Ag Tech Sunday - Covid-19 and the rise of the robot in food production

By Aidan Connolly - Chief Executive Officer at Cainthus, President of AgriTech Capital

See the original article here

In the midst of a pandemic, it may seem evident that the time has come to automate agriculture, remove the human-intensive component from the food chain, and modernize food processing. Perhaps, but the last 6 months have just accelerated trends that were already happening. Covid-19 has forced the closure of dozens of food factories, where workers are forced to work cheek by jowl, in refrigerated factory spaces designed to maximize productivity, food safety, and cost. Not surprisingly European operations, where robots are more the norm in meat processing, for example, business has not been stopped to anywhere near the same degree by the coronavirus.

So much has been written about robots, it might be assumed that we all are deeply familiar with what they are. But are all machines robots? What about drones? How does artificial intelligence fit in?

Even a decade ago, we might have imagined robots in the form of humanoid-looking machines, with funny voices and odd mannerisms. Sophia is impressive– well done Hanson Robotics, but she is certainly one of a kind! We see robots more and more, not only in the food and farming industry but in our grocery stores. They serve very practical uses and as the technology advances the costs come down; We can only expect to see their presence increase. Business Wire estimates the robotic agriculture industry to reach $20.6 billion by 2030. Clearly there is opportunity here, but the question remains: What are they truly capable of?

What is a robot?

Robots have many definitions though for purposes here, “any machine or mechanical device that operates automatically with humanlike skill” is well suited. A robot is a machine, but a self-guided, independent machine that is capable of making decisions, generally aided in this by artificial intelligence. Robots have to incorporate other technologies to do this, such as sensors, cameras, GPS. Machines, on the other hand, can do repetitive tasks and other work but are guided by humans or a program to do so. Therefore, a robot is a machine, but not all machines are robots.

There are three laws of robots defined by Isaac Asimov of Boston University, and these still guide robot development:

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.

2. A robot must obey orders given to it by human beings except where such orders would conflict with the First Law.

3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

Why are we so excited by robots?

Robots are fascinating for a good reason – they are doing some really cool stuff! From serving us dinner or providing a comforting hug. Dog-like robots play soccer and pull trucks. Sheer strength allowed one of them to haul a three ton plane. If this is what robots can do now – think about what they will be capable of in another five or 10 years.

Robots are extremely versatile. Take the exoskeleton robot, which enables one person to do the work of three. Developed first with military uses in mind these suits are primarily marketed for construction, manufacturing, and other labor or lifting-intense occupations. Applications in agriculture are still somewhat limited, though one Tokyo company is certainly thinking about it.

Their strength has now even been matched by their agility, as this robot demonstrates parkour. It is only a matter of time before they are fighting our wars. Their versatility is as wide-ranging as their size, which can be extremely tiny – less than 1 mm – and someday maybe fighting bacteria and viruses within our bodies.

Doctors won’t be injecting our bodies with disease-fighting robots any time soon, but self-driving cars are already a reality, and they are a genuinely life-changing robot, both exciting and frightening in equal measure! These cars use LiDAR (Light Detection and Ranging) to create a “point cloud” or visual map of the car’s surroundings by pulsing light outward and measuring the time it takes for the light to reflect back. The car can do this thousands of times every second. It’s like sonar, but much faster. (This technology underpins all of the applications discussed here, allowing robots to navigate their environment, fertilize fields, find weeds, etc)

At this stage, robots can pretty well go anywhere a human can, and in some cases, places humans cannot, such as environments where size or extreme temperatures may restrict humans. While space and deep oceans are the most obvious, most of us have no issue with robots performing such tasks. The concern to many of us is, how robots may replace millions of jobs and it is estimated that by 2030, 800 million jobs worldwide could be automated. According to Joel Stonington, senior producer at media company AJ+, when compared to agricultural industrialization, robotic automation is moving on a far faster timescale.

Robots and the food industry

The food industry has challenges in finding workers, since many of the key repetitive jobs are not one people enjoy, or aspire to do, such as processing plants or picking in the fields. Government labor regulations and laws controlling immigration are also driving the costs up for food producers. Robots are a perfect solution. Certainly, upfront costs can be discouraging, but in the long term, the advantages of having such a “workforce” outweigh those investments. Many producers feel they are not able to afford the technology – yet – but it’s a natural curve that robot manufacturers will solve as the manufacturing prices come down.

Covid-19 has been the perfect storm and while the concern might be that robots will be taking jobs from people, in many instances when it comes to the food industry people are no longer available to fill those roles. From harvesting to processing, to packaging, robots can fulfill and perform tasks as fast or faster than humans, and often with better results in terms of safety and efficiency.

Livestock production

From milking cows to sorting eggs, the redundancy of these tasks is perfectly suited for robotic automation. Some believe that these applications will actually save agriculture, such as the dairy industry, which is struggling to find the labor necessary to operate. Robotic milking and feeding, particularly when the whole system is interconnected with information coming from sensors and now camera vision from Cainthus, means that the milking system can tell the feeder what nutritional supplementation is necessary to improve the health of the cow, increasing yields and making the system sustainable.

In pharmaceutical medicine, robots are helping to manage the dairy herd’s health: Pharmbot has built a system to maintain the herd’s vaccination schedule effectively and efficiently, scanning each cow as she passes and either administering the shot or allowing her to pass through. The farmer is only alerted if there is a concern, and this is done in real-time.

Robots are fast making their way into poultry houses. Sensors and AI help robots detect their surroundings and navigate around objects such as chickens and feeders. They’re cleaning up the barns, collecting eggs, managing litter by aerating the ground, reducing the build-up of harmful bacteria, fungi, and ammonia. As they carefully drive around the poultry house, they keep the chickens moving. All of this improves the health and well-being of chickens, reducing the need for antibiotics and other synthetic measures used to maintain a healthy flock.

In the beef industry, robots are working 24 hours to watch over herds and are capable of covering all types of terrain. Slaughterhouses seem a perfect fit for robots, the work is hard and redundant. Just five years ago, it seemed robots probably wouldn't be really able to replace slaughterhouse jobs. Companies like JBS thought robots didn’t have the finesse to extract the most meat without cutting into the bone, and humans were more suited to the task and could be sourced cheaper, even if the turnover was high. Fast forward and Scott Technology now has a meat processor using smart technology (AI) to butcher up to 600 lamb carcasses an hour. Does it really matter if a little meat is left on the bone when a robot can churn out product that quickly? Perhaps that’s why JBS invested in the company.

Soft robots using compressed air inflation is another area that is advancing and can have a positive effect in agricultural production. These robots can employ a sensitive touch. In food production, they can delicately pick up pieces of meat and seafood for packaging quickly and efficiently. These robots can combine this touch with strength to complete difficult tasks at a speed that humans could not attempt. In the veterinary field, robots are being designed to lift animals, particularly injured ones that cannot walk.

Crop Production

From weeding to spraying, there is no shortage of opportunities for robots in crop production. The work is often repetitive and laborious. The challenge for robots is when the task requires a gentle touch, although soft robots will soon solve that. Hands Free Hectare, Small Robot Co, and Farmbot are examples today of crop automation but robotic farming also enables affordable intensification of agro-ecological models such as multi cropping in the same field.

Because of its inherent simplicity of movement, one of the most applicable opportunities for robots in crop production involves spraying. Through the use of GPS and a camera to guide the robot through rows of crops. Sensors identify and precisely aim weed killer on weeds, saving crops from contamination, and farmers from excessive costs in herbicides of up to 90 percent. Ecorobotix takes it a step further and runs on solar energy for up to 12 hours without a human operator. The same can be done for fertilizing, one of the largest costs for row crop farmers. These robots find the plants that need to be treated or the soil analyzed. Through algorithms, the robot “remembers” the location and can revisit it the following year. Interestingly Tesla, the main seller of car with autonomous driving, do not use LIDAR, believing triangulation using normal sensors is a better tech.

Humans have been gathering food since the beginning of time but building a robot with that same gentle touch doesn’t come easy. Weeding, picking and sorting, are applications that require a delicate “hand,” particularly in high-value crops such as fruits and flowers. Designers and engineers have worked for decades to find the exact level of precision and delicacy required to perform such tasks. This has been made possible through the development of soft robots, which use air compression sacks to gently remove the berry from the bush. Driscoll’s and several others in the strawberry industry have supported startup Harvest CROO, a robot designed to pick a plant in 8 seconds and can replace more than 30 human pickers. Those in the industry have seen serious challenges when it comes to labor, even watching fruit rot on the ground because there was not enough help to bring it to market.

Add in concepts like quality control, using artificial intelligence (AI), robots are able to pick and sort fruits and vegetables (even lettuce!) at the same rate, if not faster, than humans. Robots don’t tire or request time off either. Given the setting of most of these tasks, powering these robots with solar energy makes good sense and even further reduces the costs or production.

Is a drone just a flying robot?

Unfortunately, the answer isn’t a simple one. A drone’s history begins with the military, but it wasn’t long before civilian uses came into play. Also called UAVs (Unmanned Aerial Vehicles), drones can either be controlled remotely, programmed to follow a set of instructions, or use sensors and GPS to guide themselves which, like robots, requires some set of expectations for operation. There are many applications for drones in agriculture and often they coincide with a similar counterpart from terrestrial robots. Still, the opportunity to gain insight from the air is a clear distinction. Drones are used to map fields, determine crop health issues or irrigation problems, find pests, spray fields, and check livestock. There are so many uses that agriculture has really embraced this technology.

What will Robots do next?

Robots are indeed advancing quickly. Part of this is thanks to systems such as the Robotic Operating System (ROS) which allows open collaboration among robot software developers. Think of the challenges in programming robots for some of the applications above. Through ROS, programmers can learn from one another and apply others’ already developed software, significantly advancing robotic technology across sectors.

The ability for robots to self-learn through artificial intelligence is advancing fast. Also known as machine learning, we are seeing incredible advancements. Swarm technology applies this type of machine learning with the effectivity of a group effort, much like how insects, such as bees, communicate with one another to achieve more as a group. Autonomous of humans, these collections or robots work together, communicating with one another and making decisions as a group. There are some applications in agriculture, such as farming, where tractors are increasing in size and thereby compacting the soil. It is believed many, smaller robots, can be just as productively effective but have reduced noise and energy levels. Is upkeep and maintenance of several smaller robots preferable over one big tractor? Time (and economics) will tell!

As mentioned, one of the bigger challenges in robotics is replicating the delicate touch humans have, particularly the sensitivity of our nerves. A company called SynTouch has developed a robotic fingertip that uses electrodes to feel a surface through vibrations and can even feel thermal elements such as hot and cold. This type of sensitivity will move robots into a new realm of opportunities previously only accomplished by humans.

A whole new ‘human free’ work world?

The word “robot” is a Czech word meaning “forced labor” so the concern over whether they will take our jobs seems valid. However, robots still have a lot to learn before they’re fully capable of taking over everything humans can do. Take a moment and think about the activities that humans perform at work each day. Much of it is the same day-to-day, which is where robots come in, but when a new challenge arises it requires critical thinking combined with experience and the ability to research or reflect and pivot. We may not be quite ready to make farming food an entirely autonomous process; it still needs humans.

While agriculture requires some of the hardest working people on the planet, the opportunities robots afford farmers and producers will enable them to be far more productive and efficient. Adopting such technology may even allow for a little leisure time – nearly a foreign concept to farmers and producers! So, while robots are advancing frighteningly quickly, embracing this change, rather than shunning it, is the best advice.

Covid-19 has been an evil moment for our planet, bringing pain and economic strife, but for robots at least it has accelerated the interest in and the willingness of.