Research On The Mm Wave FMCW Radar Sensing

Perception is becoming an increasingly important topic and challenge in important technological fields such as future 6G wireless communication,intelligent vehicles,and the Internet of Things.In the field of wireless communication,the spatial perception ability of millimeter wave radar and its corresponding signal design and processing algorithms will provide an important foundation for the integration of perception and communication.The all-weather and low-cost advantages of millimeter wave radar in the field of intelligent vehicles,coupled with the complementary advantages of laser radar and visual sensors,have become an indispensable part of autonomous driving and road environment awareness.Based on this background,this article conducts relevant research on millimeter wave linear frequency modulated continuous wave(FMCW)radar imaging.This paper uses the Ti2243 millimeter wave FMCW multiple transmitter and multiple receiver(MIMO)radar to achieve the separation of stationary background and moving target using stationary radar in the far field environment,and high-precision imaging using synthetic aperture radar(SAR)in the near field environmentFirstly,the Ti AWR2243 millimeter wave radar was used to collect data from real scenes.A signal model for extracting distance and velocity information through distance FFT and Doppler FFT using a single transmitter and single receiver antenna was derived.Based on this,a signal model for extracting angle information using phase differences between arrays in the case of multiple antennas was developed.Finally,calibrate the radar data collection.When using radar to separate a stationary background from a moving target,it was found that when a moving object obstructs a partially stationary background during movement,the moving target may have an impact on the signal illuminating the obstructed stationary background,which may result in the stationary background being mistakenly recognized as a moving target or being completely obscured and unable to be detected.At the same time,for some stationary backgrounds with large radar cross section(RCS),the spectrum leakage phenomenon is still severe during the Doppler Fourier transform process,which may lead to some stationary backgrounds with large RCS being incorrectly identified as moving targets.For this purpose,this work proposes an offline TRPCA processing algorithm that separates stationary backgrounds and moving targets,from radar signal preprocessing to tensor robust principal component analysis(TRPCA)decomposition.The application of low rank models fully explores the relationship between pre and post time radar signals.We have overcome the problem of moving targets obstructing a stationary background in practical scenarios,which may result in the stationary background being unable to be detected or mistakenly recognized as a moving target,as well as the problem of being mistakenly recognized as a moving target due to spectral leakage when the stationary background has strong reflection.By introducing an online mechanism,real-time processing of newly received data can be achieved,and timely updates can be made through the online processing architecture when the static background changes.The effectiveness of the algorithm is verified by the data validation collected from the real environment in the experiment,and the performance of the working algorithm is superior to the existing algorithm through analysis and comparison in the simulation.Secondly,the difference of antenna array model in near-field environment is studied,and the signal model based on FMCW signal MIMO SAR is established and equivalent in near-field environment,so that the environmental target can be quickly imaged.Using SAR to sample in space and imaging targets in real environments through experiments based on the aforementioned near-field model.