Stream-Process a 64-Microphone Array to Locate Bearing Faults in Real Time
Overview
What this challenge is about.
Implement a 64-channel overlap-add pipeline to beamform bearing fault locations in real time and earn a verifiable certificate.
The scenario
The company sells subscription condition-monitoring hardware to Nordic pulp-and-paper mills, where a single unplanned bearing seizure can halt a line that costs tens of thousands of euros per hour of downtime. Their edge sensors must localize faults on-device because mill networks are bandwidth-constrained and cannot stream raw 64-channel audio to the cloud.
The Brief
What you'll do, and what you'll demonstrate.
Build an FFT-based delay-and-sum beamformer that locates a 64-channel, 48 kHz microphone array's sound source in real time on a low-power embedded ARM device.
Earning criteria — what you'll demonstrate
- Design an overlap-add streaming FFT pipeline that maintains continuity across frame boundaries
- Implement delay-and-sum beamforming over a 2D azimuth-elevation direction grid from array geometry
- Optimize signal-processing code to hold real-time throughput on a constrained ARM device
- Validate localization accuracy against labeled ground truth and report it as a defensible error distribution
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Aligned coursework coming soon.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
- Fft
Apply fft to solve real industry problems and demonstrate production-level capability.
- Signal Processing
Apply signal processing to solve real industry problems and demonstrate production-level capability.
- Beamforming
Apply beamforming to solve real industry problems and demonstrate production-level capability.
- Cpp Programming
Apply cpp programming to solve real industry problems and demonstrate production-level capability.
- Embedded Systems
Apply embedded systems to solve real industry problems and demonstrate production-level capability.
- Streaming Algorithms
Apply streaming algorithms to solve real industry problems and demonstrate production-level capability.
Careers
Career paths this challenge builds toward
Completing this challenge demonstrates skills that transfer directly to these roles:
Embedded Signal-Processing Engineer
This challenge mirrors the core of embedded DSP work: take a math-heavy localization algorithm and make it run within tight time and power budgets on real ARM hardware, with measured accuracy and throughput you can defend to a team.
This challenge sharpens
- fft
- beamforming
- embedded-systems
Real-Time Audio Software Engineer
Streaming overlap-add processing with no boundary artifacts is the daily reality of real-time audio engineering; here you build and verify that continuity end to end, exactly as you would for low-latency audio products.
This challenge sharpens
- signal-processing
- streaming-algorithms
- cpp-programming
Predictive-Maintenance Systems Engineer
Localizing fault signatures on-device under bandwidth constraints is central to industrial condition monitoring; this challenge gives you the full path from labeled fault data to a validated, deployable edge implementation.
This challenge sharpens
- beamforming
- embedded-systems
- signal-processing