Classify RF Signals Using AI (Project 245)

[robertogl]

Contribute to the discussion by asking and/or answering questions, commenting, or sharing your ideas for solutions to project 245

[Kofigo375]

Hello , I am an undergraduate student at KNUST in Ghana. I am interested in working and collaborating on this project.

[MikeMcLernon]

Excellent -- please let us know if you need guidance along the way.

[Kofigo375]

Alright sure

…

On Wed, 17 May 2023 at 17:01, MikeMcLernon ***@***.***> wrote: Excellent -- please let us know if you need guidance along the way. — Reply to this email directly, view it on GitHub <#81 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ATKRWZ7P3FA6GWNZOEYPMPLXGT745ANCNFSM6AAAAAAU22MDUI> . You are receiving this because you commented.Message ID: <mathworks/MathWorks-Excellence-in-Innovation/repo-discussions/81/comments/5930176 @github.com>

[Kofigo375]

I need access to some paid Matlab courses (eg. deep learning with Matlab). I contacted my faculty, but they have no Matlab license. What should I do? On Wed, May 17, 2023 at 5:12 PM Ephraim Osei Kusi ***@***.***> wrote:

…

Alright sure On Wed, 17 May 2023 at 17:01, MikeMcLernon ***@***.***> wrote: > Excellent -- please let us know if you need guidance along the way. > > — > Reply to this email directly, view it on GitHub > <#81 (reply in thread)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/ATKRWZ7P3FA6GWNZOEYPMPLXGT745ANCNFSM6AAAAAAU22MDUI> > . > You are receiving this because you commented.Message ID: > <mathworks/MathWorks-Excellence-in-Innovation/repo-discussions/81/comments/5930176 > @github.com> >

[MikeMcLernon]

I will check with the appropriate Ghana account representative to determine the best next step. Will get back to you as soon as I learn new information.

[Kofigo375]

Thanks. Another concern I have is about the deadline. I want it to be extended to October since I intend to actually begin on July, as I indicated on the form. I’m using from now till July to familiarize myself with the necessary background or prerequisites.

…

On Thu, 18 May 2023 at 12:37, MikeMcLernon ***@***.***> wrote: I will check with the appropriate Ghana account representative to determine the best next step. Will get back to you as soon as I learn new information. — Reply to this email directly, view it on GitHub <#81 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ATKRWZ3H43FCZBIFPTW6WITXGYJXTANCNFSM6AAAAAAU22MDUI> . You are receiving this because you commented.Message ID: <mathworks/MathWorks-Excellence-in-Innovation/repo-discussions/81/comments/5938540 @github.com>

[robertogl]

Ok, no problem with that. Thanks for letting us know!

[pqoe]

Hi, I'm trying to work out some issues in the project. I plan to use my pluto SDR for this but before when generating signals, i notice they must be same size otherwise i get an error in the frequency combiner function.?
For example my LFM signal has 250 samples as a size but my NR 5G signal has 2465000 samples as size, maybe i need to do some zero padding? but this seems very sloppy, is there any otherway to resolve this issue?
Thanks BR

[MikeMcLernon]

Hi -- If you zero pad the LFM signal to be 2,564,000 samples, then you will reduce its power a great deal. Can you generate an LFM signal that has that many samples?

[pqoe]

Done it, looks good. And btw, the multiband combiner is a very interesting function, didn't know it before :)

[MikeMcLernon]

Excellent -- and glad you like the multiband combiner. I will let the developer know that you approve. :)

[pqoe]

Hi, Mike, how you doing? i hope you help me out

1/ What length does the real world live IQ RF capture needs to be ?

2/ Finding a 20 MHz 5G in live real world spectrum is not realistic, most teleco offers 50 to 100 MHz bandwdith for 5G, which poses a serious challenges for most SDR since they are limited in bandwidth capacity, an IQ capture such as the one in the example (5G and 4G same IQ bandwidth) is not feasible. Does the capture needs to include both signals in the same time ? of can i capture a 5G signal first and a 4G second in a separate manner? doesnt this impact the AI algorithm?

thanks for your help !

[MikeMcLernon]

Hi,

Very good questions -- thanks for reaching out to me.

A few thoughts on them:

1. Several resource blocks should be sufficient. I would recommend that you look at the lengths of the signals in the example at https://in.mathworks.com/help/comm/ug/spectrum-sensing-with-deep-learning-to-identify-5g-and-lte-signals.html.
2. Yes, you should be able to capture a 5G signal separately from a 4G signal, then combine them. You can use the Multiband Combiner, at https://in.mathworks.com/help/comm/ref/multibandcombiner.html, to do this. Also, you can use 5G Toolbox and LTE Toolbox to synthesize the signals that you need.

[aswinsankesh]

Hi I am Aswin Sankesh G S, Final Year Undergraduate from Electronics and Communication Engineering Department, Thiagarajar College of Engineering, Madurai, TN, India. Our team is impressed with this project and interested in collaborating on this project. We have starting working on this project right now.

[MikeMcLernon]

Hi Aswin, Your enthusiasm is wonderful – hope that you have a great time with the project! Best, Mike From: aswinsankesh ***@***.***> Sent: Saturday, November 11, 2023 6:29 AM To: mathworks/MATLAB-Simulink-Challenge-Project-Hub ***@***.***> Cc: Mike McLernon ***@***.***>; Comment ***@***.***> Subject: Re: [mathworks/MATLAB-Simulink-Challenge-Project-Hub] Classify RF Signals Using AI (Project 245) (Discussion #81) Hi I am Aswin Sankesh G S, Final Year Undergraduate from Electronics and Communication Engineering Department, Thiagarajar College of Engineering, Madurai, TN, India. Our team is impressed with this project and interested in collaborating on this project. We have starting working on this project right now. — Reply to this email directly, view it on GitHub<#81 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/A6K4O3B4ZDSUC5OCZDPB66DYD5OQ7AVCNFSM6AAAAAAU22MDUKVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM3TKNBQGAZTO>. You are receiving this because you commented.Message ID: ***@***.******@***.***>>

[aswinsankesh]

We have inquiries related to dataset collection for our project. Our intention is to detect Wifi (2.4GHz) and Bluetooth signals. Our plan involves utilizing the ADALM PLUTO SDR to both generate and receive these signals.

However, uncertainties have arisen concerning the parameters required for generating Wifi signals. While considering the use of the Matlab Wireless Waveform generator for this purpose, we encounter numerous parameters such as packets, idle time, bit source, oversampling factor, scrambler initialization, window transition time, etc.

To proceed, we seek typical values for these parameters in order to generate authentic Wifi signals using this application. Assistance in locating relevant documentation containing these parameter values or suggesting realistic values for our project would be greatly appreciated.

[mw-mmclernon]

Hi Aswin, Please see the wlanWaveformGenerator<https://in.mathworks.com/help/wlan/ref/wlanwaveformgenerator.html> doc page to learn the valid ranges of values for the parameters you’re uncertain about. Then you can use the Wireless Waveform Generator App to generate a “prototype” WLAN waveform. Then use the “Export to MATLAB script” option on the app to create the MATLAB script for that prototype waveform. Then wrap loops around the code to vary the parameters you want to vary in order to create your data set. You can do the same thing with any RF impairments you impose on the signal. Best regards, Mike From: aswinsankesh ***@***.***> Sent: Wednesday, November 15, 2023 1:57 AM To: mathworks/MATLAB-Simulink-Challenge-Project-Hub ***@***.***> Cc: Subscribed ***@***.***> Subject: Re: [mathworks/MATLAB-Simulink-Challenge-Project-Hub] Classify RF Signals Using AI (Project 245) (Discussion #81) We have inquiries related to dataset collection for our project. Our intention is to detect Wifi (2.4GHz) and Bluetooth signals. Our plan involves utilizing the ADALM PLUTO SDR to both generate and receive these signals. However, uncertainties have arisen concerning the parameters required for generating Wifi signals. While considering the use of the Matlab Wireless Waveform generator for this purpose, we encounter numerous parameters such as packets, idle time, bit source, oversampling factor, scrambler initialization, window transition time, etc. To proceed, we seek typical values for these parameters in order to generate authentic Wifi signals using this application. Assistance in locating relevant documentation containing these parameter values or suggesting realistic values for our project would be greatly appreciated. — Reply to this email directly, view it on GitHub<#81 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AD2D5P5PUQ2W4EOMTHJZ7JDYERRVFAVCNFSM6AAAAAAU22MDUKVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM3TKNZTGM3DS>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.******@***.***>>

[aswinsankesh]

Thanks for the documentation Mike. We will try to use this information and come back to you for challenges if any.

[rajendra1124]

Hi, I am (Rajendra) a student at GMU, and would like to collaborate on this project.

[MikeMcLernon]

Sounds fantastic, Rajendra. Let me know if you need help along the way.

[Keerthana-1214]

Hi,
I am Keerthana Maremmagari, a student at G. Narayanamma Institute of Technology and Science (GNITS), Hyderabad. I am excited to inform you that we will be collaborating on this project. As we proceed, there may be times when we encounter challenges or need further clarification on certain aspects. I hope you will be there to assist and provide guidance.