By Amar Hanspal, CEO & co-founder, Bright Machines
All around us, our products are getting smarter. They respond to our touch and voice and adapt to our needs. They are responsive, flexible, and intelligent. Unfortunately, the same cannot be said about the factories that make them. Our factories are mainly analog; most are still populated by hundreds of people repeating the same task. Neither the processes nor equipment is designed to respond to change, adapt quickly to customer needs, or address quality issues at scale and with ease.
This outdated approach has real consequences: machines sit idle (utilization can be as low as 40%) while waiting for orders, material, or human intervention. Mistakes and misfires produce large amounts of scrap, and labor turnover can be as high as 30%.
The last big step-function in manufacturing capacity and capability was achieved via globalization more than two decades ago. Now, access to low-cost labor is diminishing, while consumer demand for authenticity, speed, and localization increases. Our manufacturing methods are reaching their limits.
The Next Big Leap
It’s clear that automation is the next giant leap for manufacturing, but its deployment has been stymied for the somewhat ironic reason that automation itself isn’t automated. Instead, machines have to be individually configured and independently managed by human workers. There is no connected flow of information between machines and no method of continuous improvement. This first generation of automation costs too much, takes too long to install, and depends on a small group of experts, making it difficult to scale or replicate.
In the last decade, software has evolved to address these problems. We can automate automation by connecting individual machines to a software layer that dynamically configures and continuously monitors and manages machines to improve operations, creating autonomous production lines and programmable factories. We call this approach software-defined manufacturing. It provides three clear benefits.
How Software Will Define Manufacturing
First, it makes it easier to configure, replicate and scale automation, dramatically changing the economics, speed, and flexibility of automation.
Second, by digitizing factory operations, software-defined manufacturing makes these operations more transparent and accessible, enabling the entire organization to better interact with manufacturing processes. The payoff is agility and continuous improvement.
Third, it adds intelligence. By leveraging computer vision, machine learning, and adaptive algorithms, it gives machines eyes and brains.
With software-defined manufacturing, intelligent applications can act on production data to configure and continuously improve production lines, delivering higher quality and throughput. Production lines become autonomous, and the factories that house them programmable.
Opportunities at the Back-End
The logical place to start automating automation is at the back-end of the production line, where assembly, inspection, and testing are still primarily done manually. That requires a combination of technologies to create an integrated system of hardware (such as robotic cells, automatic conveyors, and material feeding systems) and software that leverages computer vision, machine learning, and 3D simulation. The cells provide the arms and legs, and the software provides the eyes and brains. Dexterity and variability are no longer blockers.
Take, for example, a manufacturer that needs to automate the assembly of a networking device — specifically, to figure out where and how to insert screws and then to insert a high-value memory component called DIMM.
To do this, a system that truly automates automation would first take pictures of a prototype board and ingest 3D geometry describing the board’s design. Using AI, it understands the features of the product, rather than mere geometric positions. Collaborating with a human, the software creates a set of instructions—a manufacturing recipe—to assemble the components. The recipe is then simulated in a virtual environment with a physically and kinematically accurate robot model. This ensures that the proposed actions are safe, accurate, and successful.
Then, the manufacturer presses play and can watch the robot seek out the screw holes using optical sensors, then place and screw the heat sinks into place. The robot uses those same sensors to apply a precision level of force to insert the DIMM module.
That’s just the start. The abstraction of instructions into recipes means that once one software-defined manufacturing system learns how to assemble a DIMM module, all those systems, or microfactories, can instantly do the same anywhere in the world.
Enabling the Future-State of Manufacturing: Nimble, Local, Innovative
Software-defined manufacturing provides a holistic view and deep insights into the entire production line. It makes it easier to pinpoint quality issues and find their root causes, no matter where in the production line they may be. The software suggests corrective action, which an operator can implement. As a result, not only do operational metrics improve, but the production line becomes configurable and self-correcting. This is fundamental to reimagining the factory floor and essential to automating automation.
Software-defined manufacturing is already helping transform the industry, exceeding requirements for speed, scale, and flexibility and yielding a real return on investment. It is also enabling distributed, location-agnostic manufacturing. Customers can make better products in nimbler, smaller, and more sustainable factories. One can sense a future where products are made locally and on-demand.
Perhaps most exciting among the benefits of software-defined manufacturing is its potential to speed innovation dramatically. As factories become digital and distributed, they become transparent and accessible. Manufacturing is transformed into a capability that anyone can access, rather than a mysterious, dark art. Ultimately, it’s not about AI or software or increasing run rates. It’s about enabling anyone to build anything anywhere, on-demand. It’s about the democratization of innovation—and the manufacturing infrastructure to make that innovation real.