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.

  • TI Thinks Resolved

mmWaveSesnsingEstimator

Prodigy 735 points

Replies: 3

Views: 113

There is a question is about the mmWaveSensingEstimator on ti.com and its precision (tolerance consideration).

Following quotation for SNR calculation has been found in our training material.

https://training.ti.com/sites/default/files/docs/mmwaveSensing-FMCW-offlineviewing_0.pdf

 

It’s about calculationg the maximum measurement time T_meas which depends of the number of chirps and chrip duration.

Furtheremore it’s obvious by the code of mmWaveSensingEstimator that the number of virtual antennas (num_virtual_rx) is considered.

 

For calculations of different systems we used a custom tool which uses instead of the measured time T_meas the system bandwith for SNR calculation.

Furhtermore we cosidered quantization of the ADC and the process gain of signal processing.

 

It’s clear that the mmWaveSensingEstimator is “ just” a estimator for a rough estimation.

Neverthless, I like to understand why TI is using the SNR quotaion with T_meas.

 

Can you elaborate a bit more on the background and the tolerances we have experienced by this quotation.

  • Guru 61090 points

    Hi,

    Yes, the mmwave sensing estimator provides a rough estimation.

    The more chirps are used in a radar frame, the larger SNR.

    Is the question why Tmeas was used in the dmax equation on slide #50?

    Thank you

    Cesar

    ----------------------------------------------------------------------------------------------------------

    Please check TI mmWave FAQs before posting a new query
    ----------------------------------------------------------------------------------------------------------

  • In reply to Cesar:

    hi Cesar,

    how to estimate the expected error? is there any formula and guidance to estimate the error?

    What alternative ways can be used to consider range than Tmeans?

    Franz

  • In reply to Franz Schmid:

    Hi, Franz:

    Tmeans can be replaced by 1/fs * numAdcSample * Num_loop, where fs is the ADC sampling time.  Basically it includes the sampling time and the processing gain for range FFT and doppler FFT. 

    In this model, only the thermal noise is included, and ADC quantization noise is not included in this model. 

    Best,

    Zigang

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.