JAXA Japan Aerospace Exploration Agency

Q. Could you tell us about your research using the first Quasi-Zenith Satellite, MICHIBIKI?

Robotic farm machine (courtesy: Hokkaido University)

Experiment in an environment where GPS signals cannot be received (courtesy: Hokkaido University)

We are working on agricultural machinery automation using positioning signals from MICHIBIKI. By confirming its own precise location, a piece of farm equipment will be able not only to drive itself but also perform all the roles it is designed for without an operator. For instance, once you input precise instructions into the control computer of a tractor, the tractor will be able to correct its route when it goes off course, while spreading the proper amount of fertilizer and pesticides. For such a robotic tractor to operate accurately, a reliable positioning system is crucial.

Q. What are the outcomes of your experiments with MICHIBIKI thus far?

We started two projects using MICHIBIKI in 2011. One is an experiment to complement the U.S. GPS satellites; with the other, we are trying to find out whether MICHIBIKI could actually replace GPS. The GPS satellites are positioned in such a way that at certain times their navigation signals can't be received, because they are blocked by buildings or windbreaks. With agriculture, though, timing is vital, because it involves living things. If a robotic farm machine cannot receive navigation signals, it cannot drive, and that's a problem. Navigation using MICHIBIKI as a complement to GPS has extended the hours when accurate positioning is possible, and we have confirmed that it provides more reliable location information.

The other project is an experiment using MICHIBIKI's unique experimental signal (known as the LEX signal), which enhances the performance of GPS. The LEX signal provides location information that's accurate to within a few centimeters, so an autonomous self-steering tractor can operate more safely. Signals complementing GPS have typically been transmitted from ground stations and acquired via cellular phones. However, rural areas have poor cell-phone reception, and some areas cannot receive these signals at all. MICHIBIKI, on the other hand, sends signals from space, so they can be received pretty much anywhere. With this experiment, we have demonstrated that MICHIBIKI can provide more accurate positioning than the conventional method.

Q. So it sounds like there are two different functions: to reinforce GPS, and to complement it. Do these two functions require different receivers?

They do. However, JAXA is currently developing a receiver that performs both functions, and that's the one we will use in our next experiment. Mind you, I hear that it's accuracy is only about 30 centimeters, so for use in agriculture it may need further improvement. In the meantime, JAXA is also developing a next-generation positioning system with accuracy of 10 centimeters or better. It's called MADOCA (Multi-GNSS Advanced Demonstration tool for Orbit and Clock Analysis). This system could be used in almost any agricultural work. The type of farming that needs the most precision is rice planting - it requires accuracy of about 5 centimeters. We'd like to combine MADOCA with other sensors, and conduct a sowing experiment with that level of accuracy.

One of the great things about having MICHIBIKI is that when we work with JAXA, there is room for our requests to be accommodated. This is the most exciting aspect for us researchers.

Q. What is the agenda for the practical application of the Quasi-Zenith Satellite System (QZSS)?

MICHIBIKI, the first Quasi-Zenith Satellite. It is positioned near the zenith over Japan for about 8 hours a day.

The plan is to establish a constellation of four QZSS satellites as soon as possible, in order to make the positioning system available around the clock. At the moment, there is only one satellite, so positioning is possible for only eight hours a day. With this limitation, positioning is not yet completely stable and reliable. In agriculture, there are optimum periods. For instance, in seeding, there is a window that you cannot miss. If you miss it, the plant won't grow well. And there is no day or night in agricultural work; it's a 24-hour operation. This is why I hope that a 24-hour positioning system will become available soon.

Q. You have been doing a lot of demonstrations for farmers. How are they responding?

Prof. Noguchi at a demonstration for farmers (courtesy: Hokkaido University)

I hold about 10 demonstrations a year of autonomous, self-driving tractors. People of the younger generation are especially interested, and many would like to see them become available as soon as possible. This has to do with the problems Japanese agriculture is facing today, with the decline in the number of farmers and the aging of the farming population. Compared to twenty years ago, the number of farm households is down by 53 percent, and the average age of farmers is now 66. In order to maintain the industry, they need to expand the scale of their farms. But if they don't have enough manpower, it is difficult to keep up. That's why there is so much expectation placed on robotic farm machinery, which could replace human farm labour.

In addition to farm machinery automation, I am also working on "IT agriculture" - a way to improve the efficiency of farming based on computerized data. This area is also attracting a lot of interest among farmers. As the problems of aging farmers and the lack of successors continue to grow, newcomers to farming are very welcome. However, agriculture is influenced by the natural environment, by things such as weather and soil conditions, all of which are things you don't know about unless you have experience. So it is difficult for beginners without experience to grow crops successfully. But if there is data available about the history of production, the crop yields and quality from specific fields, and the amount of fertilizer and pesticides used, it gives farmers a chance to succeed. For such data to be useful though, of course, you need information about time and place, so a satellite positioning system such as MICHIBIKI is very useful.

Q. Has IT agriculture already been introduced in Japan?

Receiver system for the QZSS LEX signal (courtesy: Hokkaido University)

In Japan, IT agriculture finally started spreading slowly in Hokkaido prefecture. In contrast, in the United States, Europe and Australia, most farm machinery is already equipped with GPS receivers so it can spread fertilizer and pesticides automatically. Also, driverless steering control systems are very popular.

Hokkaido prefecture is spacious and has prosperous agriculture. In addition, thanks to the efforts of the local governments in supporting farming, many people there are interested in new agricultural technology. On the other hand, the scale of agriculture in Honshu, the main island of Japan, is still small, so people may not see a reason to use GPS. But the cost isn't as high as people may think. Rice planting and harvesting are never done at the same time, so if you have one positioning-signal receiver kit, you can use it for all tasks throughout the year, simply by swapping it between different machines, such as a rice transplanter and a combine harvester. Currently, a prototype of a robotic navigation system, including the receiver, costs about 3 million yen (about 38,000 USD). But if you consider that the work done by a piece of robotic farm machinery can be approximately equivalent to that of a person, I don't think the price is necessarily high compared to an annual wage. If mass production of the system can be achieved, the price will likely go down, too. Lately, there is also a movement in Honshu to robotize farm machines used in paddy fields. When MICHIBIKI becomes available 24 hours a day, I expect that IT agriculture will likely take off rapidly.

Q. MICHIBIKI can also provide positioning in parts of Asia and Oceania. Is there interest in your work in the agricultural sector there?

Agriculture is a major industry in Asia and Oceania, so they are very interested in agricultural technology using MICHIBIKI. This year, Korea and Malaysia ran experiments using a MICHIBIKI signal receiver. In particular, they like MICHIBIKI's capability of enhancing GPS. Enhancing GPS usually requires a base station on the ground, which transmits complementary signals. But if the signals can be transmitted from MICHIBIKI in space, there is no need to build such infrastructure on the ground. This is a great advantage. In the future, it may become possible to export agricultural technology that uses Japanese positioning satellites, and to offer diplomatic support with the technology.

Q. What prompted you to start your research on farm machinery automation using navigation satellites?

Prof. Noguchi during his time at the University of Illinois (courtesy: Noboru Noguchi)

When I was at the University of Illinois, in the United States, in 1997, research on agriculture automation was already underway there. Farms in the U.S. are so big, you cannot even compare their scale to that of Japanese farms. So they were already using navigation satellites to pinpoint the location of farm machinery. This is how I became interested in the research.

In 1998, there was the first demonstration in the U.S. of an autonomous self-steering tractor. I had a good feeling about the potential of the technology, but farm machinery in the U.S. is very big and very fast, so I thought that, in reality, letting such massive machines run autonomously without drivers would be difficult in terms of safety. But I thought it could work in Japan, considering the smaller size of Japanese tractors.

Q. What is your vision for the future of agriculture with satellite positioning?

My goal is to make robotic agriculture using remote monitoring technology a reality. This is a part of the project by the Ministry of Agriculture, Forestry and Fisheries, to develop agriculture automation and assistance systems in order to reduce manual farm labor. The idea is that robotic farm machines will operate autonomously in multiple fields, while being monitored from remote offices. These farm machines require no drivers. When they run out of fertilizer, they'll stop, and someone will go and supply it. The plan is to build monitoring rooms where two people can oversee four or more machines.

As the first step, we are looking at a system in which one person is responsible for two farm machines. There is still room for improvement in terms of safety before an autonomous self-steering farm machine can operate without a driver, so for now, to ensure safety, it will be followed by another tractor driven by a farmer, who will monitor the driverless tractor while performing another task as well. This way, the farmer can do two jobs at once, doubling efficiency. In other words, we will start the project as a collaboration between a robot and a person, with the aim of increasing work efficiency, and then we will work on practical applications.

The idea of this two-tractor process was actually suggested by young farmers. With autonomous self-steering farm machines, safety is the biggest concern, because it would be a tragedy if a person got run over. There is no such thing as guaranteed safety even when using a sensor to detect obstacles, and besides, manufactures tend to stay away from machinery that involves such risk. And when we were stuck on this problem, it was the farmers who helped us with an idea. Although we are developing technology that's absolutely necessary for agricultural work, we don't work in the fields, so feedback from farmers is extremely important - there is so much to learn from them. All the more, I would like to make systems that they are satisfied with.

Today, Japan is facing the problem of rural depopulation. If people find something attractive about rural villages, it may bring them back to the land. I hope that my efforts can serve as a motivator for people to become interested in agriculture.