Future of farming: AI-enabled harvest robot flexes new dexterity skills
Now that Root AI’s robot has mastered the art of picking oblong vegetables as well as oblate shaped berries, could this technology help enhance global food security?
In recent months, the coronavirus pandemic has highlighted frangibility in the global supply networks; particularly those involved in food security. Hallmarks of digital transformation, automation, and artificial intelligence, are being tapped to create a decentralized 21st century food chain.
On Thursday, the agricultural robotics and artificial intelligence company Root AI announced new capabilities to its AI-enhanced robotic harvester as well as investments totaling more than $7 million. Now that the AI-enhanced robotic harvester has demonstrated enhanced dexterity to tackle crops of various shapes and sizes, the technology could help shore up these vulnerabilities.
In the past, Root AI has provided glimpses of its robo-harvester, known as Virgo, picking ripe tomatoes off the vine. In the latest video titled “Going Cross-Crop,” Virgo is shown picking cucumbers and strawberries in the field. In the accompanying press release, the company asserted that Virgo was the “world’s first robot ever to replicate a person’s ability to harvest multiple crops.”
In an indoor agriculture environment, Virgo can be situated on a track in between rows of various crops. As it navigates a greenhouse, the robot leverages a host of sensors as well as artificial intelligence to analyze crop positions and ripeness and then uses a specialized gripper to pick produce once it’s ready.
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.
Whether it’s an apple in a tree or a strawberry in a bush, these are both similar challenges from an identification, planning, and picking perspective, explained Lessing. The robotic gripper and software can be swapped for different crops, however, the underlying principles surrounding planning and picking will enable application across crops.
“It’s the same gripping concept, but for your human hand, when you grab a cucumber or, which is a cylinder, or you grab a tomato, which is an orb, you pose your fingers in different arrangements to grab them.
