• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 31 subscribers
  • Views 1894 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

IWR1443BOOST: Some Qusetions about Theory of Velocity Estimation using 77GHz FMCW Radar

Xia Dai
Prodigy 70 points
Part Number: IWR1443BOOST
Other Parts Discussed in Thread: IWR1443

Hi,

I bought TI's IWR1443 BOOST EVM a few months ago and I am studyding TI's mmWave materials to do some designs using IWR1443. I have studied TI's mmWave training series, the first topic of which is Introduction to mmwave Sensing:  FMCW  Radars. The corresponding PPT of the training video is as shown below.

Introduction to mmwave Sensing FMCW Radars.pdf

But I am confused about module 3: Velocity Estimation. The PPT as below shows that ω is the phase difference of the IF signal, and we can get the value of IF signal's phase using 1D-FFT,then we can perform an operation of difference between two adjacent phases of two consecutive chirps. So, according to the velocity estimation formula, we can compute the velocity of one object just by 1D-FFT. I don't understand why there is a need for 2D-FFT.Is the 2D-FFT just for the situation that there are two objects equidistant from the radar? If objects are at different ranges from radar, is 1D-FFT enough to compute the velocities?

Besides,I am also confused about this slide shown as above.Here,the phase difference ω1 and ω2 are the result of 2D-FFT when multiple objects are at the same range from radar.If the phase of  one object's IF signal do not rotate at certain angular frequency but changes according to the law of primary function, then just for this object, there will be more than one ω in the spectrogram of 2D-FFT rather than one. If then, how can we distinguish different objects' angular frequency in 2D-FFT's spectrogram?

Best Regards,

Xia Dai

  • 0
    TI__Mastermind 36505 points

    Hi, 

    1D FFT gives you the range of targets, and 2D FFT gives you the range and Velocity information.  If you have a hard time understanding our training video, you can find some books to get more details. 

    An example of book is given below:

    Donald Barrick, "FM/CW Radar Signals and Digital Processing"

    Best,

    Zigang

  • 0 in reply to zigang Yang
    Prodigy 70 points

    Hi Zigang,

    Thanks for your answer and recommanding this book to me.

    I am request the book on ResearchGate, and it may take some time. I have deduced the formula of FMCW radar：

    I understand 1D-FFT and 2D-FFT processing here. 1D-FFT shows the frequency of IF signal, which can be converted to the range of targets. 2D-FFT shows the frequency components of IF signal's phase, but only if the phase, such as vital sign signal's phase, changes in sine or cosine,  there exists unique ω by 2D-FFT for single target. If we detect human target's velocity and the displacement x(t) doesn't vibrate, but increases or decreases all the time, we will get more than one frequency component and as a result, we can't say which frequency component is the right one. So I think velocity estimation by 2D-FFT just used for the situations that targets move in sinusoidal motion. 

    Can I understant it in this way?

    Best Regards,

    Xia Dai

  • 0 in reply to zigang Yang
    Prodigy 70 points

    Hi Zigang,

    I read other literature today and I finally understand the velocity estimation in TI's training video.

    I find that I ignored the difference between 2D-FFT's usage of vital signs detection and that of other targets' velocity estimation. In vital signs detection, after 1D-FFT, we get IF signal's frequency spectrum and the peaks stand for targets. Then according to  multiple cycles‘s 1D-FFT, we can obtain IF signal's phases which are the input signals of 2D-FFT operation. While, in velocity estimation, the input signals of 2D-FFT are the 1D-FFT's complete results which are complex signals including both frequency and phase.

    Best,

    Xia Dai