IWR1443BOOST: Run processing algorithms outside the EVM

Angelo Claudio Re

Part Number: IWR1443BOOST
Other Parts Discussed in Thread: IWR1443, IWR1843BOOST, DCA1000EVM

Hi,

Is there an example or lab that can be used as a starting point to send through the UART ports the raw data of the radar signals collected by the IWR1443BOOST EVM ?
Doing so, will It be possible to use these raw data to use algorithms run by a PC through any kind of programming language like python, go, etc.

In terms of performance, would this approach make any sense ?
It is better or worse than running it inside the EVM exploiting the dedicated hardware ?

Best Regards,
Angelo

over 5 years ago

0 Vaibhav Mahimkar over 5 years ago

TI__Mastermind 23665 points

Hello

IWR1443 like other IWR sensors is meant to post process the FMCW data and provide higher level results for decision making.

Performing post processing on the device gives faster results and response time as compared to doing similar activity off chip.

Please help us understand the objective of your end application to guide you better.

Thank you,

Vaibhav

0 Angelo Claudio Re over 5 years ago in reply to Vaibhav Mahimkar

Intellectual 300 points

Hi Vaibhav,

My goal is to build an application that will track, classify and detect drones instead of people taking insights by the People counting and tracking lab.

Up to now, I'm trying to understand what will be the devices that I will use to do so. I'd like to create a benchmark in which I can highlight the performance between running the algorithms on and off chip.

Since the IWR1443BOOST EVM does not have a DSP, I think that I will use a IWR1843BOOST EVM.

Now, Are there any demo or lab in which the raw data are sent directly to the PC to perform the off chip computation ?

Regards,
Angelo

0 Justin Curewitz over 5 years ago in reply to Angelo Claudio Re

TI__Mastermind 28055 points

Hi Angelo,

The raw data + point cloud can be sent to the device using the Out of Box (OOB) demo. This will require a DCA1000, but is very useful because you can see the point cloud generated and the raw data from exactly the same frame. Our Long Range People Detection demo is built on the OOB, so it is capable of detection and tracking.

You can find documentation if you have the SDK installed, in our doxygen: file:///C:/ti/mmwave_sdk_03_04_00_03/packages/ti/demo/xwr68xx/mmw/docs/doxygen/html/index.html#LVDSStreamingNotes

Regards,

Justin

0 Angelo Claudio Re over 5 years ago in reply to Justin Curewitz

Intellectual 300 points

Thanks Justin

Regards,
Angelo

0 Angelo Claudio Re over 5 years ago in reply to Justin Curewitz

Intellectual 300 points

Hi Justin,
I have a doubt.

You said that I will require a DCA1000EVM to collect the raw data + point cloud.

Is it not possible without this device ? Is it mandatory ? If so, why ?

Regards,
Angelo

0 Justin Curewitz over 5 years ago in reply to Angelo Claudio Re

TI__Mastermind 28055 points

Hi,

DCA is not required for point cloud, but it is required for raw data.

Regards,

Justin

Sensors

Sensors forum

IWR1443BOOST: Run processing algorithms outside the EVM