When engineering.com attended Automate 2022 this month, over 500 exhibitors shared technologies ranging from 3D cameras to artificial intelligence (AI) platforms. One interesting company was Veo Robotics, a Massachusetts-based manufacturer intent on improving human-robot collaboration.
At Veo’s booth, attendees could explore how a robotic work cell would be safe for human entry. A robot on the show floor used Veo’s FreeMove engine to scan the space, subsequently capturing, analyzing and acting on the 3D data. FreeMove is transforming how work can be accomplished through the COVID-19 pandemic.
“As manufacturers have been challenged by the pandemic, supply chain constraints and labor shortages over the past few years, they've only increased the number and use of industrial robots within their facilities,” said Patrick Toner, director of product marketing at Veo Robotics. “FreeMove is the only product currently available with the safety design needed to unlock the true power of human-robot collaboration on facility floors.”
Understanding the FreeMove system
Consider FreeMove to be the world’s first intelligent 3D safeguarding system, said Patrick Sobalvarro, president, CEO and co-founder of Veo Robotics.
“It's a computer vision system that does classification,” Sobalvarro said. “We had to build our own hardware for the safety rating. We're mostly a software company but do new releases every few months. We get them maintenance reviewed by TÜV Rheinland (a compliance partner) to continue to build in new features.”
The FreeMove system differs from Power and Force Limited (PFL) robots in that Veo’s robots move quickly rather than slowly. Veo’s robots also do not need to be caged. FreeMove does not require integrating multiple 2D sensing systems. It also frees workers from creating custom spreadsheets with complicated geometries to define the space in which the robots can move.
“You can run right up to the robot,” Sobalvarro said. “We can palletize five times faster than one of those PFL robots. That means we can build bigger… and heavier pallets.”
FreeMove works by combining a high-performance industrial computer with dual motherboards. The computer is checked by a safety processor. The system processes data captured by sensors placed around the work cell as the computer interfaces with the robot and other hazards.
In Veo’s imaging software, the robot appears as a highly detailed representation in white. A group of red dots makes up the form of each person. The robot future cloud is a group of yellow dots. This space is a calculation of all the places where a person could be and where the robot could get to before it is brought to a stop. A green line represents the minimum distance between the robot and person. This green line moves each time the person or people in the work cell move. It is recomputed with every cycle of the system.
The key to FreeMove’s success is measuring the minimum distance between people in the work cell and the robot future cloud. If the robot is closer to any person than the protective separation distance (PSD), where there might be a safety concern, the robot must stop. When the person starts to move close to the robot, the system signals a speed limit.
This causes the robot to slow down. The proximity also decreases the size of the cloud, allowing the robot to stop faster. FreeMove signals the robot to stop quickly if the person gets extremely close. When a person moves away from the robot, the system can clear the PSD violation. In many cases, this allows the robot to restart automatically.
One way a person and a robot can collaborate is by the person overseeing the robot’s placement of products on a pallet. After the robot has placed all the products onto the pallet, the person can come in and move the pallet. The robot can continue moving products to another pallet.
FreeMove is helping manufacturing to become more flexible, particularly in times of crisis. Enabling people and industrial robots to interact safely, easily and quickly helps companies be more responsive to changing markets.
Demand for Industrial Robots Is Up
The pandemic is a prominent factor in the rise in demand for industrial robots. According to the Association for Advancing Automation (A3), orders for such robots went up 28 percent in 2021 from 2020, the biggest jump since 2017. A3 statistics reveal that factories ordered 39,708 robots in 2021. Furthermore, in 2020, more non-auto companies started buying industrial robots.
Veo currently works with all of the larger automotive Road Experience Management companies (REMs), which are Tier 1 component makers.
“We also work with durable goods manufacturers who produce items like aircraft engines, refrigerators, stoves and dishwashers,” Sobalvarro said. “We’re getting pulled into logistics, like with big e-commerce retailers. We work with all the big robot manufacturers: FANUC, KUKA, ABB and Yaskawa. That enables us to meet the needs of our customers and work with their few use cases.”
Sobalvarro explained that FreeMove exists because PFL robots don’t meet manufacturers’ needs. Before the pandemic, Sobalvarro paid a visit to the BMW factory in Spartanburg, South Carolina.
“I heard, ‘We want to put an instrument panel in a car, and it weighs 200 kilograms,’” Sobalvarro said. “At the time, a human worker had to clamp the instrument panel and push it into the vehicle body once every 75 seconds. This was hard work. Then, other people had to jump in and connect the panel with the wiring harness.”
PFL robots could not accomplish the job because they move very slowly. They have to be caged because their end effectors are hazardous. PFL robots typically cannot pick up more than 10 kg. The only way they know how to stop in the presence of a person is by bumping into the person. This is dangerous if the robot is engaged in a hazardous task, such as sanding.
Veo Robotics’ Beginnings
Veo Robotics was founded in 2016 by Sobalvarro; Clara Vu, chief technical officer; and Scott Denenberg, vice president of engineering. Sobalvarro got his start in industrial robots at the MIT AI lab in the late 1980s.
Between 2008 and 2010, Sobalvarro sold a company that specialized in tracking people in large retail stores to Tyco International, which at the time owned ADT and American Dynamics.
“I thought, ‘What if we could track people around the robot and use the robot controls to put it into a safe state?’” Sobalvarro said. “When people are moving around it, they get too close, and it creates a hazardous situation.”
At the time, the technology would not allow such innovation. By late 2015, Sobalvarro saw parts come down in price. That year, he was an entrepreneur in residence at Siemen’s Venture Capital. Sobalvarro considered creating a company that realized his vision to make large robots collaborative.
He then reached out to Vu and Denenberg to form Veo Robotics.
“Our goal was not to be like other startups that wanted to move fast and break things,” Sobalvarro said. “We wanted to be good citizens and participate in the standards committees. It took five years and $40 million just to get to our first product.”
Sobalvarro’s advice for engineers who want to innovate is to consider working for a systems integrator or a major manufacturer. The key to building a manufacturing work cell is to be able to run millions of parts.
“You want to create a robot that’s built for 100,000 hours of service,” Sobalvarro said. “That’s 10 years, 24-7, and maybe 20,000 or 25,000-hour service intervals every couple of years. Compare that to a car you’ve owned for five or six years that hasn’t gotten 20,000 hours on it yet.”
According to Sobalvarro, discipline to respect functional safety helps Veo stand out from other robotics manufacturers. This concept has proved essential to building health monitoring, fault tolerance, document control and failure mode analysis into FreeMove.
“You got to find the right kind of engineers who see that as a positive challenge,” Sobalvarro said. “They also have to be brilliant enough to be able to develop the kind of stuff that we develop. We like living up to that challenge.”