• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 30 subscribers
  • Views 1039 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

IWR1642BOOST: Where are tracking and clustering at real code ?

user5205609
Intellectual 960 points
Part Number: IWR1642BOOST

Hi,

I use IWR1642BOOST(ES1.0),with Tool Box 2.3.0 and SDK 1.2.0.5.
I get source codes of lab0011-pplcount (dss and mss) by CCS 8.xx.

In the previous question, I received an answer that mms is in charge of tracking and clustering.
Are they inside gtrack_step ( )?

I checked tidue71a.pdf,then I thought that Scenery Parameters and Allocation Parameters
are parameters related to tracking and clustering.

MmwDemo_appTask()( of task_app.c) calls gtrack_moduleStep()(Is Real code at
"...\\mmwave_sdk_01_02_00_05\packages\ti\alg\gtrack ?) .

The gtrack_step() do tracking and clustering by calling gtrack_module? (cf:gtrack_moduleAllocate())

Regards,
user5205609

  • 0
    TI__Mastermind 28055 points

    Hi Hanako,

    The tracking and clustering happens in gtrack_step(). This is in "...\\mmwave_sdk_01_02_00_05\packages\ti\alg\gtrack, but the People Counting project in CCS includes the files found at C:\ti\mmwave_industrial_toolbox_2_3_0\labs\lab0011-pplcount\lab0011_pplcount_pjt\radarDemo\chains\RadarReceiverPeopleCounting\mmw_PCDemo\gtrack. The code at these locations is identical.

    Inside gtrack_step(), 5 functions are called in this order:

    1. gtrack_modulePredict(inst);
    2. gtrack_moduleAssociate(inst, point, mNum);
    3. gtrack_moduleAllocate
    4. gtrack_moduleUpdate
    5. gtrack_moduleReport

    Predict() predicts where the tracked people will be in the next frame using a Kalman filter. Associate clusters points in the point cloud with predicted location of the tracked objects. Allocate clusters points that haven't been associated in an attempt to create new tracked objects. Update corrects the predicted location of the target and updates the Kalman filter. Report puts the data in the output struct. 

    When a person walks into the field of view of the sensor, their points will first be clustered in the allocation step. After that, predict, associate, and update steps will be responsible for maintaining their location.

    Regards,

    Justin

  • 0 in reply to Justin Curewitz
    Intellectual 960 points

    Hi,

    Thank you very much.

    I will check the codes under mmwave_industrial_toolbox_2_3_0.

    Regards,

    user5205609