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.

IWR6843: mmWave Sensing Estimator Discrepancies

Part Number: IWR6843

I do not understand how the TI mmWave Sensing Estimator calculates range interbin resolution and velocity interbin resolution. 
If I manually calculate distance resolution via d = c / 2B I get a different result from the tool. Likewise, calculating velocity resolution via v = lambda / 2T also results in a different value.

Example: Range Resolution

c = 3x10^8 [m/s]

B = 375.60[MHz]

d = c / 2B = 3x10^8 / (2 * 375.60x10^6) = 399.4x10^-3[m] or 39.94[cm] which does not match the above highlighted value.

Velocity Resolution

lambda = c / f = 3x10^8 / 60x10^9 = 5x10^-3[m] or 5[mm]

T = 66 * 69.07x10-6 = 4.55x10^-3[s] or 4.55[ms]

v = lambda / 2F = 5x10^-3 / (2 * 4.55x10^-3) = 548.4x10^-3[m/s] or 0.548[m/s]


Where do these differences come from? Am I misunderstanding the above parameters?

Cheers,

Curtis Belknap

  • Former Member
    0 Former Member

    Hello Curtis,

    Your calculations are correct for range and velocity resolution respectively (based on the physics). However the two values you are comparing to can be thought of as the effective resolutions (subsequent to sampling/FFT processing).

    So for velocity resolution if we take the resolution you calculated and multiply by (Nchirp_loops/#doppler fft bins) we arrive at the value indicated in the estimator.

    The same should apply for range: where deltaRbin = detaR * (#samples per chirp/#range fft bins).

    I would recommend that if you want to see how any of the other calculations are performed you can view the source code of the visualizer. The relevant calculations are in the input.js file:

    	var range_inter_bin_resolution = function(range_resolution, number_of_samples_per_chirp, num_range_fft_bins) {
    	    return range_resolution*number_of_samples_per_chirp/num_range_fft_bins;
    	};
    	
    	var velocity_inter_bin_resolution = function(velocity_resolution, min_num_of_chirp_loops, num_doppler_fft_bins) {
    	    return velocity_resolution*min_num_of_chirp_loops/num_doppler_fft_bins;
    	};

    These steps apply to the sensing estimator:

    https://e2e.ti.com/support/sensors/f/1023/t/797254?-FAQ-AWR1843-Where-can-I-find-the-source-code-for-the-mmWave-Demo-Visualizer

    Amanda

  • This, along with the realization that the BW used in the calculations is (Slope) * (# samples * sample period), fixes everything.
    My calculations still are not exactly matching the estimator, but I assume that is from numerical round-off/truncation issues.

    Thanks!

  • Former Member
    0 Former Member in reply to Curtis Belknap

    Curtis,

    Yes - that's referred to as the valid bandwidth (the bandwidth during which we sample) the excess ramp time refers to the period in which we're continuing to ramp but not sample. 

    Amanda