O.K. Carter: Creating machines that not only do work but also learn

Cloud 9 Perception

130 S. Collins St., Arlington 76010



- FWBP Digital Partners -

Repetitive steps in the supply chain process can use machine learning to reduce cost and speed up boring but necessary processes.

Put a couple of computer engineering wizards together puttering with gadgetry in a garage and sooner or later, if they’re both smart and lucky, they’ll come up with a “What if …?” question they’re compelled to answer.

For computer engineering senior lecturer Christopher McMurrough of the University of Texas at Arlington and his former student James Staud (also a computer engineer), the question was, with apologies for extreme simplification: “Can we teach machines to see and manage complex work flow by themselves, learn as they go, fixing problems and increasing efficiency?”

Their company, Cloud 9 Perception, evolved from that question, the Arlington company today emerging as a leader in a blended discipline involving stand-alone collaborative robotics, machine visioning and machine learning. You may know the last part by another name: artificial intelligence.

- Advertisement -

But back up. It began a few years ago with a robotic arm.

“One of my associates had an extra one and offered to give it to me to experiment with,” Staud recalls. “I said sure, why not?”

He stashed it in a barn, where it sat. For a while.

But when Staud mentioned that he had the robotic arm, McMurrough – then working with a blend of 2D and 3D visioning instrumentation – suggested they bring it to the garage.

- Advertisement -

They began a long process of blending computer programing, the ability of the optics to see and interpret, and the robotic arm’s functioning.

Think of it as an assembly line that can do the usual automation required but which also recognizes variances with 3D visioning, fixes them based on consultation with its computer interface and also learns along the way with little or no need for human interaction.

Eventually their fascination morphed into the creation of Cloud 9 Perception with the help of fellow entrepreneur and funding expert Don Addington.

“Christopher was inspired to develop a truly industrial solution to 3D machine vision,” says Addington. “One that was to lower cost and bring the possibilities of vision application to every level of the supply chain. James saw the potential of using modern hardware and software methods to reduce costs and to introduce new features to industrial vision applications,” Addington said.

He is now the CEO of Cloud 9 Perception, working with McMurrough and Staud at their downtown Arlington headquarters.

Much of the company’s focus in recent months has been on the grocery industry and product flow. Most of the products arrive on old-fashioned wooden pallets and conveyor systems, but the old-fashioned part ends there.

Once there, 2D sensing takes over tasks like barcode scanning and label reading, while 3D image data recognizes dimensions, defects and variances in the norm. Pallet-packed products skim along to their proper destinations and functions with minimal human interaction. Or interference.

In the Cloud 9 work lab, Staud points to a pallet loaded with boxes filled with cans of tomato sauce.

But there’s a problem. Somehow along the way, a case of the sauce has been partially crushed, creating not only a cleanup problem but an unbalanced pallet that might tip and become a safety hazard.

The visioning gear picks up the issue right away, shuffling the damaged pallet aside for a quick cleanup, restack and – inventory adjusted, problem reported – the unbalanced pallet goes back quickly into a system in which thousands of pallets with thousands of products are moving to thousands of locations.

All autonomous without need of remote computer hosts. And in a hurry.

“The automation industry is in the midst of a second robotics revolution,” McMurrough points out. “What we’re doing is facilitating that transition.”

The future, he adds, “is machine learning. We’re working to build an industrial-grade solution that can learn and improve as it works on a task. This will make lead times to integrate vision much smaller and make the possibilities much more dynamic.”

McMurrough, Staud and Addington split marketing responsibilities, promoting their innovative approaches not only on line but through ongoing appearances at trade shows and the inevitable word of mouth.

They’re also slowly expanding their artificial intelligence assembly line robotics to other industries, so much so that they’ll eventually run out of room at their current location.

“I give us another year here, and then we’ll need considerably more space,” Addington says. “We don’t know where that will be, but we know it will be in Arlington and with a continued close connection to UTA.”

O.K. Carter is a former editor and publisher of the Arlington Citizen-Journal and was also Arlington publisher and columnist for the Star-Telegram and founding editor of Arlington Today Magazine. He’s the author of the definitive book on Arlington’s colorful history, Caddos, Cotton and Cowboys: Essays on Arlington.