Augmented reality comes to factories

In front of me is an optical color sorter, one of dozens of giant machines on the floor of a plastic recycling plant in Laverne, Montreal. At the time of my visit, I didn’t know what it was for, let alone how to operate or clean it. But with a HoloLens 2 augmented reality headset placed in front of my eyes, the images appear almost above the device and tell me what to do to save it.

The two virtual arrows indicate, for example, the screws that need to be turned to open the machine housing. In the next step, more arrows appear to show me where to shoot the compressed air. I am learning how to clean the optical sorter, but I might as well learn how to operate it or perform preventive maintenance.

“Our goal is for an employee who has never completed a task to be able to put on a headset and complete it quickly without error,” explains Nicholas Bearzatto, co-founder and general manager of DeepSight, the Montreal-based company that developed the headset. platform, the software used by an augmented reality headset to display instructions.

Another learning tool

The DeepSight software platform, launched in February, comes at just the right time, when the need for training in the manufacturing sector is clear. According to a recent survey by Quebec Manufacturers and Exporters, 98.5% of companies in this sector are experiencing labor shortages. And the training of new personnel is often carried out in cooperation with experienced workers who do not have a minute of free time in their shift and therefore they are doubly in demand.

To solve this problem, DeepSight has developed a set of software to develop instructions without programming, much like creating a PowerPoint presentation and then displaying them on a Microsoft HoloLens 2 headset. At Lavergne, the person responsible for creating instructions puts them together throughout the day, and the next day reviews them with staff and adjusts based on feedback.

Augmented reality has two big advantages over the paper instructions that are usually offered in factories: it allows you to perform tasks with your hands free and see exactly where to perform operations.

And it seems to work. Instructions aren’t finished at Laverne, but Quebec-based aerospace company Avior, which deployed DeepSight technology ahead of its official launch, says augmented reality has reduced employee training time by a factor of two to three.

However, recruits are not the only ones who benefit from augmented reality training. “Sometimes someone with experience can be temporarily transferred to another department, and night workers don’t always have a warden when they have questions,” illustrates Patrick LaViolette, who was a trainer at Laverne at the time of the transition. Newsat the start of this winter.

“We are developing a console that will be added to the system to collect data that will help determine, for example, how long tasks take and which ones are the most problematic, which will allow factories to improve their instructions,” notes Nicolas Bearzatto.

There will be other use cases

For Joan Laverne, director of marketing at Lavergne (the company founded and still run by his father Jean-Luc), learning is just the beginning of the journey to augmented reality.

“I would like us to buy glasses for each of our big clients so we can help them remotely,” he notes. Currently, Lavergne often has to send employees to Asia to help customers use certain machines and the company’s recycled plastics. An augmented reality headset that allows you to see what employees on the other side of the world are doing and guide them can help avoid certain trips. “The potential is incredible,” says Joan Laverne. However, at around $5,000 per helmet, the project is on hold until the technology improves and its price drops.

Other anticipated technological advances could also be useful in the manufacturing sector, such as image recognition. In front of the optical color sorter, the helmet indicates, for example, where to screw it on and asks you to press a virtual button in the air with your hands to signal the end of the stage. The software will once be able to verify that the operation has actually taken place, making it easier to use and minimizing the risk of error.

“It is possible to train a similar AI model in a controlled environment to achieve such a result, but currently the technology does not allow it to be available in any company,” Nicholas Bearzatto clarifies, adding that DeepSight intends to offer such functions in the future. .

Leave a Comment