[nzlug] Embeded problem - Dairy

[Simon Bridge]

Boy are you going to get a lot of interest...

On Thu, 2007-12-27 at 11:01 +1300, Simon Lyall wrote:
> However this happens when they are in the middle of the milking and in the
> process of attaching the cups to the udders of the cows (from below) so it
> is hard to interrupt this and separate the cow or even climb up and find
> it's ear number.
> 
> The idea they have is to somehow mark the cow and then as it is leaving
> the shed an automatic gate will trigger and divert it into a separate
> area.

When sanity-checking a tech solution, it is a good idea to consider how
to solve the problem with as little tech as possible.

There are a number of things we don't know, like the kind of milking
setup (rotating, fixed, mixed...), how the operation is routinely
carried out (how many people, extent of existing automation...) and what
the farmer currently does with the problem at hand.

A successful deployment will fit as closely as possible with the
existing setup while actually solving the problem under consideration.
If the solution also makes life easier in other areas, this is a bonus.


* The low-tech version is to have a can of spraypaint or dye - mark the
cow that needs the attention, and the joe watching the gate shunts the
marked cows out.

There are existing methods for marking cows that farm-hands are used to.
The marker stuff can be worn in a sling or kept in a pouch on a rail
(rotating shed - the hands don't move as much so a fixed location is
more feasible) until needed. This is especially viable if there is
already a hand watching or operating the exit gate. (Even automatic
gates need to be watched...)


* Next up the line are the passive methods mentioned - reflective or
magnetic cuffs which are detected by the gate mechanism. This ads a new,
but minimal, process. The extra equipment to be carried can be less
cumbersome than markers. Magnetic systems are already used in stores to
set off alarms - how hard would it be to modify a store system to open
and close a gate?


* Active systems - a small radio transmitter on a cuff. Doesn't have to
be an rfid, but rfids have less overhead for this. A detector on the
gate selects out the cows. There's not likely to be much by way of
programming involved as there are commercial solutions already... it's a
matter of rigging the positive result output to a gate servo.

The main advantage would be that it is not as bulky as the previous
systems and less prone to environmental effects (reflectors getting
muddy etc).


* Intellegent systems - rfid the entire herd during routine tagging. The
hand carried a small, lightweight appliance. See's the suspect cow,
pushes a button on the appliance, continues working. The appliance reads
the rfid and sends the tag number to a computer that operates the gates.
A detector or the gates reads the rfid, and shunts the selected cow off
to the side.

This method is a lot of tech for a simple problem - really only sensible
if the farmer was planning to employ rfids anyway. There are lots of
things you can do with rfids, once they are in place, so this could be
an attractive additional selling-point.

The farm manager could use a handheld computer to access a database on
individual animals - especially useful in a vet visit, and those
wee-small-hour tramps through the pastures ("Who do we have here?")


Note: the first solution has an advantage in that, when it doesn't work,
it is easiest to spot - the cows in the main herd that have a bright
orange stripe obviously shouldn't be there - hands will spot it in
seconds.

OTOH: mistakes are more likely - the mark will likely be on the
hindquarters. The cow may be approaching the gate-hand head on.
Sometimes the mark won't be as visible.

The high tech solutions have less likely but hard-to-spot errors - what
happens if a cuff slips off in the jostle or the gate timing is a bit
out?

The main advantage in low tech is that it is easy to understand and
there are already systems in place that can be leveraged to make it
work. The main advantage of the high tech is that it involves fewer
humans - maybe leading to more ambitious, productivity enhancing,
deployments (it's also more buzzword compliant <sigh>).


I'd need to know a bit more to properly evaluate any systems - but
that's not what you wanted here. BTW: I grew up around dairy farms.

Farming is different now though... there is less of the Wal Footrot and
more of the corporate manager about the whole thing.