In a recently renovated office building in Watertown, Massachusetts, 100 robots are slowly learning how to pick up random objects, including everything from hand lotion bottles to packages of Welch’s Fruit Snacks.

Each human-like machine is equipped with four cameras — one at the head, at the chest and on both claws — and fixed to a stationary box. A few employees carrying smart devices inspect the machines’ work to see if they complete their tasks correctly.

At this point, most of them do not.

The robots, named Sonny after the “I, Robot” movie character, have only been operational for a few days, and it’s going to take several months of research and development before Tutor Intelligence has collected enough data for industrial applications, CEO Josh Gruenstein said at a facility tour April 22. Some day Sonny could be deployed at a Fortune 500 company’s operations.

Tutor has just moved to its new headquarters at Watertown, Massachusetts, located at the former home of Boston Scientific, along the Charles River. The historic site was previously a cotton mill in the 1800s.

“This is really an instrument of discovery for us to determine what are the best robot learning methods, and to start to build that capability set where we might be comfortable going to a customer,” he said about the robot fleet.

DF1 and the unique way Tutor teaches its robots Data Factory 1, or DF1, is what Tutor is calling “the largest data factory in the United States.” The fleet of robots are learning autonomously how to move and manipulate objects with their claws, powered by the startup’s first vision-language-action model, Ti0, and through large-scale human supervision. In addition to the onsite team, remote workers in the U.S., Mexico and the Philippines monitor and correct the robot behaviors.

Tutor, founded in 2021, was started by Gruenstein and Alon Kosowsky-Sachs, chief technology officer, out of the Massachusetts Institute of Technology’s Computer Science and Artificial Intelligence Laboratory. As students, they developed keen interests in the robot brain, and later set out to discover a “scalable recipe” for teaching robots how to do general tasks through their company.

The startup has raised a total of $42 million to scale its AI-powered robot workers, including a series A funding round of $34 million that closed in December.

Instead of using the latest wearable or sensor technologies, Tutor chose its recipe — a mix of cameras, vision-learning software and human supervisors — because it was more cost-effective.

“Actuators are expensive, sensors are expensive, even just like the physical structure of the robots … are not that big and could be 10 times cheaper,” Kosowsky-Sachs said. Cameras are cheap, in part, because of the iPhone and its massive scale of production, he said. He and Gruenstein hope that economies of scale will eventually take hold for other parts used in their robots.

“There’s this fundamental problem with robotics, which is, you can build the smartest robot and the world’s most capable robot, but if its [return on investment] isn’t extremely competitive, your business is uninvestable and will probably go out of business,” Kosowsky-Sachs said.

Support from AWS and its cloud services Last year, Tutor participated in the inaugural cohort of the Physical AI Fellowship, an eight-week startup accelerator supported by Amazon Web Services, Nvidia and MassRobotics, a nonprofit robotics hub located in Boston’s Seaport district.

As a member, Tutor received dedicated science and engineering support from AWS and Nvidia, as well as $200,000 in AWS credits to help “get their solution off the ground and then, of course, go to market,” said Alla Simoneau, global operations leader at AWS.

Tutor’s approach to advanced robot learning differentiated it from other startups in the robotics and physical AI space, Simoneau said.

“What they were trying to do was build a brain for robots that could adapt to different environments,” she said. “And that generality was something that was quite interesting.”

The main goal of the fellowship, which has expanded globally this year, is to advance leading technologies. However, some of the participants have matured and scaled to the point where they can sell their services to customers through the AWS Partner Network, Simoneau said. Deployment in Amazon warehouses is not the end goal, she said, but it could be a byproduct.

“It’s critical to have a cloud partner that can support massive amounts of compute,” Gruenstein said. All of Tutor’s robots are connected to AWS.

Gruenstein said he thinks of Tutor as two halves: One is focused on developing fundamental robotics research and seeking how to build robot models and software, while the other is focused on robot deployments in factory and warehouse settings.

“I don’t think there are any other companies that are building the robot brain and then actually deploying it into the real world and doing productive work in that manufacturing context or that logistics context,” he said.

Tutor deploys its Cassie robot at US manufacturing and logistics facilities While the Sonny robot is still in development, Tutor has deployed its Cassie robot for autonomous case-picking and palletizing with U.S. customers.

The 2,000-pound industrial robot can manipulate boxed goods and materials, among other tasks, for a wide range of customers, Gruenstein said. He declined to offer specifics about the cost and number of Cassie deployments. Early adopters in the food manufacturing and logistics sectors attended the tour and said they were already seeing the benefits from their pilot experiments.

Paul Baker, CFO at Productiv, a third-party logistics provider, said Tutor’s robot is operating at or better than the level of a human at its Dallas warehouse. The company is testing about 15 other robots that perform at sub-human levels, doing high-variability, high-dexterity tasks, with the goal of improving speed and efficiency.

“We’re working with groups, including Tutor, to try to [improve dexterity] as fast as possible,” Baker said. “That’s kind of that next frontier, and that’s why they have the data center working.”

Jeff Pulley, a facility manager at BetterBody Foods — a maker of peanut butter powder, avocado oil, chia seeds and other health-focused goods — said the company in July added two of Tutor’s palletizing robots to its Lindon, Utah, facility and one to a factory in Greenfield, Massachusetts.

Since implementing the robots, Pulley said BetterBody Foods has saved 36% compared to the cost of a worker that would normally be loading and unloading boxes from the pallets. He said the company hopes to add three more Cassie robots to its facilities over the next year and a half, and eventually add Sonny robots when they are ready.

“Robotics and automation, I believe, is … where everything’s going. If you want to delay it, you’re slowing down your own progression,” Pulley said. “So [you] might as well embrace it and learn from it.”

By: DocMemory
Copyright © 2023 CST, Inc. All Rights Reserved