We recognize broken machines using sound

Prague, Czech Republic

Artificial Intelligence


Neuron soundware has developed a diagnostic technology which provides an early detection of broken machines. Humans are frequently used during the quality control of air-conditioning, gearbox, etc. A sound of the car engine can indicate the root issue to the serviceman. We can automate this analysis using our digital brain which emulates auditory cortex.

Neuron soundware technology focuses on the early detection and prediction of mechanical malfunction of machinery in manufacturing industry. The unique pre-processing of the input data via complex physical sensor model allows neural network to learn and identify the important features quickly with high confidence. Sounds and vibration are often the simplest method for early detection of mechanical malfunction. Neuron soundware developed intelligent IoT device, which provides

- Remote real-time detection of mechanical malfunction in machinery

- Cost reduction by more effective predictive maintenance

- Reduce unplanned downline of the machine

- Quality control of equipment diagnostics based on device sound, vibration, etc.

We participated in 3 months acceleration program StartupYard and won competition Idea of the Year 2016. Our solution allows you to optimise operation via more effective predictive maintenance and real-time anomalies detection. That is achieved via sophisticated audio analysis algorithms powered by methods of machine learning.