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.

AWR2944EVM: I can only see one target of a metal object is 60cm in width & 180cm tall that is 2 meters away

Part Number: AWR2944EVM
Other Parts Discussed in Thread: AWR2944, AWR1642, AWR1843

Tool/software:

I have AWR2944EVM, which I was able to successfully connect to the Demo visualizer tool after flashing TDM application.

I put the sensor in a setup where there is a big metal object in front of it. This metal object is 60cm in width & 180cm tall. It is around 2m away from the sensor. On the Demo visualizer, I can see only 1 target!! I would expect to have more targets from this object.

- Do you think that I need to make some changes to the configuration?

- Just to confirm, is the default mode a target mode, or is it an object mode "Cluster of targets" ?

  • Hi Ahmad,

    We do not have any clustering algorithm included in the SDK TDM demo application. DDM demo has better performance when compared to TDM demo. Since all transmitters are active at the same time in DDM application, SNR will be high. In general, you observe more detections using DDM when compared to TDM. I suggest you to use DDM demo instead of TDM demo for the same scene. Refer to (+) [FAQ] What are the benefits and SW migration effort moving to AWR2944 from AWR1642/AWR1843? - Sensors forum - Sensors - TI E2E support forums for the high level difference of both the processing chains.

    If the Radar Cross Section area is too large, you can receive multiple reflections from a single target. You can refer to Programming Chirp Parameters in TI Radar Devices (Rev. A) to change your configuration from long range to short range if you are focusing on Short Range Detections.

    Regards,

    Samhitha

  • Hello Samhitha,

    I did it based on your recommendation and I have the following points (Sorry in advance if they a bit a lot):

    • I believe that the number of points are now more using DDM. But I notice that the spread of the points on the elevation level is below my expectations as I mentioned the Object is 2m long and I was putting the sensor on 1m high so I would expect that the points should be between -1m: 1m. Current output is as shown in the image.

    • I notice also that I always need to go to visualizer and send the config to mmwave device so that it starts sending the points out. I would like that the output comes out directly without the need to send the config.
    • I used the following configuration. Is there any other configuration or SW binaries that you would recommend. My Target FOV max range is 35m but I seek high resolution to be able to detect small objects as of size of a ball of 5cm radius. 

  • Hi Ahmad,

    Firstly, AWR2944 demo is intended for Automotive applications. The configurations that are provided in the SDK, are for long range detections.

    Elevation angle resolution at boresight is given by θ = 2/N = 2/4 = 0.5 radians = 28.6 degrees. (N is number of virtual antennas in elevation direction). For detailed understanding, refer to MIMO Radar (Rev. A). To get a fine elevation resolution, you need more antennas in elevation direction.

    I notice also that I always need to go to visualizer and send the config to mmwave device so that it starts sending the points out. I would like that the output comes out directly without the need to send the config.

    If you want to use the demo visualizer, you have to send the configuration via CLI for every run. If you are not going to use demo visualizer, follow the steps given below to bypass the CLI -

    • Navigate to C:\ti\mmwave_mcuplus_sdk_<ver>\mmwave_mcuplus_sdk_<ver>\ti\utils\cli\src\cli.c
    • Uncomment CLI_BYPASS macro
    • Rebuild cli library
    • And rebuild the demo application
    I used the following configuration. Is there any other configuration or SW binaries that you would recommend. My Target FOV max range is 35m but I seek high resolution to be able to detect small objects as of size of a ball of 5cm radius. 

    Detection of the target is based on the RCS, Tx output power, etc. I suggest you to check about the Radar Equation. We haven't tested AWR2944 demo detecting a small ball at 35m. Fine tune your parameters by referring to Programming Chirp Parameters in TI Radar Devices (Rev. A) based on your requirements.

    Regards,

    Samhitha

  • Hi Samhitha,

    Thanks for your support.

    Detection of the target is based on the RCS, Tx output power, etc. I suggest you to check about the Radar Equation. We haven't tested AWR2944 demo detecting a small ball at 35m. Fine tune your parameters by referring to Programming Chirp Parameters in TI Radar Devices (Rev. A) based on your requirements.

    I had a look on the document that you shared. For example, I'm interested in using USRR. I would like to test the performance with it, since it has a highest range resolution and the max unambiguous Range is 22.5m, so around the max range that I'm looking for with is 30m.

    Now, how could I force the USRR configuration? :

    • Should I make some changes over DDM package? If yes, how to do so?
    • Should this be similar to using Demo visualizer and changing the scene selection resolution?
  • Hi Ahmad,

    We have USRR application developed using AWR1642/AWR1843. If you are looking for the USRR application, you can check Auto SRR Users Guide and Medium Range Radar Users Guide. You have separate visualizer for these applications at radar_toolbox_<ver>\tools\visualizers.

    AWR2944 is mainly used for long range detection of around 200m. If you want to use AWR2944 for USRR application, you have to finetune the front end configuration sent via UART. You can check the config files present at C:\ti\mmwave_mcuplus_sdk_<ver>\mmwave_mcuplus_sdk_<ver>\ti\demo\awr294x\mmw\profiles\ddm_awr2944. Configuration can be updated based on your requirements without any additional changes in the DDM processing chain. 

    Should this be similar to using Demo visualizer and changing the scene selection resolution?

    You should be able to use this feature present in the latest Demo Visualizer. Visualizer uses same math to provide the scene-based parameters.

    Regards,

    Samhitha