This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

How to increase the number of detected points of an object in a frame

Part Number: IWR6843ISK-ODS
Other Parts Discussed in Thread: IWR6843ISK

Tool/software: TI C/C++ Compiler

Hi TI Team,

We are using IWR6843ISK-ODS device and 68xx_Sense_and_Direct_HVAC_Control lab in our application.

We are trying to detect human when passed in-front(parallel to radar) of the radar(from -x to x direction) at a distance of 0.5 to 1m from radar(y distance).

We are using existing KNN algorithm given by TI. But during our testing, we found that most of the time RADARDEMO_featExtract_run function is failing at below condition

if ( featExtrInput->numPntPerTrack < featExtractInst->minNPtrPerTrack )


We found that number of detected points are less(mostly varying between 4-7) than minimum number of points required. But we are getting around 80-100 points in the point cloud. When we checked the point cloud, we found that radar points are distributed in the space rather than our ares of interest.

We would like concentrate our radar points to be focused in the area of interest as shown in the below figure.

(X : -1 to 1), (Y: 0 to 4) (Z : -2 to 2)

We can filter out the data using scenary parameters, but requirement is focus the chirps in that box dimensions only. This will help us to get more detected points from the object.

Could you please help how to achieve this requirement using the above demo application.?

Could you please suggest suitable configuration for this application?

Thanks

Sudharshan

  • Hi Sudharshan,

    The HVAC demo only runs on the IWR6843ISK - it does not run on the IWR6843ISK-ODS. Because there are different antenna patterns, the ODS will not have a "correct" point cloud.

    Regards,

    Justin