Robotic technologies have penetrated almost every sector from the medical through manufacturing and space exploration. One sector consistently lags behind the rest, however ?agriculture.
This is slightly mystifying, because farmers have used tools for millennia and complex, heavy machinery for a century or more. Further, by embracing genetically modified foods, farmers have shown that they are not completely adverse to technological advances.
Some factors that have impeded adoption of robots in the agricultural sector include: the availability of cheap agricultural labor, the complicated processes involved in farming (which make robotic duplication a real technological challenge), a lack of research funding, low seed prices, and the somewhat complicated relationship with risk that pertains amongst the farming community.
But there is cautious optimism amongst agricultural robotics experts, because all of these factors are currently in transition.
Agricultural robotics projects are attracting government funding. Venture capitalist money is starting to flow. Labor is no longer always readily available, nor cheap. Seed –and food– prices are rising. And agribot technologies are rapidly advancing.
If these trends continue over the coming years, say experts, agribots of all shapes and sizes will become more attractive to farmers and the opportunity for widespread growth of the agricultural robotics market (in both the U.S. and overseas) could increase significantly.
So, what are the key indicators for future growth and where can robotics developers find opportunities in the agricultural sector?
One set of opportunities is found through partnerships with established agricultural equipment manufacturers. Jaybridge Robotics (specialists in systems integration for autonomous vehicles), for example, collaborated with agricultural equipment manufacturers Kinze, on the ‘Kinze Autonomy Project,’ which resulted in the development of an autonomous tractor for row crops. The robotic tractor is expected to become commercially available next year.
The ability to plant seeds with precision (and without any wastage) is more important than ever, with seed prices increasing rapidly over recent years, explains Luc van Herle, global sales manager, Kinze.
?The cost of the input has increased dramatically. Not that long ago, a bag of corn seed was costing somewhere between say $20 and $40 a bag. That same bag of seed now costs between $300 and $400. That has created the need to be 100 percent precise and accurate with how much you plant and where you plant it,? explains van Herle.
?Everybody can plant in a straight row for 8 hours. Real hardy souls can plant in a straight row for 16 hours. Nobody can plant in a straight row for 24 hours. Robots can. So it’s a timing issue, as much as efficiency and speed.?
As seed prices rise, so have values for crops like soya beans, sugar beet, and edible beans ?further encouragement for robotics developers.
An international assessment of research and development in robotics sponsored by the National Science Foundation, NASA, and the National Institute of Biomedical Imaging and Bioengineering of the United States Government, published in 2006 estimated that a total of 885 service robots (with a value of $117 million) were in use across the U.S. agriculture, forestry, and mining industries.
?We don’t know the size and value of the U.S. agricultural robotics market. It’s a bit like asking ‘What’s the market for VCRs?’ before there were such things. We’re at the leading edge of this technology with no real understanding at this stage of just how deep this will penetrate the market segment,? explains van Herle.
One of the key indicators for likely adoption of agricultural robotics is reduced availability of cheap, qualified labor. Seeding and harvesting take place within strict time-frames –up to ten days in most cases– with large farms in particular requiring a lot of labor in a short time-frame.
In 1991, 1.1 million farm workers were hired in the U.S., according to the USDA’s National Agricultural Statistics Office. By 2011, that figure had fallen to around 700,000. As qualified farm labor becomes more difficult to find, robotics solutions become more attractive.
?When a large farmer plants in the third week of April, every other farmer within a 500 mile radius is also planting at the same time. So, when the farmer is looking for labor, everybody else is looking for labor too. It’s not as much a cost issue as it is an availability issue,? explains van Herle.
Wouter Saeys, from the Department of Mechatronics, Biostatistics and Sensors at Katholieke Universiteit Leuven (KUL) agrees.
?We hear it not only within Europe, but also from people in Latin America and the U.S. – that getting seasonal workers is becoming harder and harder. Especially reliable ones,? says Saeys, who worked on a collaboration involving researchers from Flanders? Mechatronics Technology Center (FMTC) and KUL, to develop a prototype self-steering, automated tractor.
Ten years ago most Belgian fruit growers used students and retired people to come pick the fruit, explains Saeys. Following European Union (E.U.) expansion (a move that enabled Polish workers to work legally in the E.U. region) they started employing a lot of Polish labor. Today, Polish farm workers are harder to find.
?The whole logistics of getting the seasonal workers is becoming more and more tricky. So you feel that the mindset is starting to get right to invest in robots for that work,? says Saeys.
Although the conversation about agricultural robotics often centers around the availability of labor, it’s worth noting that farm workers and agribots are not necessarily mutually exclusive. In fact, opportunities exist in robots specially designed to assist farm workers.
In Japan, for example, where half of farm workers are 65 years old or more, the Toyama Lab in Tokyo University of Agriculture and Technology is working on the ‘Wearable Agri Robot’ ?an exoskeleton being developed to assist Japan’s aging farm workers with manual labor. The device is expected to go on sale in Japan in 2012, with a price tag of approximately USD$10,000.
To fully exploit opportunities in the sector, companies need to understand the demographic and farmers’ complicated relationship with risk.
Forty percent of U.S. farmers are 55 years or older?not exactly the optimum age for adopting new technologies. Further, some 90% of the 2 million total U.S. farms are family owned.
On the positive side, farms are getting larger as the sector consolidates. In 1992, for example, 62,000 U.S. farms accounted for 50 percent of sales of agricultural products. By 1997, 46,000 farms accounted for 50 percent of sales ?a positive indicator for adoption of new technologies, since large-scale agriculture requires machines to seed, irrigate, cultivate, and harvest very large areas of terrain.
Companies need to understand that farmers are some of the ?biggest gamblers on the planet,? says Kinze’s van Herle, but having already invested heavily in a risky and unforgiving business, they seek to minimize their exposure at every opportunity.
?Farmers are enormous risk takers. The risk is always going to be the weather: how much sunshine, how much rain, how many storms at the wrong time of the year and so on.’
Understandably, farmers fear the unknown effects of new devices.
?Farmers plant once a year and they don’t get a second chance. If the planting operation didn’t work for whatever reason, you have just zeroed out their income for that year,? says van Herle.
If robotics companies prove that their equipment can provide extra reliability in the midst of agriculture’s inherent uncertainty, then farmers’ fear of new technology can be overcome.
‘Farmers like to control as much as possible of the risk if they can. So what they will do is try and control the facts that they can control, which is mechanization, automation, and autonomy.?
There’s a social and cultural aspect to the adoption of any technology, particularly robotics, says Daniel Schmoldt, national program leader at the National Institute of Food and Agriculture (NIFA), an agency of the United States Department of Agriculture (USDA).
?Whatever you introduce in terms of technology has to fit in with the rest of the existing practices that the producer is using. If it doesn’t mesh well, it’s not going to help them or save them money, it’s just going to be counterproductive,? says Schmoldt.
‘In many cases, even though they might like to adopt newer technology like robotics, if their sons and daughters are not going to take over the business, maybe they won’t see the investment as worth it, because they’ll be getting out of the business: many non-economic factors come into play that are somewhat unique to agriculture.?
Farmers may appear conservative, Schmoldt concludes, but if you can show them that robotics works to their advantage, it’s possible to get past their aversion to risky new technologies.
Another positive sign indicating the likelihood of increased adoption of agricultural robotics in the future is growing investment from governments and venture capitalists.
Comprehensive Automation for Specialty Crops (CASC), for example, is a USD$12m research project funded by the USDA Specialty Crop Research Initiative with 100 percent matching funds from industry and university partners.
Established by the 2008 Farm Bill, CASC is dedicated to developing comprehensive automation strategies and technologies for the USD18b U.S. deciduous tree fruit industry and the USD17 billion U.S. nursery and landscape industry.
Project participants are working on technologies from autonomous vehicles through digital insect traps and robotic calipers.
Meanwhile, the NIFA has a budget of approximately USD5m to invest in robotics-related projects as part of the U.S. government’s National Robotics Initiative, says Schmoldt.
A call for agricultural robotics proposals closed late this year and submissions are currently being reviewed by experts at the National Science Foundation (NSF). [Schmoldt couldn’t comment on the specifics of these confidential proposals as a final decision has not yet been reached, but notes that they cover a wide range of areas from forestry through food production processing and crop production.]
U.S. government investment in agricultural robotics doesn’t begin and end with the USDA however.
?My expectation is not all the agricultural projects will be funded by USDA. I hope that the NSF will want to fund agriculture-related projects that are ranked high by the peer review panel because it’s good science and engineering. So, the amount that we have available isn’t really necessarily all that’s available for agricultural robotics under this solicitation,? says Schmoldt.
Schmoldt expects his agency to fund between 3-7 projects ?depending on how much money they ask for,? with NASA and the NSF funding other, larger projects with potential agricultural applications.
Meanwhile, in Australia, a team led by Professor Peter Corke at the Queensland University of Technology, was recently awarded almost USD400K by the Australian Research Council to develop teams of lightweight agribots.
If successful, the three-year project will see the creation of agribots with advanced navigation capabilities and the ability to work cooperatively in swarms. On-board cameras and image-recognition software will enable the agribots to work out precisely where herbicide needs to be sprayed –bringing an end to the costly, largely-inaccurate, and environmentally unfriendly methods currently employed by farmers worldwide.
Large swathes of Australia are not particularly hospitable environments for agriculture, with arable land in extremely short supply ?a positive indicator for agribot adoption in that country and in other countries where arable land is at a premium.
And in late December, the Taiwanese government’s Council of Agriculture (COA) launched an initiative to help agricultural technology companies to gain access to capital markets. To qualify for the scheme, companies are required to have spent at least 3 percent of net revenues (or 5 percent of paid-in capital) on R&D.
Venture capital money is flowing too ?at least in the case of Boston-based Harvest Automation. The company has raised more than $12 million in venture capital funding over the past few years, including investment from the Massachusetts Technology Development Corporation.
Their flagship product is a lightweight, knee-high, mobile agribot designed to assist workers in nurseries and other agricultural settings. Aimed at the North American ornamental plant market (estimated to be worth somewhere in the region of 17 billion according to the USDA).
There are also opportunities for developers to create hardware and software packages –such as computer vision and obstacle detection systems– that can be used on existing machinery and devices.
In October 2011, for example, researchers at the U.K.’s National Physical Laboratory announced the development of an imaging technology that can identify whether strawberry crops meet specific predefined criteria for ripeness. Systems like this could be implemented on fruit-picking robots to improve productivity and ensure that fruit is picked at just the right moment in time.
The most exciting trend –for both his agency and agricultural robotics in general– that NIFA’s Schmoldt has noticed over recent years is heightened interest in agricultural robotics amongst prestigious universities that have a proven track record in successful robotics development.
?Our primary R&D partners up to now have been colleges of agriculture and Land Grant universities, but now we’re getting other people involved –Georgia Tech Research Institute, Carnegie Mellon, MIT, Stanford– that normally before wouldn’t pay any, or at least very little, attention to agriculture,? says Schmoldt.
‘The National Robotics Initiative, and some of the other things that we’ve been able to do recently in our agency, has really brought a whole new mix of partners with different ideas and different ways of approaching problems. We’re very excited about these new partners.?
For Saeys, the most important trend is increased use of robotics for high value crops, like fruit and vegetables.
?There has been some interesting research, but they always stayed pretty much on the academic side. Some companies have tried to get a hold, but typically they have had trouble competing with humans. The time seems to be right for these things really to get into practice in the next 5 to 10 years,? says Saeys.
For robotic technologies to become feasible, the price of components needs to drop.
?Competition is still difficult with equipment and maintenance costs, compared to human labor. But as labor costs increase and technology costs decrease, the balance will be different. So, in some senses it?s only a matter of time.?Read More