with, and is
still much smaller than a conventional mower so shouldn't have any
trouble getting in places a mower can get.

> Cutting Rate:
> I Made the following objectives: Cut yard once a week.
>
> 400' x 100' = 40000 square feet.
> divide by 24hours/60minutes/60secs i got around a half a cubic
> foot per second.
Square foot?
Doing it backwards using Don's figures and a variation on the
formula I posted 1996-09-24 gives efficiency e=1/7=0.14,
time to mow all t=600,000s (1 week), forward speed v=0.15m/s (0.5 foot/s),
cutter width w=0.3m (1 foot), and area
a=evwt
=3780m^2 (0.93 acre). Close enough.

> Cutters:
> Weed Terminator blades.
This sounds like a possible alternative to the opposed blades scheme.
Having parts that wear out quickly seems like a bad idea, but if the
replacements are cheap, perhaps we can live with it.

To get your 0.3m coverage needs two such blades. That's going to
be quite a current drain. Can you get figures?

> Power:
> Electric of course.
> Don't want no robot with combustibles running around the place.
> 6 or 12 volt systems for 2 reasons.
> Very standard to find motors etc.
> Can use car or motorcycle battery for testing
Agreed. And for these same reasons I'd go further and consider only 12V
parts. Car batteries are a bit big, but motorcycle batteries are cheap
and look ideal for the job.

> Will go to better and more expensive battery later as I don't
> want acid leaking when mowbot flips in the ditch.
All the small to medium lead-acid batteries in the Maplin and RS
catalogues are completely sealed, so unless you punctured them there'd
be no acid spill.

> Design and Testing Stages: ( the nitty gritty )
I like your ideas and progression from simple to complex design.
Testing the motors, batteries, cutter and sensors early on seems vital
if we are to proceed with any solid basis in fact. Having simple, cheap
and quick-to-build platforms to run tests like this is a good plan.
If anything, I think your very first design is too ambitious since it
does pretty much everything needed except recharge! I think I'd insert
an even more primitive version 0 that was just chassis, battery and
motor that could have various things we wanted to test bolted to it.

> Add a wash hands before dinner logic. Build cleaning
> station which sprays water for 1 - 2 minutes while bot spins.
Interesting idea. The parts that are likely to get mucky are underneath,
so Mowbot's underside would have to be waterproofed against jets,
instead of just its top protected from rain. That does add to the
engineering difficulties.

> HMM. any idea how to build a gate that mowbot can go thru and
> Fred can not?
How about something like one of those magnetic-key cat-flaps? Or,
since Mowbot is going to be bristling with IR emitters, make the gate
respond to an IR code.

[lots deleted]
> communication with base station (home PC) thru RF transmitter.
This would be good. Low-power RF modules that carry 9600 baud data
are cheap enough.

> check with base every so often and report okay.
> Base would call me if he disappears.
I expect this would be a feature of the charging/washing station.

> quick cam photos of intruders
> Send photos to base.
Nice idea.

> respond to call to arms from base station detection of perimeter
> violations.
I can see it now: a herd of angry Mowbots descend on the intruder
and squirt him off the premises:-)

> water spray of intruders, fred.
> reload water and pressure chamber at charge station.
I like this plan too. I wondered if animals might take an unwelcome
interest in Mowbot, so adding a simple deterrent like this might be
needed. And I have a plague of cats which might move elsewhere once
they get the idea that this is Mowbot's territory!

> Does anyone out there have some motors or batteries or other things
> collecting dust for testing? I have many things, and don't mind
> buying some, but don't want to be eaten up buying and testing
> motors and batteries to determine the requirements we
> will need.
The trouble with doing this is that old parts are often of unknown
origin and specification, which means that information gained from the
tests might not be much use. And if the part turns out to be perfect for
the job, where are you going to get another like it? Now I don't
suppose that we are all going to buy exactly the same battery, but
at least we can try to define the sort of battery closely enough
that test and experience are reproducible around the world.

> Will ultrasonic transducers upset a dog's ears? I would hate
> for Fred to start howling every time mowbot goes to work.
40kHz is a lot higher than we can hear (about 20kHz is the human limit),
and though I don't know what the range of dog's hearing is, it
seems unlikely that it could be that high.

> The electric guys here also carry a ultrasonic do hicky
> that keeps dogs at bay. Will that thing blow mowbots mind?
Unlikely Mowbot will even notice it. The ultrasonic transducers
are somewhat directional and (not very sharply) tuned, so unless
the gadget is pointed right at Mowbot and happens to be the right
frequency, it probably won't even be noticed. We might have
two sorts of ultrasonic sensors in Mowbot: one downward-looking
to sense the sort of ground surface and one sideways-looking
to measure distance. The ground sensor is more vulnerable to
interference as it is measuring signal strength; the distance
measuring sensor measures time intervals, so will easily reject
interference that comes at the wrong time.

> Will sensors go straight into 386 type boards i/o ports, or
> will I need something. If so, what, and how much?
The printer port on a PC can serve as an 8-bit digital I/O port (or
with a bit more cunning, a 12-bit port), but that's about it.
You can buy cards that add all sorts of interesting I/O, but
all the ones I have seen are many times more expensive than a
microcontroller that does everything we need.

> Ideas for dog proof, but mowbot accessible gate?
Modified cat-flap.

> Do buried wire lines require power?, if so how much?, can they
> work DC for backup?
Yes, they would use low voltage AC (which could be derived from DC).
I don't know how much power is required; not much I should think.

> How can 2 PC's communicate by RF. Any off the shelf type of
> stuff? I know there is a standard for I.R., but there are many
> trees, sheds, etc in this yard.
IR for fast data outside is impractical. A TV remote might manage to
work outside but only does a few bits a second. For any real data we'd
need radio. You can buy fairly cheap RF modules (tx=17UKP + rx=28UKP =
45UKP = 67US$) that just give and take TTL data, so all you need is a
voltage level converter (maybe just a resistor) and something to control
TX/RX and you can stick these on a PC serial port.

> Could a sensor be built cheaply to detect a strobe light like a
> cop car.
Doubtful. Unless you know the precise colour and frequency of the light,
I'd guess any detector would be very prone to error.

> Is there anyone out there willing to show me how to build
> circuits in a way that a idiot like me can understand?
That's certainly one way the group can help. I'm sure there will be
clear descriptions of what we build when the time comes, and people
willing to fabricate specialist stuff for others should the need arise.
I've started writing a step-by-step guide to the construction of a useful
HC11 board I built a while ago; I'll post a pointer to it when it's done.

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/