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.

IWR1443: Sensor precision at close range

Part Number: IWR1443

Hello,

Our application of the IWR1443 sensor requires high precision (~1mm) at close range (<1m) to sense a somewhat rapid moving liquid level. This is something we thought the sensor could achieve based on the "micron accuracy" statements made in many of the videos describing the product.

However with the EVM that we are currently using - and according to some of the online papers - the range resolution cannot be set to less than 3.75 cm. I understand that with radar sensors, accuracy is a function of chirp length, so I was wondering if there is a way to reduce the chirp length, possibly compromising accuracy at range, to increase accuracy at close range. In addition, based on tests using the EVM, it seems that even though the sensor will only show us peaks (representing recognized objects) at 3.8cm increments, it theoretically might be able to estimate more precise distances between as, if the true distance of an object is, say, 9cm away, it might show a high peak at 8cm and a slightly lower but still high peak at 12cm. This suggests that even though it can only really detect objects in increments (by its physical chirp-length limitations), at close range specifically, it always seems possible to estimated the true distance based to two peaks. So essentially, we were wondering if either there is a way to increase its accuracy directly - perhaps at the cost of long range abilities - or, if there is a more efficient way to reliably calculate the true distance between two peaks shown on the distance/signal strength graph.

Thank you 

  • Hello,

    This thread might be of interest for you to read :

    e2e.ti.com/.../599055


    Hope this help !

    Christophe
  • Thanks Christophe !

    Hello Saurav,

    In addition to the thread mentioned by Christophe, please also look at the following post which talks about the considerations related to high accuracy measurements.

    e2e.ti.com/.../597434

    Accuracy is a function of range resolution, SNR and also your FFT resolution (also known as inter range-bin resolution). 1 mm accuracy should be achievable but as mentioned in the above post, for micro meter accuracy, the target needs to be stationary for a few seconds while multiple measurements are taken. You'll have similar considerations for mm level accuracy as well.

    We have the 4K FFT Lab available (along with a step-by-step user guide) on TI Resource Explorer. This lab demonstrates how you can stitch 4 1024 point FFTs (hardware limit) in software to achieve a 4K FFT for higher range precision. You can download and import the project directly in CCS using the resource explorer.

    dev.ti.com/.../


    Regards
    -Nitin