Soft Robotics Inc. has learned a lot in the years it has spent primarily focused on automation in food processing.
The Bedford, Mass.-based company makes grippers and high-speed 3D vision software using artificial intelligence.
Harley Green, director of business development at Soft Robotics, said the company initially took more of a “shotgun approach” and tried to go after the automotive and pharmaceutical industries. For many businesses, however, the price-to-value proposition didn’t exactly add up, he said.
Soft Robotics expands from food to logistics
Soft Robotics realized that its best bet was to start with the food-processing market. The company sells standard end effectors through its mGrip product line and integrated systems that combine 3D vision, compliant grippers, and AI through its mGripAI line.
“The product market we found to be really successful is food,” he said. “When we talk to customers in that space, they don’t ever question pricing because they know it’s going to cost more for a food-grade solution,” he said. “We took a step back and said, ‘Let’s not get into the noise. Let’s just double down and go after food.’”
The company recently announced that it is now marketing to the e-commerce and logistics markets. With a few years of experience in the food space and advancements it has made in its technology, Soft Robotics believes it can make the pivot.
“Focusing on one specific industry that was asking for our technology rather than trying to sell to companies that weren’t asking for our technologies was a great move on our part,” Green said. “It also helped us make that transition to e-commerce and logistics types of applications and help us develop our technology further and go after other opportunities.”
Tool changing increases in speed, importance
As end-of-arm tooling [EOAT] technologies improve in precision and more industries look to automation, gripper makers have kept busy. Software advances are big driver, as user interfaces have become easier to use.
On the hardware side, there are new engineering feats such as the introduction of seven-axis collaborative robots and new grippers outfitted with force sensors.
Customers are always asking for more precision, speed, and flexibility, Green said. Getting customers the systems they desire takes a combination of many different types of technologies.
“You really need to marry a number of different technologies into an end-of-arm tool to be able to handle all of these different types of products that you're going to run into,” he said.
Green mentioned advancements in automatic tool changers, noting that one of the biggest customer requests is the ability to eliminate the need for tool changes at all and to do it all with a single tool.
In the meantime, vendors such as Stäubli provide offerings such as the Safety+ System, which includes hardware and software for reliable and safe tool changing.
Accurate to the 99th percentile for e-commerce
Vince Martinelli, head of product and marketing at RightHand Robotics Inc., said pick-and-place automation really exploded about five to 10 years ago, in parallel with industry labor challenges felt across the board and the rise of online and mobile ordering.
The Somerville, Mass.-based company’s RightPick System uses a gripper, a vision system, and control software. It’s used for picking applications in the materials handling space. RightHand sells its system as a package, rather than each component piecemeal, since Martinelli noted that users want an integrated system.
When RightHand Robotics launched RightPick 2 in 2019, it could do somewhere around 750 to 800 picks per hour at trade shows. Its latest version, the RightPick 3, can do 1,200, he said.
Accuracy and precision are of utmost importance on production lines, said Martinelli.
“If you get it right 90% of the time, seven or eight years ago, that’s great,” he added. “You’d see demos of the robots moving things, and people would say, ‘Wow, that’s incredible.’ If you gave it to people who run warehouses, they’d say, ‘What do I do with the other 10% of the time when it fails?’”
99% isn’t even good enough, Martinelli added.
“You start [calculating]: If a robot does 500 to 600 tasks an hour in a particular job, if you fail one out of 100, that’s five or six failures an hour,” he said. “If you have 10 or 20 robots, someone’s spending a lot of time resolving issues. So, when we talk about accuracy and precision, it's about hammering out edge cases and issues and problems to get to 99.9% every time the robot is asked to do a task.”
Testing is of paramount importance to making faster and precise systems. Martinelli explained that over the past two to three years, RightHand Robotics has focused on those capabilities.
“I’m not spilling any secrets here,” he said. “If you want a highly reliable system, you have to invest in that.”
RightHand spends a lot of time feeding data back to its grippers to help them grasp items more efficiently, said Martinelli. Machine vision, sensor, and camera refinements and dropping prices have also played an important role.
Martinelli cited Intel’s RealSense depth cameras.Taking advantage of computer vision hardware and software, Intel said its cameras were “designed to give your products the ability to understand the world in 3D.”
“We’ll see where the camera technology goes because I think there is a big enough market across robotic applications,” Martinelli said.
Electric versus pneumatic grasping
At the Automate 2022 trade show, OnRobot A/S showed off its new OnRobot Palletizer. Kristian Hulgard, the Odense, Denmark-based company’s general manager for the Americas, said the EOAT maker is moving toward offering application-focused products rather than only selling grippers.
Hulgard said the palletizer uses both OnRobot's hardware and software.
“The hardware is a telescopic lift that can lift the robot up to palletize tall stacks of boxes,” he said. “It also comes with two different tools to handle different types and different sizes of boxes.”
OnRobot is trying to expand its horizons and sell more than just grippers, acknowledged Hulgard. “We've come a long way introducing the cobot as 'plug and play,' but there are still a lot other items or pieces of application that might not be easy to use,” he said.
Hulgard said he is seeing increasing adoption of electric grippers over pneumatic ones because of their greater accuracy and control.
“More and more companies have a need to minimize their costs on the use of air,” asserted Hulgard. “Being able to control your grippers with software is extremely key part of modern applications.”
“In terms of accuracy, in terms of operations, the fact that you can control your gripper and monitor it to see how it performs as it should is definitely something that has a positive effect on the accuracy, operations, and quality,” he said.
Still early days
Martinelli compared the automation space and the potential of EOAT to the airplane industry, noting that it is still “early days.”
“Sometimes, we jokingly say we’re past the Wright Brothers stage,” he said, “but we’re not at the point where you can book an international flight and get a meal and watch a movie and be on Wi-Fi.”
Editor's note: This article has been updated to specify that Hulgard was referring to pneumatic grippers.