2007 July 15 Sunday
Robots Can Replace Immigrant Field Labor
An article in MIT's Technology Review reports on yet another promising agricultural robot that demonstrates we can eliminate the need for low skilled labor in agriculture.
Scientists in Denmark are developing an agricultural robot for identifying and eliminating weeds. While this might seem like a relatively easy task, it actually requires a lot of machine intelligence to pick out the weeds among the crops. The robot is still in the early stages of development, but the researchers hope that it will ultimately lead to a reduction in the amount of herbicides used by farmers and therefore cut costs.
Called Hortibot, the semi-autonomous robot is a navigational platform designed to have different agricultural tools fitted to it to either mechanically remove weeds or precision-spray them with herbicide.
The cheap labor lobby argues the United States must let in huge numbers of low skilled illegal aliens from Mexico to do grunt work. But necessity is the mother of invention. Take away the supply of cheap (really subsidized) labor to farmers and then the market will produce many more robots to do every job in farm fields.
This robot can identify rows of crops and navigate the rows.
At a recent Field Robot Event, held in Wageningen, in the Netherlands, Hortibot was able to follow furrows and autonomously turn in the appropriate direction when it reached the edge of the crop rows.
The ability to recognize different kinds of leaves in order to selectively spray herbicides only on weeds reduces chemical use and reduces the need for labor. The navigational ability has applications beyond spraying of herbicides. Initial crop planting and harvesting both would benefit from autonomous machines that can navigate farm fields and stay oriented in row direction.
This robot still needs a human handler to watch it in case it gets off track. But cameras mounted on such robots could allow a single handler to sit at a desk and track multiple robots watching for mistakes that require manual correction.
"But cameras mounted on such robots could allow a single handler to sit at a desk and track multiple robots watching for mistakes that require manual correction."
That isn't quite true. It is very hard for humans to understand what occurs in video streams from remote locations.
It is also very hard to handle multiple video streams. It is also very hard to have reliable video broadcast from multiple robots out in the middle of a field -- at least it is hard in Iraq.
I think I read somewhere that this unit is $70,000. What is that, around 3 farm workers' salaries?
The difficulty of operating outside and in potentially wet conditions means this won't necessarily come down too much. But eventually, we'll get better at it.
On a side note, one way to reduce the number of immigrants interested in working in the US is to free up trade for agriculture and remove subsidies. Both would drive production into other countries, which would be mutually beneficial for a number of reasons. I've probably mentioned this before.
You bet, if a machine can assemble a computer, perhaps it can learn to pick a cherry or pull a weed. The challenge is increased by the fact that weeds often resemble the crop plant and prefer similar cultural conditions. An example is velvet leaf (weed) in soybeans (crop plant.) If we look back a few decades, we note that the history of agriculture is largely the history of the mechanization of harvest. The hay baler and the cotton picker are inventions that were only made in the last century.
Ivan said: Both would drive production into other countries, which would be mutually beneficial for a number of reasons.
agree entirely Ivan.
Randall, using current and near-future technology this just isn't sufficiently reliable or cost effective - especially when you've got lots of cheap labour to fall back on. It might be easier to approach the problem from the other direction and genetically engineer machine-friendly plants. For instance:
- plant identification could potentially be solved by incorporating a gene that causes the target plant to be more visible to a machine compared to weeds - maybe the plant is modified to absorb a particular easily detected chemical that non-modified plants don't absorb, or maybe it has a jellyfish-like gene that causes bio-luminescence. Imagine a pear tree where the fruit had a bio-luminescence, it would be relatively easy to build a machine that scans for the luminescence and delicately plucks fruit until no more luminescence is detected.
- plants could also be engineered to grow in more regular ways - apple trees that are guaranteed to bud a branch every 20cm along the stem, and every branch is guaranteed to have four positions where an apple can grow. The regularity makes machine design relatively easy.
Perhaps you have already hit on the solution to the 'similar weeds' problem. A machine can filter for very specific wavelengths of light. The 'green' of two similarly green plants might look as different as black and white at some specific wavelength.
Machines are not limited by the properties of human physiology.
Here's an economist championing mass immigration of menials, while admitting that restriction of immigration into rich countries induces labor-saving innovation:
" The primary policy pursued by every rich country is to prevent unskilled labor from moving into their countries. And because unskilled labor is the primary asset of the poor world, it is hard to even imagine a policy more directly inimical to a poverty reduction agenda or to “pro-poor growth” than one limiting the demand for unskilled labor (and inducing labor-saving innovations)." quoted from Lant Pritchett via Econlog:http://econlog.econlib.org/archives/2007/05/lets_increase_p.html
Though he doesn't admit it that I know of,
there is a commutavity to this reasoning:
large enough immigration, of low enough quality, into enough rich countries, can
disastrously reverse the increase of productivity and its concomitant overall technological level.
It's not really that necessity is the mother of invention, but that investors want to use land and facilities in a more productive way, rather than share too much with labor.
Greed is the mother of invention, unless it is greed for masses of low-quality labor
which would allow those employers to hold still technologically, or even fall back,
while still keeping their expected share of value-added.