• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 1 answer
  • Subscribers 29 subscribers
  • Views 358 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

IWR1642 vehicle detection One vehicle is identified as two vehicles.The vehicle was lost after stopping and starting at a traffic light

ctc ctc
Intellectual 290 points
Other Parts Discussed in Thread: IWR1642

Hello,

I use IWR1642 for vehicle detection and my routine is lab0013_traffic_monitoring_16xx.

In the process of vehicle tracking, one car will be identified as two cars, which parameters are related to this.

Part of the vehicle will be lost when it stops at a red light and starts at a green one.What configuration parameters does this relate to and how do I modify it.

      I'm in a hurry,Hope TI official guidance, thank you.

  • Former Member
    0 Former Member

    Hello,

    Please see the Gtrack tuning guide here https://dev.ti.com/tirex/explore/node?a=VLyFKFf__3.6.2&node=AN0AJy9ls3N096oILhoGZw__VLyFKFf__3.6.2

    The tracker does not track non-moving objects this is why you are losing the vehicle you can increase the static 2 free threshold to maintain the track at the light and prevent it from being lost.

    For the one car being identified as two cars you can change the allocation parameters and increase the range so the points aren't split into two tracks.

    Amanda

  • 0 in reply to Former Member
    Intellectual 290 points

    Hello ,

    I want to know my maxNumPoints set to 250, and drive the vehicle a lot, but I'm rarely amount of point cloud data, this is what reason, I maxNumTracks set to 60, and the actual road, 60 metres away from the radar, there are about 20 vehicles four lanes, but at the same time visual inspection to identify to vehicle difference is very big, what reason is this?The farthest point cloud data is also about 100 meters, and that distance point is very small, to about 70 meters point to become more, about 50 meters to identify the vehicle, this has obstacles to do the length of the vehicle queue, the length of the vehicle queue also to more than 80 meters, how to solve?

  • 0 in reply to ctc ctc
    TI__Mastermind 23570 points

    Hello

    Are you saying the point cloud data  is in tune with what you see in the scene

    but the tracker is not showing you the required grouping..

    Thank you,

    Vaibhav

  • 0 in reply to Vaibhav Mahimkar
    Intellectual 290 points

    Hello ,

    My radar sensors now reflect less point cloud data every 50 milliseconds，And Point cloud data appears only when the distance is close to the radar.This is the most serious problem,I need reflection point 100 meters away, and I need a lot of reflection points.

    For example, if I want 120 meters, I have a lot of radar reflection points

    Which parameters do I need to configure?Help me circle it?And how should it be configured?

    thanks!

  • Former Member
    0 Former Member in reply to ctc ctc

    Hello,

    Configuring a maximum number of tracks and points doesn't mean that you will automatically get that many points or tracks.

    If you are not seeing enough points or tracks please refer to the previous tuning guide to lower the threshold to enable easier allocation of tracks.

    Amanda

  • 0 in reply to Former Member
    Intellectual 290 points

    Hello，

    I refer to the previous tuning guide,Here are some questions:

    1、how to replay data captured using the an xWR1642 EVM?

    2、Whether the files under the GUI\sample_fhist directory are replay data for the traffic Monitoring lab？

    3、using the an xWR1642 EVM，How do I generate files such as GUI\ Sample Fhist？

  • 0 in reply to Former Member
    Intellectual 290 points

    Hello，

    And The previous tuning guidelines have nothing to do with the distance from which point cloud data begins to appear?

    Here are the tuning guidelines,I import the default configuration following the steps in the tuning guide,

    The number of point clouds is below 80 meters,I want 120 meters to appear how do I optimize it?

  • Former Member
    0 Former Member in reply to ctc ctc

    Hello,

    To increase the range of the points you can detect you can decrease the CFAR threshold until you see the points needed. Note this would likely increase the noisy points you see in the near range. You are effectively reducing the SNR threshold needed for point detection.

    You should also consider that it may not be possible to easily detect a vehicle at 120m using 1642. You can instead consider an 1843 or 6843 device which supports 3TX beam forming which can be used to more easily detect long range objects. You can see that using beamforming in the long range people detection lab enables detection of humans at 100+m so vehicles should also be easily achievable.

    Amanda