Research On Millimeter Wave Radar Imaging Based On MIMO Array

With the rapid development of millimeter-wave radar technology,millimeter-wave radar is receiving more and more attention as an imaging sensor in the civil field.Traditional optical sensors are limited by environmental factors leading to limited application scenarios,while millimeter wave radar emits electromagnetic waves to collect information and its small and lightweight features make millimeter wave radar have the advantages of all-weather operation,high measurement accuracy,far detection and low cost in imaging applications.Multi-input Multi-output(MIMO)array technology combined with millimeter-wave radar can improve the resolution accuracy by optimizing the antenna array,and make the millimeter-wave radar imaging effect greatly improved.This paper focuses on the principle of millimeter wave radar signal processing and imaging technology,and uses MIMO technology to improve the angular resolution of radar.Based on the study of millimeter wave radar to achieve two-dimensional imaging and point cloud imaging,we propose a method to achieve three-dimensional contour imaging using millimeter wave radar based on MIMO array and a point cloud imaging of gesture trajectory using millimeter wave radar,and classify and identify the trajectory images.First,the signal processing process of millimeter wave radar is introduced for millimeter wave radar imaging technology,and the principles of radar ranging,velocimetry,and angular measurement are investigated.The MIMO array is introduced to improve the angular resolution of millimeter wave radar,and the virtual aperture technique,time division multiplexing(TDM)signal emission method and signal processing model of MIMO array are studied.The Constant False Alarm Rate(CFAR)technique,point cloud imaging,and interpolation algorithms used in millimeter wave radar imaging are also investigated.Secondly,the two-dimensional range-angle imaging method of millimeter-wave radar based on MIMO array is studied,and the commonly used Direction of Arrival(DOA)estimation algorithms: Capon algorithm and MUSIC algorithm are mainly introduced.In order to improve the two-dimensional imaging effect,a linear sparse array antenna is proposed for two-dimensional imaging,and the linear sparse array antenna is studied and designed,and a physical prototype is made for experimental verification of the imaging effect.Based on the realization of two-dimensional imaging,a three-dimensional imaging method for millimeter-wave radar based on MIMO arrays is studied.Firstly,the signal processing process of three-dimensional point cloud imaging using millimeter-wave radar is investigated,the angle is calculated using the 2D-Capon goniometric algorithm,and the target information is extracted using the two-channel CFAR detection technique is proposed.On the basis of 3D point cloud imaging,this paper further proposes a new method for 3D contour imaging of millimeter wave radar based on MIMO arrays,which improves the angular resolution and increases the number of point clouds by increasing the number of antennas,interpolates and flattens the point cloud images to generate 3D contour images,and performs experimental verification using millimeter wave radar.Finally,an extended study of millimeter wave radar imaging is conducted,and a new method of generating hand gesture trajectory point cloud images based on MIMO array of millimeter wave radar is proposed,and the gesture recognition application is realized by image processing with the trajectory point cloud images as the original data.Unlike the traditional gesture recognition using Doppler images,the gesture recognition using gesture trajectory point cloud images proposed in this paper can not only greatly expand the gesture types,but also the differences between the trajectory images are obviously more conducive to improving the recognition rate.Finally,this paper uses millimeter wave radar to test a variety of gesture actions to generate point cloud images and uses classification algorithms to verify the accuracy of gesture trajectory point cloud image recognition.