Neuron Soundware: The Machine Whisperer

Rosina Geiger

Neuron Soundware uses artificial intelligence (AI) to listen to a machine’s sounds, analyzing and interpreting them inside a “digital brain” to identify emerging breakdowns so humans can prevent them from happening.

Squeal, hiss, grind, wheeze, rattle, cough, pop, whirr, bang

We can often hear a problem even before we know exactly what’s wrong. For example, most parents don’t need to see lab results to know that their baby has a cold—they can tell by the way it cries. Similarly, you can probably anticipate that your car’s brakes are about to fail by the gut-wrenching squeaking sound they make.

Hearing with your digital brain

Crying babies and squeaking brakes fall within the “sweet spot” of human hearing—the band between 2,000 and 4,000 Hz. Beyond this range, however, is the secret sound spectrum of machines. Although humans cannot hear it, it is critical for predictive maintenance because certain sound frequencies reveal specific internal malfunctions. They tell technicians that a failure is likely to occur in the near future.

Like investigators recovering DNA evidence from the scene of a crime, the Neuron Soundware platform hears what humans can’t. It then provides useful information about what is happening—or what will happen—to a machine.

A living neural network that listens and learns

Anything with moving parts has multiple points of potential failure. Fortunately, each of these points has its own telltale signature. Emulating the human brain’s auditory cortex, Neuron Soundwave automatically performs sound analyses. Unique data input pre-processing techniques allow the neural network to learn and identify a sound’s important features quickly and with high confidence. Using a machine-learning algorithm, the system can then proactively monitor components that are likely to break.

Sound as a service

Neuron Soundware, a Prague-based startup, provides its sound analytics algorithms as a service, delivering a clear and early warning of emerging mechanical problems on even the most complex equipment.

“Waiting for a failure can be expensive—even dangerous,” says Neuron Soundware CEO and co-founder Pavel Konecny. “We can’t wait for an airplane engine to just stop working. You can’t have a printing press suddenly fail an hour before the trucks arrive. The ability to correct issues before they happen offers tremendous value for any industry.”

Breaking the sound barrier

Neuron Soundware technology works offline and online, transforming conventional production floors into smart factories. The system can also be integrated into existing software or third-party IoT platforms, which presents a rich set of opportunities for enterprise business software users. Sound and vibration sensors (microphones) can be quickly and cheaply installed on all types of machinery, enabling legacy equipment to be digitized without expensive upgrades.

From engines, turbines, and planes to escalators and printing presses, the Neuron Soundware approach works on a diverse set of machines.

The technology has caught the attention of a wide range of global clients, including Airbus, Siemens, E.ON, and Deutsche Bahn. “By hearing what humans can’t, we can quickly cut to the root cause of a potential machine breakdown—before it ever becomes a breakdown,” Konecny explains. “It’s adding a whole new dimension to the smart factory revolution.”

Why it’s big

The Neuron Soundware system represents the next generation in predictive maintenance. It works for a diverse set of machines, including engines, turbines, planes, escalators, and printing presses. It can be integrated into existing software or third-party IoT platforms, including enterprise business software landscapes. Early customers include Airbus, Siemens, E.ON, and Deutsche Bahn.

To learn more about how SAP is working with startups, get in touch with the SAP IoT Startup Accelerator and SAP.iO.


Rosina Geiger

About Rosina Geiger

Rosina Geiger is the Director of Startup Engagement at SAP. She has worked at the Hasso Plattner Institute in Potsdam before joining SAP in 2016 to establish the SAP IoT Startup Accelerator in Berlin and Palo Alto.