Content
Skip to Main ContentAccessibility Help
Spark

The Shazam of machines

A new tool can diagnose problem machinery - just by listening!
CBC Radio ·
Figuring out why a machine is malfunctioning is often a lot harder than fixing it. What if the machine could tell you what was wrong? (Michelle Parise)

Machines are all around us...how would we live without them? But they're always breaking down and in need of repair!

Fixing problem machinery is one thing, but figuring out what part of it needs to be fixed? That's quite another.

Saar Yoskovitz (Courtesy Augury, Inc.)
The sounds machines make can be good indicators of whether something is wrong, or about to go wrong. If you know what the sounds mean.

Now new technology promises to use those sounds to diagnose what's wrong with our machines.

Saar Yaskovitz is the co-founder and CEO of Augury.

They've developed a device and software that he likes to call "The Shazam of Machines."

"We can tell you if they're working properly, if they have a malfunction, and even what type of malfunction they have."

Much like Shazam uses a database of music to detect and report what song you're hearing, Saar's device uses a database of malfunctioning sounds.

To build that database, Saar and his team had to listen and record a LOT of sounds, using sensor technology and machine learning to learn more about problem machines.

The auguscope, what Augury uses to listen to appliances - and diagnose mechanical problems. (Courtesy Augury, Inc.)

Right now they're mostly focused on recording the sounds of industrial machines. "We listen to the noises that machines emit, and then we try to understand what they're saying to us," he says.

It's a pretty cool idea. And hopefully, we could soon use Augury's technology to identify why the darn washing machine is making that weird thunka-thunka sound.

More from this episode