Hivemind

There have been signs that automation is becoming a bigger part of our daily lives, with 10 million driverless cars predicted to be on our roads as early as 2020 as just one example of how an entire industry is about to be turned on its head.

But perhaps one of the greatest beneficiaries of these new innovations is one that wouldn't be considered straight away. The agriculture industry has seen a seismic shift since mechanisation took place and now the industry is set to be turned on its head again as further research is going into AI. But in order for AI in agriculture to be achievable, some experts are looking to bees in order to learn more about harvesting crops, pollination and limiting the impact of diseases. 

In the Economist published book Megatech Geoffrey Carr, the Economist's Science Editor, talks about what the world of farming could look like in the far off year of 2050. In it, Carr describes how drones can be used to monitor and maintain crops.

The world of tomorrow that Mr Carr talks about in Megatech sounds like the work of science fiction. After all, if you were to ask the average person in 1967 what today would look like they may also be as optimistic about the progress we would be making. But we may not have to wait until 2050 for Mr Carr's vision of our world to become a reality. The technology is already on the horizon and we could be seeing these drones and satellites being implemented in industrial agriculture as early as the next decade. As driverless cars are already being tested on some roads, driverless tractors are not far behind and the potential investment that can be made in autonomous machinery have been turning heads in Big Agriculture.

But this all begs the question, how is it possible for technology to not only navigate through fields but also identify different crops and other anomalies such as weeds and diseases? It turns out that this technology can be achieved by taking a closer look at the humble honeybee and research is already being conducted all over the world to see how we can turn the bee's brain into a multi-billion dollar industry.

Before looking at the technology capable of imitating a honey bee, it is important to first understand how the bee's brain works and why it is more suited to the tasks a robot is capable of performing rather than the human’s brain. Professor Andrew Ratnieks heads up the Laboratory of Apiculture and Social Insects (LASI) at the University of Sussex and he explains what humans can learn from bees.

"Human society is based on cooperation and mutualism but we do face some of the same challenges as bees and this is where we can learn because we won't necessarily be able to say we will be able to make human society better by becoming like bees but we can say well, bees have faced various challenges that we also face and we can learn from them. Not copy them, but get an insight. One of the things we can get insights on is reducing conflicts managing diseases and coordinating complex systems.

"Many modern human systems like the internet or traffic systems are bottom up and are composed of many interacting units like cars or electricity stations and that's how insect societies operate. Nobody is in charge. There is no leader bee but they coordinate their activity within the colony in a self-organised approach. So those are systems that humans have started to produce in the last few hundred years and especially interacting complex systems that we really don’t understand so maybe we can get some insights."

Looking at the beehive gives us an understanding of managing these complex systems, but if we narrow this down to the individual worker, we can get a better understanding of what a single drone is capable of.

"A bee is a wonderful thing. An individual organism does so many things and any time a bee is doing her own thing, they have the ability to learn very quickly. Individual bees possess huge abilities too and I think one lesson with navigation, for example, are that they have several mechanisms with which to do that, they learn landmarks, visual signs of their colony. They also can navigate using the position of the sun. Obviously with our human machines we can make optical sensors which are much more sensitive than an insect's eye in fact we can leave the insects behind to some degree in our hardware. I suppose it’s about how we can use that information that we can learn from them."

The navigational techniques and solutions that Professor Ratnieks talks about are well underway as researchers are attempting to mimic bee’s work patterns. One thing honeybees are most known for is pollination, which many crops rely on in order to flourish and survive. 

A third of our food and 90 per cent of flowering crops rely on pollination, so being able to replicate this process in an industrial scale could not only feed us but improve crops as well. In Japan a researcher, Eijiro Miyako from the National Insitute of Advanced Industrial Science and technology developed a miniature drone that was capable of pollinating flowers, which could be the first step in the next generation of pollinators and crop watchers. By using a non-harmful adhesive material on the bottom of the drone, Miyako-san was able to mimic the pollination patterns of honeybees. Artificial pollination can be a time-consuming and costly process for humans to do by hand, so creating robots that are capable of the same simple can help improve efficiency in the process. He hopes that this new technology can be mass-manufactured in the near future. "So far, I could achieve the plant pollination by flying robots with sticky gels. It's is very simple idea but it is a world-first demonstration for flying robotic pollination," according to Miyako-san. “It can up-scaled for mass production. Actually I have the best protocol for making [ionic liquid gels]. That is also very simple."

Even though this newly developed technology is still in its infancy, Eijiro Miyako is well aware of the potential uses that it could bring.

"I believe it will possibly revolutionize agriculture although the research has just been started. Personally hope this idea and technology will help to motivate somebody to create more crazy ideas for pollination and other applications. That's my dream as a scientist."

The research that was conducted within a controlled environment, however and people were quick to state the technology would be useless if it were not able to function in an outdoor environment when factors such as wind need to be taken into account, which Eijiro acknowledged.

"I'm trying to control my new drone on a farm. For now, my targets are apples, pears, tomatoes and strawberries. I'm waiting for the approval and season for these experiments to take place. I'm also developing other materials for more efficient pollination."

While this technology can be considered a breakthrough in drone-assisted farming, there is one fact that should be acknowledged. The drone that Miyako-san has developed has to be piloted by a human with an immense amount of precision. Until the technology allows for artificial intelligence capable of controlling drones over crop fields, there needs to be a focus on sustaining the pollinators that are capable of self-piloted flying themselves. Eijiro Miyako sounds hopeful about the fast development of artificial intelligence.

"It may take a long time, especially for developing sophisticated AI in the robots to control them autonomously. Ten years, perhaps? I don't know... But everybody, in particular, robotic scientists are doing their best for the future. It's surprisingly their research speed is so fast. So, I'm sure it will be coming soon in ten years."

In fact, we might be seeing this technology as early as 2022. 

At the University of Sussex, the Brains on Board project is developing technology to try and recreate the navigational functions of colonial insects such as honeybees and ants. Dr Andrew Philippedes is a senior lecturer at Sussex and is heading up the project.

"[Brains on Board] study the visual learning of insects and in particular ants and bees. The thing that we look at with them is how they manage to go out, find food and navigate back long distances with a very small brain, how they do it well and to look at a route navigation."

Like Eijiro Miyako, Dr Philippedes sees the potential of incorporating autonomous technology based on colonial insects into agriculture. The Brains on Board project has only just started and is being funded for the next five years but Andrew Philippedes work has already helped create practical solutions for problems faced in agriculture in China, where soil pollution is affecting one fifth of farmlands, using this navigational technology.

"We got funded to do some work looking at automating agriculture the focus was China so what we were looking at was a robot that would go round and repeatedly collect soil samples from the same place, so they used my algorithms to navigate and collect soil samples because soil health is a really big issue in China."

“I think Agritech is going to be the biggest use of robots in the next ten years. Absolutely. There's a lot of money going into it. Lincoln University is another big leader in this area and they've just invested a lot of money on a new centre there and I think its going to be huge.”

There is a lot of enthusiasm for the future of robotics in agriculture, but how does all of this apply to how bees navigate? Dr Philippedes did his postdoctorate on insect-inspired visual learning strategies at the Centre for Computational Neuroscience and Robotics and explains how technology has come far enough to mimic honeybees' ability to identify different plants and states.

"There's [autonomous] tractors that already exist and are used but people are already looking at how we would have deployed that in drones. People are already using drones for monitoring and I think that is the most obvious use right? Next thing people want to look at is the ability to pick different fruits and there's a lot of autonomy in that. Recognising fruit in the first place but then being able to manipulate it and being able to cut it. I think that is achievable in five years easily. A lot of it is already happening."

While the project may seem to solve a seemingly simple problem with an overly complex solution, one of the goals of Brains on Board is to make these solutions as cost-effective as possible by using components that can be bought off the shelf, which should encourage the adoption of autonomous technology throughout the agriculture industry in the next few years. 

Alex Dewar is a postdoctorate researcher on the project with Dr Philippedes and has a background in behavioural psychology and computer science. 

"The thing that we are specifically focusing on is trying to build effectively a robot honeybee that can imitate certain features that a bee can do, in particular navigation so yes at the end of those five years we will have an autonomously navigating drone that was operating in a honeybee like fashion."

As Professor Ratnieks stated, humans are capable of developing sensors that are much more sensitive than a honeybee's. But a drone with sensors similar to a bee’s might be more effective:

Honeybees have become a big topic of conversation as their evident decline in the United States has started a debate about the overuse of pesticides. A petition from Avaaz was even posted in May of 2013 and has now garnered over 4.4 Million signatures in time for a rumoured total ban of all bee-harming pesticides that was leaked to the Guardian last month. According to Mr. Dewar, we can use the bee's navigational systems to help save them, almost as much as they help to save us. 

"Robots in agriculture can mean you can be more selective of pesticides and herbicides for example people are trying to build robots to carry out weeding. If you can get rid of them without blasting them with chemicals then that has benefits. The potential benefits of robots in agriculture are pretty enormous."