July 5th, 2017 | PRODUCT , PEOPLE

How far can we evolve? A report from Hokkaido on the forefront of agricultural robot research.

With sixty-five percent of the working population over 65 years of age and an average age of 66.8 years, agriculture in Japan has pressing needs. Its workforce, made up largely of aging professional farmers, faces a challenging shortage of human resources. To address this, "agricultural robot" research is progressing throughout the entire country.

The Ministry of Agriculture, Forestry and Fisheries has set objectives for achieving the commercialization of autonomous driving systems on farming lands by 2018 and unmanned remote monitoring systems by 2020. The future is rapidly approaching. So, how far has the frontline of agricultural robotics research progressed?

Professor Shin Noguchi of the Agricultural Research Institute, Hokkaido University Graduate School, took us on a guided tour of the verification test site. We asked Professor Noguchi his views regarding the possibilities of robotics for advancing the future of agriculture.

Shin Noguchi
Professor Noguchi was born 1961 in Mikasa City, Hokkaido. He is a Professor at the Agricultural Research Institute of Hokkaido University’s Graduate School, specializing in Agricultural Information Technology and Agricultural Robotics Engineering. He is also the Program Director of the Cross-ministerial Strategic Innovation Promotion Program (SIP) "Next-Generation Agriculture, Forestry, and Fisheries Industry Creative Technology" and a Cooperating Member of the Science Council of Japan, Board Chairman of the Japanese Society of Agricultural, Biological and Environmental Engineers and Scientists.

See the future of agricultural robots at Hokkaido University!

Approximately one-third of Hokkaido University's vast campus is occupied by two farms. On this occasion, we visited "Farm No. 1," which covers an area of about 35 ha. Noguchi showed us three kinds of "unmanned robot technologies for agriculture," which Noguchi's research laboratory is promoting. We'll begin with a demonstration of the "Multi-Robot (coordinated robot)."

Larger farms through smaller tractors
The world's first "Coordinated Robot Tractor"

Four robot tractors are standing by in a field. The operator is a graduate student holding a tablet. By pressing the Start button on the screen, one-by-one the tractors start moving. Before long, four tractors begin plowing the field while maintaining a uniform distance. Naturally, all the robot tractors are unmanned. Equipped with high-precision GPS receivers, the tractors travel while performing different preprogrammed tasks accurately to within 5 cm. Multiple units seamlessly carry out combined work processes by "coordinated" understanding of the progress of and distance from other units.

In this manner, these represent the world's first unmanned robot tractor technology united in a coordinated formation. In addition, being programmed to work unmanned, each robot is equipped with a coordinating communication system. In theory, coordinating any number of units is possible.

Although no one was riding the tractors for today's demonstration, normally the operator rides on one of the units, monitoring the work of each tractor and, if required, operating the tractors. The most difficult technique is making them turn without colliding. At the present stage, if a seamless turn is judged to be difficult, they must stop and turn in sequence.This leads to downtime. In fact, one of our current areas of focus is to reduce this downtime.

Going and returning unmanned!
State-of-the-art "automated guidance" also holds great potential for night work

Note: Playback is at double speed

Next, we were shown a demonstration of robot tractors that were programmed to work unmanned and autonomously from the agricultural shed to the field. Although agricultural robots optimized for farmland work are at the commercialization stage, because this research originated to save labor, the theme is to also fully understand automation outside of the field. To achieve this, certain nontechnical issues remain. However, as you can see, the technology itself is already pretty accurate!

They go to the fields, work, and come back again. Up to this point, robot tractors capable of autonomous, automated travel are currently not sold by any manufacturer world-wide. Because these are useful during agricultural seasons when night work is required, there is a lot of interest among farmers. Besides automation in the field, where coordination is being developed, we are also aiming for automation of movement between fields. It can be said that this is also one form of the future of agriculture.

Equipped with front/rear/left/right obstruction detection sensors. The installation of safety systems, such as sounding an alert, slowing down, and stopping if an obstruction is detected within a specified distance, has progressed. However, to put these into practical use, many hurdles remain, such as negotiating sinuous farm roads and overcoming issues with traffic laws, etc.

“Drones" for converting growth conditions into data
Improving agricultural productivity from the sky

Drones are now completely familiar. Their introduction into the agricultural sector is no exception. Besides their use for pesticide spraying, which was formerly done by agricultural helicopters, attention is focused on using them to obtain aerial image data. At Noguchi's research lab, cameras and "rider" surveying components are installed onto the drones. Here, researchers are working on gathering image and height data and converting it to 3D to ascertain crop growing conditions. Soaring more than 100 meters high, the drone travels back and forth above the field following a very accurate flight path.

The drones fly completely automatically with preprogrammed flight speed, altitude, and flight path installed in the drones. Maneuvering the drone by manual operation makes them unstable. Naturally, they are dependent on the flying conditions and are impacted by wind, however, having the capability to follow a highly accurate programmed flight path offers a distinct advantage over manual operation.

With the aging, dwindling professional customer population, issues of converting to larger-scale agriculture and monitoring larger areas can be addressed by drone technology.

“Automatic Pest Control Boats" for accurate automatic pesticide spraying
Compatible also with large-scale rice fields

Land, sky, and next comes water. At the Noguchi laboratory, work is being carried out to automate radio-controlled boats for pesticide and herbicide spraying. Programmed navigation routes are installed in these normally hand-maneuvered, radio-controlled boats. Again, this research aims to save labor by leaving the work to the robots. Although this research has only just begun, it looks as though it will be in demand.

The main use of the radio-controlled boats, which are currently used for pest control, is in large rice fields. To avoid trouble, such as crashing against concrete dikes, etc., the operator must maneuver the boat by walking with it.

Therefore, although the boat can travel faster, it has to move at a human pace. If the boat can navigate unmanned and autonomously, the speed of the boat can be fully utilized. Therefore, this allows farmers to perform other tasks, such as grass-cutting, etc., while watching the boat from the dike. Autonomation and automation help extract the full capability that the technology holds, although there is some coordination with humans.

How far has the technology come? “Current" agricultural robotics

Agricultural robots for land/air/water. Although these are works in progress, the future impact of agricultural robots feels very real. It has been 25 years since Professor Noguchi started his research on agricultural robots. Today, just exactly how far has agricultural robot research progressed? To clarify the current situation, we asked Professor Noguchi about his forecast for the future.

— Although we’ve asked you before, how far has agricultural robot research progressed with respect to the ideal scenario?

Presently, our agricultural robots can still only perform simple tasks. Therefore, the next step is to use AI to transform the agricultural robot itself into a smart robot. For example, systems that facilitate optimal fertilization and spot pesticide spraying, using the drones shown here to interpret crop growth data. These can be made smart using IoT (Internet of Things) by further combining the data gathered by these robots with weather data acquired by earth observation satellites, etc. Ultimately, the expert knowledge that has been cultivated by the farmers themselves will be transferred to the robots. I think this is one picture of the future.

— Indeed. Robots will become smarter. Are smart robot R&D activities being promoted?

Of course. The affinity for IoT using robot technology and big data is very strong. By using robots to collect information, the robots can develop while doing work based on the amassed data. That said, because only one to two farming tasks can be performed each year, it will take a very long time to be able to convert all that know-how into big data. To make practical use of AI in the short term, it would be necessary to control optimal tasks by converting considerably condensed information into codified knowledge.

In addition, incorporating multiple functionality into robot agricultural equipment is another challenge. Another idea is to add harvesting and carrying of heavy objects to the original function of tractors used for tilling, planting, and fertilizing seeds. Using the current robot tractor as the feet, we can also robotize data-linked attachments that become the hands to give them intelligent dexterity. For example, using drone-like technology, AI is also very effective at performing tasks while screening, because the state of the crops can be identified from images.

— Agricultural robots will become smarter, gather agricultural crop data by themselves, and expedite tasks while interpreting the data. How will agriculture change once the future of smart agriculture becomes commonplace?

As sensing and data analysis technology evolve, it will become possible to precisely determine differences in cultivation states within the field and detect signs of pests that are unobserved by the human eye. Accordingly, the productivity per capita and yield per unit area will increase and, simultaneously, we will be able to save material input, such as fertilizer, etc., and make farming more efficient. By alleviating the labor shortage and inheriting the wisdom of skilled farmers via data, robot technology will contribute significantly to establishing agricultural sustainability and halting the decline in self-sufficiency rates.

Presently, we are considering creating fields for performing unmanned agricultural verification testing by cooperating with autonomous communities within Hokkaido. Where there are closed spaces, we can freely conduct verification tests for inter-field movement and remote monitoring robots. The pace of development will quicken by involving different industries, such as manufacturers of agricultural equipment, of course, and vehicles, etc., and by being able to conduct as much testing as possible under unrestricted conditions.

For those of us in Japan who are currently involved in this research, we also view this as a very significant opportunity. The trend for farmers is toward larger farms as the number of farm workers in all countries is decreasing. However, if we attempt to increase the size of tractors, farming will become extremely expensive because the farmers will have to replace all their working machinery. Due to issues of safety, large tractors are difficult to convert to unmanned operation. Another problem is that the soil will not be able to support overly large tractors. Consequently, the realistic option is to have small tractors converted to unmanned operation with technology for coordinated movement. In the future, there may also be a significant shift to an overseas large-scale agribusiness model that will allow Japan's small robot technology to sweep the global market.

Unmanned work by remote monitoring, coordinated work by multiple robot tractors, and work optimization by smart robots... The agricultural robots we saw at Hokkaido University certainly gave us the feeling that the age of robot farming will soon be upon us. It was also very exciting listening to the prospects for the future.

Troubled as it is by labor shortages and faced with fewer young people entering the profession, will the agricultural robot be worthy of the title of "Savior” by revitalizing Japan's agriculture and enhancing international competitiveness in an industry facing significant challenges?

Related Articles

Will the robot be our Savior? The future of agriculture changed by technology