Virgo by Root AI is a robotic arm that uses machine learning to identify when a crop is ripe and ready to harvest, and then with remarkable agility, maneuvers amidst other non-ripe crops, and carefully and gently harvests the crop.
Intelligent robots that will help growers build the farms of the future?
While there’s considerable potential for automation and robots in agriculture, there’s also a lot of variables that influence not only possible success, but whether such systems are viable in the first place.
After all, technology is not magic, but generally requires specific circumstances and operating conditions for it to work as intended. Which is why automation in agriculture can be a challenge, as nature can pose all sorts of problems and unintended issues. Farmers have to face a wide range of environmental events and incidents, it’s ludicrous to think a robot could be programmed to anticipate and manage all that the natural world can offer.
Yet what if nature itself could be controlled or at least programmed? That’s part of the appeal of greenhouses and the use of technology to control and maintain specific environmental and growing conditions.
For example, in southwestern Ontario, greenhouses have become a mainstay of agricultural production. While tomatoes are a dominant crop, these advanced buildings, using sophisticated technology, are able to grow a wide range of crops, throughout most of the year, with yields that continue to improve.
However these greenhouses still face the same labour issues and shortage that plague the agricultural sector as a whole. In particular harvesting crops still requires significant manual labour that in the midst of a pandemic has been in short supply.
Is this an opportunity for robots and automation? While robots may be questionable in a natural environment, a greenhouse provides enough environmental control and technological infrastructure that an automated approach not only makes sense, but may actually be possible.
This is the focus of the company Root AI, which is building a robot named Virgo. Virgo is a robotic arm that uses machine learning to identify when a crop is ripe and ready to harvest, and then with remarkable agility, maneuvers amidst other non-ripe crops, and carefully and gently harvests the crop. Here’s an example with tomatoes:
And another with strawberries:
Building a robotic platform for greenhouses makes sense, as the controlled environment provides clear parameters for that robot to work. In this case the arm is able to move around precisely because the ground is even and the space controlled.
This allows the company to spend their time focusing on the gentle touch of the harvesting device, rather than worry about general issues of movement and mobility. Instead that motion can focus on the arm’s ability to move among vines and plants to find and identify fruit or vegetables to harvest.
Here’s a paragraph from the company’s website:
We believe tomorrow’s farms won’t look like the endless outdoor crop rows we remember and that the security and quality of our food will no longer be limited by an increasingly unpredictable climate or access to land. Instead, hyper efficient indoor farms will grow delicious and chemical free produce. At Root, we are on a mission to create the intelligent robots that will help growers build the farms of the future.
This company is part of a genre of technology companies focusing on a vision of agriculture that is arguably more science fiction than reality. They emphasize the benefits of greenhouses, in particular when it comes to minimizing water usage and resources, but also mitigating climate volatility and climate change by growing in controlled environments.
However this vision ignores the reality that the vast majority of our food is grown on traditional farms, using soil, and that there’s no way greenhouses could scale to replace this, nor could most farmers or agricultural operations afford converting to greenhouses.
Soil is an essential resource, one that we cannot take for granted, and something that we need to better understand, both when it comes to regenerating soil, but also appreciating the nutrients and life it brings thanks to the bacteria and microbial life that exist in our dirt.
Yet soil is dirty, literally and figuratively, and working in soil is complicated, which is why so many tech startups focused on agriculture choose to ignore soil (at their peril).
Though maybe we shouldn’t blame these startups for following the money, as there is a lot of investment capital flowing towards this kind of technology.
By all accounts, the COVID-19 pandemic has accelerated the adoption of robotics and automation by months, if not years. The reasons are fairly clear — robotics don’t call in sick, nor are they disease vectors in the same manner as their human counterparts. As food production and agriculture are looking to be among the biggest winners from the trend, it’s little wonder that Root AI is announcing a new funding round.
The Boston-based startup has already been getting a fair bit of press (including on these very pages) with the promise of produce-picking robotics. This week it announced a seed round of $7.2 million, bringing the company’s total funding up to $9.5 million, courtesy of Rob May of PJC, First Round Capital’s Josh Kopelman, along with Jason Calacanis and Austin McChord of Outsiders Fund.
Don’t get me wrong, there’s tremendous potential for this kind of robot, not just in greenhouses, but converted to work in fields as well. The machine learning technology and robotics being developed here will have a wide range of additional applications.
The Virgo is a self-driving robot with sensors and cameras that serve as its eyes. Because it also has lights on board, it can navigate large commercial greenhouses any hour of the day or night, detecting which tomatoes are ripe enough to harvest. A “system-on-module” runs the Virgo’s AI-software brain. A robotic arm, with a dexterous hand attached, moves gently enough to work alongside people, and can independently pick tomatoes without tearing down vines.
The robot’s “fingers” are made of a food-safe plastic that’s about as flexible as a credit card, and easily cleaned. Josh Lessing, founder and CEO of Root AI, says that easy-to-clean trait is important.
“People don’t think about this -- you have to manage disease on a farm. Just as if I picked with my own hands, there’s a risk of spreading around mold, viruses or insects with a robot. That’s why you want these to be washable. It is part of the work you do to keep the plants safe.”
One of the most unique things about the Virgo, he notes, is that the company can write new AI software and add additional sensors or grippers to handle different crops. “It’s a complete mobile platform enabled to harvest whatever you need,” says Lessing.
This platform approach to technology is relevant given the dominance of platforms in the technology industry as a whole. It makes sense that Root AI is looking to expand the functions and abilities of its robot, in contrast to other companies who are developing application specific machines.
It is this platform based approach that suggests Root AI has an eye beyond the greenhouse, to smart robots operating in fields.
The machine sees its environment in 3D using intelligent motion sensing. This data allows Virgo to determine an optimal route, through vines, leaves, other unripe crops, to pluck its target. As Root AI CEO Josh Lessing explained, Virgo uses more than computer vision to see its environment and plan accordingly.
"We need to go beyond computer vision that finds fruit in three-dimensional space. We do that, but on top, we have a layer of computer perception that then plans how to go about grasping that fruit. How do I navigate through the environment and then land my fingers on that target to effectively pick it? To move with authority, the same way people look at an object that they want to pick, the mind needs to create a plan," Lessing said.
The company is building solutions to enable its fleet of systems to learn on the job, so to speak, and then share these insights with other robotic harvesters in the field.
"We're building artificial intelligence algorithms that understand how to do a task, but as it works, learns how to do it better, and then shares those learnings across a fleet of systems," Lessing said.
When most people think of robots, they think of hardware, the physical robot itself. Yet this is misleading, as the reason we’re on the cusp of a revolution in robotics, is because the software is what’s currently driving innovation. AI and machine learning are enabling new capabilities and functions, and these advances are not static, but ongoing.
The advantage for a company like Root AI stems from their desire to focus on different crops, which allows for the development of smarter and more diverse systems, that are capable of expanding and harvesting more complex crops and growing environments.
We’ll continue to explore and profile companies engaged in the automation of agriculture and the development of farm robots, and keep an eye on Root AI as the development of their robotic platform progresses.
Finally, here’s a video that breaks down the methodology behind Root AI’s robotics: