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: A problem about the Doppler-FFT on Design Guide: TIDEP-01000 "People Tracking and Counting Reference Design Using mmWave Radar Sensor".

Part Number: IWR6843
Other Parts Discussed in Thread: TIDEP-01000

Dear TI engineers and other experts, thank you for your attention.

The Design Guide: TIDEP-01000 uses range-FFT, capon beamforming in azimuth dimension, elevation beamforming and doppler estimation to get the 4D information of objects. But I have some questions about the Doppler estimates. Consider a situation, the distance information of multiple points is the same, but their velocity information is different, which means that there will be multiple peaks in the RD spectrum obtained by Doppler FFT at the same distance bin. Correspondingly, RA spectrum also has multiple different orientation information at the same distance. I would like to ask how to solve the problem of matching the azimuth information with the speed information? 

I am a beginner to FMCW mmWave radar and look forward to and appreciate the TI engineers and other experts taking the time to answer my questions.

  • Matching azimuth information with speed information in FMCW mmWave radar systems can be a challenging task, especially when dealing with multiple peaks in the RD and RA spectra at the same distance bin. One approach to addressing this issue is through advanced signal processing techniques and algorithms.

    One potential solution is to use sophisticated signal processing methods such as multi-target tracking algorithms, which can help associate the correct Doppler velocity measurements with their respective azimuth angles. These algorithms can help in tracking and associating multiple targets based on their range, velocity, and angle information over time.

    Another approach involves implementing advanced beamforming algorithms, such as adaptive beamforming techniques, to improve the resolution and accuracy of the azimuth and elevation angle estimates. By enhancing the spatial resolution of the radar system, it becomes easier to distinguish between multiple targets at the same distance but with different velocities.

    Furthermore, leveraging machine learning and pattern recognition techniques can also aid in accurately associating velocity information with azimuth angles. By training models on a diverse set of radar data, these techniques can learn to accurately match velocity information with azimuth angles, even in complex scenarios with multiple targets.

  • Thank you. But can you explain to me the approach TI used in this design guide(TIDEP-01000)? Looking forward to your professional and popular answers.

  • Hello,

    Personal understanding:The process described involves detecting points in range and azimuth space and then estimating the Doppler using capon beam weights and Doppler FFT for each detected point. The resulting Doppler information is stored in the L2 memory. This Doppler information is then combined with the point cloud generated during CFAR (Constant False Alarm Rate) and Elevation Estimation. The output for each point includes information on range, azimuth, elevation, Doppler, method of application, and signal-to-noise ratio (SNR). This method allows for the comprehensive analysis and visualization of the detected points in the radar data.

    Regards,

    Gary