In its latest master plan, Tesla Inc. revealed a vision of the future in which robots handle work, while humans have “more time to do what they love.” After creating fully autonomous, electric vehicles and designing a general-purpose, artificially intelligent, bipedal humanoid robot, “we are combining our manufacturing capabilities with our autonomous prowess to deliver new products and services that will accelerate global prosperity and human thriving driven by economic growth shared by all,” the company writes.
Tesla officials say its humanoid robot, Optimus, “is changing not only the perception of labor itself, but its availability and capability.” But can robots like Optimus handle the complexities of doors and windows? With the latest advancements in door and window machinery and artificial intelligence (AI), they may not have to. The latest machinery lines don’t have humanlike hands, and they don’t walk upright, but they don’t need humans to operate.
“Robotics have really given us that flexibility and dexterity, very similar to the human hand,” says Tim McGlinchy, executive vice president of engineering and R&D for GED Integrated Solutions. Additional axes of movement and proper tooling give them the ability to carve out changes as small as 0.010 inch or even 0.021 inch, he says.
Now companies are exploring how they can use AI-powered data gathering and analysis to enable machines to make those finite adjustments instantaneously, with no human involvement. In this way, robots and AI are quickly becoming symbiotic, McGlinchy says.
“It’s so highly iterative and faster beyond a human that it’s not even [like] being human,” he says.
When it comes to sealants application, for instance, machines can now inspect and analyze in real time—amid application—making the best decisions regarding how and when to adjust. The system can be adaptive, cutting back on flow or pressure to apply just the right amount.
Robot Who?
In this way, the door and window industry isn’t crawling with humanoid robots, but it does employ its fair share of gantry robots and other robotic features. Robots already facilitate the movement of materials in and around the manufacturing floor. For instance, Oz Machine employs robots in its corner welding and corner cleaning machines for loading and unloading profiles on its machining center, says Adil Sasmaz, the company’s managing director.
“These are the robot applications we have,” Sasmaz says. “Not like the Tesla robot walking around, installing things.”
As products get larger and heavier, gantry robots have become increasingly important, says Morgan Donohue, president of Erdman Automation Corp.
“In my estimation, everything in our industry seems to be continuing to get larger,” Donohue says. “It seems like all of the customers are continuing to push the envelope on size.”
For instance, “In aluminum profiles, they are heavy and can be roughly 24 feet long, and you need to grab them and put them on the machine for cutting or machining,” says Sasmaz. “We use robots in such applications.”
For situations such as these, “I think there’s going to be a continuation of adoption for cobots,” Donohue adds.
Designed to work in tandem with human operators, machinery makers estimate that cobots can handle as much as 50-60% of labor, but they can also be used to eliminate repetitive injuries and to increase safety, augmenting humans when size and weight are factors. For instance, a custom-engineered robotic cell by Augsburg, Germany-based Kuka handles door inspection, flipping and stacking. The robot precisely maneuvers and flips doors, handling a motion that, according to the company, is difficult to perform for humans without risking damage or strain.
Meanwhile, AI can ease the need for constant mental focus and decision making.
Not Very Chatty
The type of AI making its way into the door and window industry isn’t the web-based chatbots such as OpenAI’s ChatGPT we’ve all grown to know. Instead, it’s more behind the scenes. But there are use cases in which machinery companies plan to deploy chat-like features, such as for customer service and troubleshooting.
“We’re going to take AI one step further from a service standpoint,” says Joe Shaheen, vice president of sales for GED Integrated Solutions. “We have phone support here, with four guys who have 25 to 30 years of experience. They are great at what they do. We also know they’re not here for the next 30 years.”
GED is loading its manuals into an AI engine, Shaheen says.
“We have to test it, security-wise, and validity-wise, because we have to ensure that it’s right,” he says. “But we’re going to start stepping it up, one by one, as we train somebody new. We’ll let AI communicate with our person. They ask the question and AI gives them an answer. Eventually, our customers will be able to chat with it directly.”
Other than those uses, for the most part AI will be used in less obvious ways, mostly in data collection and analysis. But with the right machine-learning mechanisms, that can go a step farther. AI can gather and process information in an instant, analyzing such things as temperature readings, vibration, sound and pressure readings, to detect issues and early indicators of breakdown. That sort of instantaneous number crunching could lead to big savings for door and window manufacturers, machinery companies suggest—especially when it comes to keeping just-in-time production on course. By the same token, artificially intelligent machines can help manufacturers to maintain their equipment, not on a prescribed basis, but to the needs of exact usage scenarios.
“As we continue to collect data, we can start to spot trends that help our customers minimize downtime by proactively and even predictively maintaining their machines,” Donohue says.
Another area in which AI has the ability to take robotics next level includes programming, which can now be done in natural language.
German robotics manufacturer Kuka now uses a new, text-based AI assistant, dubbed iiQWorks.Copilot, to facilitate easier robotics simulation and programming. The initiative was launched as a development project with Microsoft. The goal includes facilitating natural language commands, such as, “Pick up boxes from the conveyor belt … and palletize them in three rows, two columns and three layers.” With that simple, straightforward command, Copilot is able to generate a fully functional program for robots in just a few minutes. That program is then validated through simulation before it’s transferred directly to the robot controller.
“That’s the kind of stuff that a 20-year programmer does over the course of a week, and it can do it in minutes,” Donohue says. “Then that long-time programmer can double check and ensure that, yes, it does work, then apply the program.”
Prepare for Takeover
Ultimately, the door and window industries remain a far cry from the humanoid work environments envisioned by Tesla Inc., but lights out manufacturing is almost here, Sasmaz and others suggest. So far as a takeover, Tesla officials say we need not be concerned about human decline. “Growth in one area does not require decline in another,” they assure.
For example, they point to the industrial revolution, which “widened our economic landscape,” they suggest. Similarly, semiconductors, computers and the internet “have expanded—not diminished—social and economic opportunities,” they point out.
If Tesla has its way, though, one day doors and windows could be manufactured 24 hours a day, seven days per week, by robots. Meanwhile, company employees could be out taking a jog or fishing.
