IWRL6432: Machine Learning on the Edge with the mmWave Radar Device IWRL6432

Bruce Hu

Intellectual 570 points

Part Number: IWRL6432

Tool/software:

Hi,

For "Case-Study-1: Motion Classification", could you share data set(train/test) and source code for reference?

We'd like to run Machine Learning for mmWave Application.

Thanks

BRs

Bruce

10 months ago

0 Sharan S S 10 months ago

TI__Expert 4450 points

Hi,

We are looking into your query. Please allow us a couple of days to respond.

Regards,

Sharan

0 Pedrhom Nafisi 10 months ago in reply to Sharan S S

TI__Mastermind 24197 points

Hello Bruce,

For Motion and Presence demo, we currently only provide the model itself as a binary, and source code to interface with the model which all runs on-chip. In the upcoming weeks the Radar Toolbox will have the Machine Learning Package update where we provide all the steps and open-source/TI licensed tools in order to create your own model. It will include all the code for surface classification as an example so one can use the ML tools to go from recorded data all the way to a C optimized model implemented into a flashable demo.

Best Regards,

Pedrhom

0 Bruce Hu 10 months ago in reply to Pedrhom Nafisi

Intellectual 570 points

Hi Pedrhom,

Since IWRL6432 has limited computing power and RAM size, is it possible to run ML on the chip?

Thanks

BRs

Bruce

0 Pedrhom Nafisi 10 months ago in reply to Bruce Hu

TI__Mastermind 24197 points

Hello Bruce,

Yes it is possible to run ML models on chip on the main Arm Cortex-M4F. We have several examples of this which can be found in the Radar Toolbox in TI's Developer Zone.

https://dev.ti.com/tirex/explore/node?a=1AslXXD__2.30.00.12&node=A__Adf-bOuT9vYneQoGUdh4Vw__radar_toolbox__1AslXXD__2.30.00.12

Best Regards,

Pedrhom

Sensors

Sensors forum

IWRL6432: Machine Learning on the Edge with the mmWave Radar Device IWRL6432