AudioLog Predictive Maintenance

What is your project about?

AudioLog is an "industrial stethoscope" that monitors machine health by listening to internal sounds. Using the Arduino UNO Q, it detects high-frequency acoustic anomalies in motors and pumps, identifying early signs of failure like bearing wear or misalignment before a breakdown occurs.

Why did you decide to make it?

Unplanned industrial downtime is incredibly expensive. I created AudioLog to move maintenance from "reactive" to "predictive." By using Edge AI, the system analyzes complex audio data locally on the factory floor, providing a low-cost, non-invasive way to protect multi-million dollar machinery without needing cloud bandwidth.

How does it work?

A PDM microphone captures raw audio, which the UNO Q's Linux processor transforms into a spectrogram using FFT (Fast Fourier Transform). An Edge Impulse AI model analyzes this "sound fingerprint" for anomalies. If a fault is detected, the Arduino side instantly triggers a safety relay and updates a local OLED dashboard to alert maintenance crews.