by
Patrick Crehan, Crehan, Kusano & Associates sprl, Brussels,
Director, Club of Amsterdam
TTF2035 Part 2
A new report from Ireland explains how technologies
will transform food and agriculture between now and 2035:
This article focuses on the role of robotics and the automation
of physical tasks in enabling this transformation.
See also Teagasc
Technology Foresight 2035.
The attention of the
world has been recently been captured by the rapid progress being
made in the development of the autonomous car. Almost all major
car companies now have roadmaps for the introduction of autonomous
cars, in a progressive fashion that has already started for many
companies with driver assist and accidence avoidance features,
and is intended to culminate in a not too distant future in cars
that require no driver, and which take-over all tasks that required
the intervention of licensed human beings. Google in particular
has captured the imagination of the public with the introduction
of its self
driving car project
started in 2009 and since then many others have followed.
Most people are less aware that self-driving car concepts have
a much longer history.
One of the pioneers of this domain is John
Deere
the US based maker of tractors, mowers and farm machinery. John
Deere is currently the largest operator in the world of autonomous
vehicles and you can see one of its tractors in operation on YouTube.
Others include ATC, the Autonomous
Tractor Corporation.
Vehicles such as these reduce the need for labour, or allow available
labour to achieve higher levels of productivity.
Note that no one is driving the tractor that is pulling the trailer
taking the load of corn.
Whereas John Deere
and other companies like it target large farms others such as
Kubota of Japan target the needs of small farmers, farms on rugged
hillsides with irregularly shaped fields or water logged land.
It is interesting to browse
You
Tube videos
of Kubota machinery operating in the context of fields typical
to places like Thailand. It gives an idea of how different is
the challenge in these countries compared to the giant prairies
of the American mid-west or the vast expanses of Poland. Kubota
intends to embed as much intelligence into their small format
relatively low cost machines, as is already to the big clients
of John Deere. In this way, technology progress in Asia may provide
a "frugal innovation" dividend for European producers.
IBEX Automation of the UK has developed a robot for spraying
weeds
that can navigate and manoeuvre in difficult terrain and on marginal
lands. The Lettuce
Bot
developed by Blue River Technology is designed to thin plants
such as lettuce, keeping only those that are likely to provide
the best product. Energid has developed an autonomous system for
picking
citrus fruit.
Agrobot of Spain has developed an autonomous system for harvesting
strawberries.
Wageningen University in the Netherlands has developed an autonomous
harvester for picking
cucumbers.
A team at Harvard University is currently working on an autonomous
robotic bee
capable of pollinating a field of crops. US based Vision
Robotics
has developed an autonomous system for removing weeds and pruning
vines. Its French competitor Wall-Ye
is capable of performing a range of tasks including pruning.
Wall-Ye demonstrating its ability to prune vines
Kubota
proposes another approach to problems such as the pruning of vines.
One that is increasingly well known in Asia, but almost unknown
in Europe, the use of exoskeletons or wearable robots. Video testimonies
from users of its Arm-1 exoskeleton on the website of the Kubota
Cyber Farm Machinery Square
illustrates how they are currently being used by a wide range
of users including seniors who have to carry out physical work
on a farm. This system is not intended for lifting heavy loads.
It gives relief to workers who may lack physical strength and
resilience such as the aged owners of farms in rural Japan. The
testimonies indicate that "you really have to try it to understand
the value it can provide."
In Australia autonomous
drones
and
four
wheeled robots
are being tested for herding cattle and sheep. Farmers in Ireland
have been experimenting with this as well. A recent FT article
describes how
drones lift farming to a higher plane in Bulgaria.
It is interesting to note that Japan has employed drones in agriculture,
already for the last 3 decades. They are mainly used for spraying
and it is estimated that 1 in 3 bowls of rice eaten in Japan have
been treated at some stage by aerial drone. Yamaha is the main
producer and it has only recently started to go over-seas in its
sales efforts. In the US, it recently won an exemption from the
FAA to a law banning commercial use of unmanned aerial drones.
It intends to market them both
in the US and the EU,
for use in the production of rice and wine. For now these drone
systems require human pilots, but future versions will operate
autonomously without the need for adult supervision. As cars and
tractors and other machinery act with greater degrees of autonomy,
so will drones. With the help of AI and machine learning technologies
these systems will respond to problems in real time, based on
direct commands, scheduled programs of operation or as part of
bigger more complex systems that generate alerts from cameras
and sensors dispersed throughout the fields and other places of
production.
Yamaha drone
getting ready to spray vines in Nappa Valley
The
Danish Meat Research Institute is working on the application of
robotics in meat processing. It already claims superior results
in the use
of robotics to process pork.
According to Food Quality News a Scottish abattoir has recently
made a breakthrough in the
use
of robotics to measure the "eating quality" of meat.
Scott
of New Zealand
is a world leader in the automation of work in abattoirs, in particular
in the processing
of sheep and lamb
and in the execution of complex tasks such as the removal
of meat from the bone.
Its most recent You Tube video talks of how it aims to develop
a system capable of processing
a whole lamb carcass in under 12 seconds.
Scott is not the only actor in this space and the technology is
also applied to the processing of beef. To get an insight into
what is being done in Australia it is worth having a look at this
report on the Bordertown
R+D project.
https://www.youtube.com/watch?v=6j0KELd_ZAM
The "twelve second goal" of Scott Technology
Nowadays
there are other ways of producing meat. Modern
Meadows
is a US based venture backed start-up that intends to revolutionise
agriculture by growing animal cells in test-tubes with a view
to producing both meat and leather. Perhaps this has the potential
to make a "meat eater" out of the most ardent vegetarian.
Rearing sheep is a very important commercial activity. They are
reared not only for their meat and milk but also for their wool.
They are sheared once a year. In some places, if they don't get
sheared, they are at risk of dying from heatstroke. The job requires
a lot of skill as can be seen from this short video of a sheep
shearing competition.
The fact of having to work with a squirming and often confused
live animal results in small nicks to the sheep's skin. When these
happen in the vicinity of an artery close to the surface of the
skin, it occasionally results in the death of the sheep. The goal
of "good shearing" is therefore to reduce trauma to
the animal, maximize the yield of wool by shearing as close to
the skin as possible, reduce damage to the skin from accidental
cuts, and reduce the loss of live animals. Automation is not only
as a way to increase productivity but also a hedge against a possible
shortage of people who know how to do it well. Research in Australia
was conducted on the application of robotics to shearing as early
as about 1978. These efforts were successful in that they resulted
in the development of a system that was demonstrably faster, better
and safer for the animal. However the system was never adopted.
It seems it raised many ethical and cultural issues, in particular
in relation to the difficulty of ensuring "0" fatalities
with the technology available at that time.
The cow is trained to use the robot whenever it feels the urge.
She has no need for assistance.
Traditionally,
one of the most labour intensive activities on a dairy farm has
the milking of cows. The installation of a modern milking parlour
is also among the biggest investments a dairy farmer will ever
make. Nevertheless it remains a labour intensive activity and
the ability of a farm to grow is often conditioned by the number
of cows a single labourer is able to handle on their own. Research
in the automation of milking has been going on for many decades
and companies such as Lely,
De
Laval,
Boumatic
and Fullwood
all provide commercial robotic milking systems. These systems
make most sense however only for very large farms and so adoption
of robotic solutions in milking has been limited by the relatively
small size of farms. This is an area ripe for revolution based
on new thinking that will bring the benefits of automation to
the majority of farmers, for example to those with as few as 100
cows or even less. Maybe it is time for an open source hardware
+ open source software approach that will create opportunities
for local providers competing on science and technology rather
than on hardware.
Moving away from the countryside and onto the subject of urban
agriculture, it is interesting to note the adoption of robotics
in the production of small green vegetables, lettuce, cucumbers
tomatoes and small fruit such as strawberries. Sky
Greens
of Singapore claims to be the world's first commercial vertical
farm. California based Urban
Plot
claims to be able to grow 16 acres worth of food in a 1/8 acre
plot. A joint venture involving an indoor farming company called
Mirai
and GE of Japan
has led to the development of systems that produce up to 10,000
heads of lettuce every day. They claim that lettuce grown in this
way has eight-to-ten times more beta-carotene and twice the vitamin
C, calcium and magnesium of competing products grown outdoors.
Although these systems employ human workers, a Japanese company
called SPREAD
has developed a fully automated indoor farming system entitled
the Vegetable Factory. There are already more than 200 such systems
in operation in Japan and in 2017 SPREAD will open its newest
and largest facility near Kyoto, intended to produce 10 million
heads of lettuce a year. These large scale indoor systems have
been developing slowly over the last few decades arguably since
the publication of Dixon Despommier visionary work on "The
Vertical Farm."
Vegetables are grown in a "pink houses" so called because
of the colour of the LED lights adapted to the biology of the
plant
On a
smaller scale but still outdoors FarmBot
is a 100% open source robotic system suited to small scale local
food production, suitable for small produces and people with vegetable
patches in their garden. Targeting an even smaller scale grower
we have companies like Urban
Cultivator
and Grove
Ecosystem
developing indoor garden concepts based on aquaponics and LED
lighting that grows food indoor at home or at work to deliver
fresh food all year round.
Retail
and logistics will also benefit from robotics and the automation
of complex physical tasks. At the Future Food District of the
2015 Food Expo Milan, ABB demonstrated its versatile robotic arm
capable of picking and packing fruit such as apples.
Finally
we come to the role of robotics in consumption both at restaurants
and in the home. In 2009 a number of YouTube videos appeared demonstrating
the future application of robotics in the kitchen. One of these
called FuA
Men
(short for a "Fully Automated raMen") was a robotic
noodle shop in Nagoya. It is not clear if this shop still exists,
but it featured a robotic chef and its robotic assistant. They
not only prepared ramen, but talked to customers and performed
(rather bad) stand-up comedy. Another demonstration of that time
involved a weirdly human looking robot hand that made
sushi.
Since then a group of MIT students has recently developed the
Spyce
Kitchen
concept as the basis for a new style of fast food restaurant,
and UK based Moley Robotics in collaboration with the Shadow
Robot Company
has developed a very impressive robotic
kitchen
that can emulate
Michelin starred chefs
and that has already learned by "observation" more than
2000 recipes. This opens up a whole new dimension in food related
intellectual property, as well as a new pricing possibilities
for the sale of food service systems. Moley intends to make it
available commercially from 2017.
The Moley Robotic Kitchen
We may
be on the verge of a breakthrough where automation will enable
in all parts of the agri-food supply chain. Autonomous tractors
and the smart tools they deploy can help producers or farmers
to achieve higher yields with lower inputs through greater accuracy
and a greater ability to optimise aspects of production such as
weather or time windows for optimal yield. Automation can help
with produces of meat and dairy production, the production of
cereals, fruit and vegetables. It can help in picking and packing
as well as processing, distribution and retail. It is even pretty
good in the kitchen.
So much has already been achieved, in the automation of physically
complex tasks right across the agri-food supply chain, as if by
stealth. Although in many cases the potential of automation has
been clearly demonstrated, uptake has been limited until now.
For many of these systems issues of affordability, maintenance
and reliability may need to be addressed. Perhaps a service oriented
approach is the way forward based on new forms of ownership, uberisation
or sharing and co-ownership. The coming decade should see waves
of innovation not only in the transformation of work via new tools
and systems, but in the transformation of business via new organisational
and economic models.
The Innovation Landscape
Research in agri-food related science and technology is going
on all over the world. The two issues I want to highlight here
are the role of innovation related competitions and the role of
open source hardware and software, as well as the emergence of
platforms to help accelerate development.
The FRE or Field
Robot Event
was founded by University of Wageningen in 2003. It appears to
be the longest running such event in the world. The last one took
place in Slovenia and the fourteenth such event will take place
in Germany on 14-16
of June 2016.
Nevertheless other parts of the world are starting their own independent
initiatives. The US based ASABE
agricultural robotics competition for example started in 2007.
Even more recently the Ag
Bot Challenge
starts in 2016. It encourages teams to design, develop and demonstrate
autonomous agricultural machinery that address specific on-farm
challenges. The 2016 competition focused on planting seeds. In
2017 it will focus on weed and pest identification and removal.
In 2018 the plan is to focus on autonomous harvesting.
The gantry-like Farm Bot system
Earlier
on we mentioned the FarmBot
initiative aimed at growers that have access to small outdoor
gardens. The MIT
Open Ag Initiative
is an "open source ecosystem of food technologies to create
healthier, more engaging and more inventive food systems."
One of its platform projects is the development of a desk-top
Food
Computer
that can be programmed to grow anything by simulating an approach
climate, soil and nutrient conditions. FRoboMind
is an open source robot control system that originated in Denmark.
It was developed with a view to accelerating progress in the application
of robotics in agriculture. It is currently employed in more than
10 field robot projects.
It is easy to imagine future initiatives based on synergies between
efforts such as these and those who participate in or are inspired
by other open source movements such as Hong Kong based OSVehicle
which has developed an Open Source platform for electric vehicles,
and US based Local
Motors
which claims to produce the world's first 3D printed commercial
automobile.
In this article we focused on the role of robotics and the automation
of physical tasks in enabling a transformation of global agri-food
systems. In the next one we will discuss the role of ICT, sensor
networks and big data in the automation of cognitive tasks and
the work of management.
Disclaimer:
The author was employed by Teagasc to assist in the development
of the Teagasc Technology Foresight 2035 initiative. Nevertheless
the views expressed in this article remain those of the author
and do not necessarily reflect those of the client.
See
also: Transformation
of Food and Agriculture- Part 1 by Patrick Crehan