Use of Machine and Deep Learning on RF Signals

Keyboard shortcuts

Action

Key

Play / Pause

K or space

Mute / Unmute

Toggle fullscreen mode

Select next subtitles

Select next audio track

Show slide in full page or toggle automatic source change

Seek 5s backward

left arrow

Seek 5s forward

right arrow

Seek 10s backward

shift + left arrow or J

Seek 10s forward

shift + right arrow or L

Seek 60s backward

control + left arrow

Seek 60s forward

control + right arrow

Decrease volume

shift + down arrow

Increase volume

shift + up arrow

Decrease playback rate

Increase playback rate

Seek to end

end

Seek to beginning

beginning

Loading Click here to add:

Information on this media

56 views

An RF Signal is an element that a human cannot see nor hear, but could be measured with many means today. Particularly, the Software-Defined Radio allows even people with a low budget to observe radio frequencies in real-time, and so make they capture different types of communications: AM/FM, Mobile & LPWAN communications, etc. There are many ways to classify all the technologies depending on the used frequency, used bandwidth, duty cycle, and patterns, but it is sometimes hard and/or time-consuming to recognize these technologies.
To resolve these types of challenges, we thought about using Machine & Deep Learning tools to optimize our classification, and we wanted to share with you our successes, mistakes, and other feedback. In addition to proper classification, RF emanations are also permanent in the air, and we will see that the same techniques can be applied to match harmonics, but also for side-channel attacks as well.

In this presentation, we will go through the steps of observing a signal, doing capture, talking about challenges to classifying the signal, and show techniques of using ML & DL from making a model, to using algorithms and available functions.

This will be an opportunity to talk about our infrastructure, today's results, failures, and future improvements.

Sébastien Dudek

Sébastien Dudek is a security researcher at Trend Micro and is also the founder of the PentHertz consulting company specialized in wireless and hardware security. He has been particularly passionate about flaws in radio-communication systems and published research on mobile security (baseband fuzzing, interception, mapping, etc.), and on data, transmission using the power-line (Power-Line Communication, HomePlug AV) like domestic PLC plugs, as well as electric cars and charging stations. He also focuses on practical attacks with various technologies such as Wi-Fi, RFID, and other systems that involve wireless communications.

Creation date: July 5, 2022

Speakers: Sébastien Dudek