Edge-AI Adaptive Noise Cancellation

Inspiration

We were inspired to improve defence communication by reducing disturbing noises such as gunshots, explosions, aircraft, and vehicles while keeping speech clear.

What it does

Our system uses AI-based speech enhancement and adaptive filtering to suppress background noise in real time and improve speech intelligibility.

How we built it

We integrated NVIDIA Jetson Nano, two microphones, a communication headset, and an RF module. We use NLMS adaptive filtering and DeepFilterNet3 for real-time audio processing.

Challenges we ran into

Our main challenges were handling sudden noises, adapting to changing noise conditions, preserving natural speech, and achieving low-latency processing on embedded hardware.

Accomplishments that we're proud of

We developed an integrated hardware-software approach for edge-based noise cancellation, combining adaptive filtering and AI-based speech enhancement for defence communication.

What we learned

We learned about adaptive filtering, DeepFilterNet3, embedded AI, microphone integration, and real-time audio processing.

What's next for Edge-AI Adaptive Noise Cancellation

We plan to optimize performance on Jetson Nano, test the system under different noise conditions, improve speech clarity, and enhance reliable wireless audio communication.

Built With

• ai/ml
• deepfilternet3
• jetson-nano
• microphones
• nlms-adaptive-filtering
• python
• real-time-audio-processing
• rfmodule
• usb-headset