Data processing

[shibiana]

Hi,

I get the raw data all less than 1 from BGT60TR13C development kit(the scale factor is not known), how can I reccover the  complete data from the data I get step by step? Sorry, I'm new in this field and not very familiar with the data processing steps. Are there any data processing examples? 

[Honey_D · Accepted solution]

Hello @shibiana ,

You can extract and record raw data from Radar Fusion GUI with the help of recording option available in the GUI. Regarding the data format details, please refer to 'Help' section in fusion GUI, it describes the details on all the parameters visible on the GUI. 

For data processing example, please refer to software documentation from \SDK\doc which describes all the theory of the algorithms offered. Also, refer to the application examples from \SDK\apps which you can build and test the board.

Hope this helps!

Best Regards,
Honey

[shibiana]

Hi, Thank you! 

I have some questions about the demo presence_detection.m. I want to ask about the meaning of /chirp_samples and *2 when we do the range_fft in figure.1.  In figure.2, I don't know the meaning of each parameter and these three fomulas.

[Honey_D]

Hello @shibiana ,

For figure 1, once you run the application, go through the variable workspace to understand the matrix arrangement of the data received. The /chrip_samples is to normalize the FFT operation and *2 is to eliminate the half (redundant) spectrum from the FFT out vector. 

For figure 2, please refer to the image attached as a explanation

Also, please refer to the sdk documentation for all the algorithm and data processing explanation /SDK/doc/html/pg_rdk_matlab.html#sct_rdk_matlab_running_the_algoexample 

Hope this helps!

Best Regards,
Honey

[shibiana]

Well! Thank you for your information! By the way, now, I try to measure the amplitude(displacement) of  human breathing.  I have 2 questions about this.

1. I can get the phase information by doing 128 FFT with zero padding to the signal (num_chirps_per_frame: 8 and num_samples_per_chirp: 64), then I choose the first 64 columns so it is a 8*64 matrix for a frame. Because the displacement is small, I want to use 8 chirps for the phase information. However, each row for the 8*64 matrix can get a phase information, do I need to average the phase information of these 8 rows? 

2. Suppose I get a Frequency phase diagram for a chirp,  how can I get the phase difference between successive chirps? 

[Honey_D]

Hi @shibiana ,

Please refer to the Algorithm supported by SDK section of documentation on how the parameters are evaluated from the IF beat frequency 

Hope this helps!

Best Regards,
Honey