AI & SDR - Analysing Radio Transmissions with Machine Learning

Could you make a ann app that listens to radio, then summarizes what happened around you when you were away?

My students Dianne, Pauli, Mikko and Juho are going to give it a try. The goal is a proof of concept, not an operator grade box with its own theme song.

Update: MuikkuRF blog, kajami/SDR-project on GitHub.

They are using HackRF software defined radio (SDR) and free software such as Linux, gqrs and maybe GNURadio. For the machine learning (ML) and AI part, they are looking at Python and Tensorflow.

You can follow them on MuikkuRF blog and read the code in SRD-project GitHub repository.

This project is part of infrastructure project, a course I'm teaching with Harto Holmström in Haaga-Helia.

They have collected some samples, taking the first steps in building a signal library. Here is some early training material.

A pre-trained AI was available to recognize persons (seems that I'm not a person). Now, just make it recognize some signals...

MuikkuRF blog

kajami/SDR-project on GitHub - The main source repository for the project.

Some older machine learning material for the project kajami/SDR-image-classification GitHub

Adminstrivia: this article has been updated after publishing.