Precision Farming | Remember Hard Tomatoes? Well, Get Ready for Bland Lettuce
 |
| This CGI of the robotic weeding of what surely must be GMO corn was developed for a commercial featuring the herbicide Halex GT, a weed-control herbicide produced by Switzerland’s Syngenta AG. Credit: Kunochan. |
Lettuce
Be
WILL GROWERS FINALLY SHIFT TO MORE FULLY AUTOMATED
AGRICULTURE?
Devon G. Peña | Seattle, WA | January 22, 2014
I first became interested in agricultural robots around 1997. I had justread the amazing RAFI
Communique on Bioserfdom: technology,
intellectual property, and the erosion of farmers’ rights in the industrialized
world. This is not light reading and the discussion of precision farming in
that RAFI[i]
report, that I believe was prepared by the inimitable Hope Shand, is what got me
thinking about broader applications of harvest mechanization and automation. The rest of the food system was getting
automated, I remember thinking: Why not the harvesting of a broader range of
what are currently labor-intensive row crops?
Those of us who came of age in the sixties remember the story about the
mechanization of tomatoes. In the early years of the United Farm Workers of
America, when it was still an organizing committee (UFWOC), Cesar Chavez and
Dolores Huerta focused their efforts on three crops – grapes, lettuce, and
tomatoes.
The organizing of farm workers was gaining momentum despite attacks by
hired thugs from the Teamsters and other groups including police. It looked
like the farm worker union would gain recognition and start winning some battles
under threat of strike or boycott. The corporate agribusiness growers became
increasingly weary of the longer-range possibilities posed by this workers’
movement. This fear and angst came to a climax when the organizers turned their
attention to tomatoes.
Rather than engaging with the union to arrive at union recognition and
collective bargaining agreements, the growers decided they would invest in the
development of mechanical harvesters. The most popular account of this
story is a classic essay and later book by Jim Hightower condemning the narrow-interest
politics of the land grant universities and agricultural extension services: In
“Hard Tomatoes, Hard Times”
mechanization was not just about industrial efficiency; it was about
undermining the possibility of workers’ struggles; it was about power and
profit.
The California Tomato Growers Association made its first research and
development grant to UC-Davis in 1956 for the development of a mechanical
harvester system (De Janvry et al, 1981:13). This has been
described as a response not to union threats but to “labor shortages” resulting
from the end of the Bracero Program and arguments about “comparative advantage”
(Thompson 2000:52). These were contrived
shortages. The growers repeatedly turned to the always-reliable California land
grant college, the University of California-Davis. This was not new activity
for UC-Davis as it had been involved with previous efforts in agricultural
mechanization and agro-industrial chemical research in the 1940s and 50s.
 |
| An early U.C.-Blackwelder mechanical tomato harvester moves through a Tracy area field in the 1960s. Tomato vines, cut by a blade at the front of the machine, are brought into the machine on a wide belt and shaken to remove the tomatoes. |
With funding from the growers, the UC-Davis researchers and engineers
implemented crash programs to develop mechanical harvesters for tomatoes.
Within a year, they had developed the first prototype and these were
immediately deployed to selected fields. The results were a disaster. The
mechanical harvesters worked but they crushed too many of the tomatoes,
rendering them useless for sale as fresh or canned whole tomatoes. It looked
like the growers were going to have to cave to the union demands. I venture
that these changes happened around the same time that these very same growers
first marketed tomato purée (sauce) and diced tomatoes in a can.
The growers went back to the UC engineers. They had a new idea: Why not design
a new tomato, one that could withstand the stress and handling of a mechanical
harvester?
The Davis engineers returned to their labs, but this time in search of a
new tomato. In a lesser known and rather obscure study, The Political Economy of
Technological Change: Mechanization of Tomato Harvesting in California (1981), Alain de Janvry, Phillip LeVeen, and David Rumsten argue that
the quest for a new tomato was part of a “political issue that involve[d] sharp
conflicts between capital and labor as well as among fractions of capital.”
These early efforts at mechanization eventually succeeded in eliminating
harvest labor for tomatoes and resulted in larger grower operations. Between
the first steps toward mechanization (1956) and the drive toward mechanization
(1964-75), the number of farmers with tomatoes declined. In 1964, there were
about 1072 farmers with tomatoes on about 143,000 acres. In 1975, there were
only 845 tomato growers producing on 300,000 acres. While the number of farmers
declined the average size of the farm increased by almost three times – from
132 to 354 acres (Janvry et al 1981:38). The union remained incapable of
challenging this mechanization. This is an example of erosion and concentration
(of assets and power), as the ETC Group likes to conceptualize this process.
The rest of the tomato story is that these new varieties that are picked
by mechanical hands don’t taste like tomatoes. They don’t taste like anything
much at all, actually. Wet cardboard, perhaps? But the central tragedy of
mechanization is that it devalued skilled labor by reducing farm workers to so
many “substitutable inputs of production” and thus forcing many into deeper
poverty and deprivation. The workers may yet get their alterNative world (forthcoming blog)
In the 1997 report, RAFI
defined precision farming in the following terms:
Precision
farming…is a bundle of new information technologies applied to the management
of large-scale commercial agriculture, mostly in the industrialized world.
Precision farming technologies include: personal computers,
satellite-positioning systems, geographic information systems, automated
machine guidance, remote sensing devices and telecommunications. Various
combinations of these tools will enable the gathering of unprecedented levels
of information about every square metre of the geographic area to be
cultivated...Though still in its infancy, this type of computer-driven
automated application will be possible for all kinds of inputs – fertilizers,
herbicides, pesticides, etc. (RAFI Communique 1997:5)
 |
| The Weedy robot from the Faculty of Engineering and Computer Science University of Applied Sciences Osnabrueck, Germany. Source credit: Unibots. |
Since 1997, these
developments have been further perfected in designs realized but not generalized. There is
a journal that specializes in Precision
Agriculture. A recent article in the journal focused on “Agricultural
Robots–system analysis and economic feasibility” (Pedersen, et al. 2006). According
to this study, one development – entirely associated with the advent of the
IT-based systems that RAFI predicted would become a dominant force – is that
researchers and designers are focusing not so much on what might be
characterized as the servo-mechanical dimensions of design including things
like hydraulics, and are instead focused on the complex algorithms it will take
to get these machines to conduct their semi-autonomous operations: “There are a
number of research projects dedicated…to…sensing systems and navigation
algorithms” but oddly these are so far limited to three types of systems: “...field scouting in cereals; robotic weeding in sugar beet, and grass
cutting on golf courses” (Pedersen, et al. 2006:296).
The study also suggests
that a major limiting factor in the widespread adoption of these robotic
technologies to a broader set of row crops is, quite simply, cost: “The high
cost of real time kinematics Global Positioning System (RTK-GPS) and the small
capacity of the vehicles are the main parameters that increase the cost of
these robotic systems.” (Pedersen, et al. 2006:295)
However, some people think
all this might finally be changing. This past summer the Associated
Press (July 14, 2013) published a report about the so-called “Lettuce Bot,”
which is basically a semi-autonomous machine. It is actually rather bizarre: This
robot is used not to harvest the crop but to help with thinning and weeding as
part of crop quality control.
 |
| The Lettuce Bot being prepared for a thinning operation. Unlike other agricultural robots, this is not a semi-autonomous machine. The current model must be hitched to a tractor PTO. Credit: Townhall.com. |
The machine basically uses
pattern recognition technology (artificial vision), while being remotely
piloted from a laptop computer, “to identify which lettuce plants to eliminate
with a squirt of concentrated fertilizer that kills the unwanted buds while
enriching the soil.” This expensive machine, developed by a Mountain View-based
company known as Blue River Technology, is meant to replace farm workers who,
with the skilled use of long hoes, usually thin out the crop to create more
room for the growth of the right-sized heads of lettuce. The AP report also
notes that another company, San Diego-based Vision Robotics, “is developing a
similar lettuce thinner as well as a pruner for wine grapes.”
The designers of this Ag
Bot claim it “can ‘thin’ a field of lettuce in the time it takes about 20
workers to do the job by hand.”
 |
| The agricultural robotics industry is trying to solve the challenge of grasping. Here is a strawberry harvester being developed by Robotic Harvesting, LLC. |
However, the same problems
that faced earlier technologies – as far back as the days of the earliest
tomato harvesters – have to do precisely with something that is actually quite
simple, elegant, and human: Grasping.
Most of these robotic systems do not yet have a very gentle grasp and this involves
limits related to much more than just servo-mechanical design.
This is really more about
mind and intelligence: It goes to the center of what it really means to be a
human being and a farm worker – the melding of brain with hand; the unity of
the mind’s eye and grasping fingers, the motions of which are so elaborately
played out at near speed of light through our neural and bodily nerve networks
and yet so easily executed in a physical action with a great deal of precision
and purposefulness. Machines can’t seem to replicate this, no matter so far the
sophistication of the algorithms that are to remote control the
servomechanisms.
For this reason, a great
deal of research is being done today precisely on the problem of robot grasp.
There is even a journal dedicated to this field called, Grasping in Robotics: Mechanisms and Machine Science. Writing in
this journal, Rodríguez, Moreno, Sánchez, and Berenguel (2013) note how
The
automation of agricultural tasks has been widely developed in recent years due
to labor requirements and its high cost for growers. The majority of
agricultural tasks is of the “pick and place” type, which has led to
robotization’s coming onto the scene as a solution for growers. In turn, they
have witnessed how robots have begun to work on their farms or in their
greenhouses. A foundation on the agricultural tasks is shown, as well as major
developments to produce them. For this, the tasks, where they are used, are
explained and the main organoleptic characteristics are set out as a basis to
consider for future development. Within such developments, this work is focused
on the grasping end-effector used in agriculture. (2013:385)
A note to my readers
before your eyes glaze over: This is actually a very disturbing development, if
you believe that the lack of a human touch in farm work is an ethical problem
with serious social, political, cultural, and ecological consequences. I will be
exploring these consequences in subsequent posts as I continue to dig into this
dystopian futureworld of agricultural robotics.
For those of you still
with me: According to the USDA, organoleptic
properties are “the aspects of food or other substances experienced by the
senses, including taste, sight, smell, and touch, in cases where dryness,
moisture, and stale-fresh factors are to be considered”. If the engineers and
capitalists that finance these technologies can figure this out, watch out.
 |
| ASME promotes development of robotic harvesters. |
Sources cited
De Janvry, A., P. Le Veen,
and D, Runsten 1981. The political
economy of technological change: mechanization of tomato harvesting in
California. Working Paper. Bib.
Orton IICA / CATIE. (Print)
Pedersen, S. M., S.
Fountas, H, Have, and B. S. Blackmore 2006. Agricultural Robots–system analysis
and economic feasibility. Precision
Agriculture 7:296-308.
Rodríguez, F., J. C. Moreno,
J. A. Sánchez, and M. Berenguel 2013. Grasping in agriculture: State-of-the-art
and main characteristics. Grasping in
Robotics: Mechanisms and Machine Science 10:385-409.
[i] RAFI (Rural Advancement Foundation International) is now the ETC Group which stands for Erosion,
Technology, and Concentration. It has long been one of the most widely respected groups maintaining a stalwart presence in global social
movements for conservation and sustainable advancement of cultural and
ecological diversity and human rights. According to a mission statement the
organization addresses “the socioeconomic and ecological issues surrounding new
technologies that could have an impact on the world’s poorest and most
vulnerable people.”
Comments
Post a Comment