Research On Super-resolution Algorithm Based On Traffic Radar

As a fundamental perception device for the intelligent application industry,millimeter-wave radar is widely used in the fields of autonomous driving,robots,and security monitoring.However,due to factors such as manufacturing costs,detection range,and detection accuracy,its promotion and application in the field of intelligent transportation still face significant challenges.In recent years,Ministry of Transportation of has proposed a package of solutions such as vehicle-road coordination and intelligent transportation for the intelligent application of highways.Therefore,there is an urgent need for high-performance millimeter-wave radar as a basic roadside intelligent perception device to solve problems such as high cost of long-distance traffic radar installation,short detection distance,and low resolution,and to provide support for the intelligent application of highways.This thesis conducted research on signal processing technology of high-performance millimeter-wave radar in the context of roadside intelligent perception on highways,aiming to improve the detection performance of radar in terms of distance and angle.The specific research contents are as follows:Firstly,this thesis addressed the issue of resolution loss when frequency modulation continuous wave(FMCW)millimeter-wave radar detects high-speed moving targets,and introduced the radonfourier transform(RFT)into the radar signal processing framework,and further proposed RFT approximation algorithm and blind speed sidelobe suppression algorithm.Simulation results showed that compared with traditional radar signal processing methods,the proposed RFT algorithm/RFT approximation algorithm based on FMCW radar has higher accumulation gain and higher distance/velocity resolution ability when detecting high-speed targets.Under the premise of reducing computational complexity,the degradation degree of accumulation gain of the RFT approximation algorithm is within 2 d B,and the maximum unambiguous speed that can be achieved by the proposed blind speed sidelobe suppression algorithm is above 200 km/h.Secondly,this thesis studied the distance super-resolution estimation technology of millimeterwave radar,proposed a time-domain multiple signal classification(MUSIC)distance super-resolution estimation scheme combined with RFT algorithm,and proposed a time-subarray smoothing technique in the time-domain MUSIC distance super-resolution estimation,which achieves distance superresolution estimation of multiple high-speed moving targets.Simulation results showed that the distance super-resolution estimation scheme proposed in this thesis can achieve resolution better than0.5 m,exceeding the Rayleigh limit of 0.66 m.The time-subarray smoothing technique can achieve super-resolution estimation of multiple moving targets,with a resolution better than 0.5m,and the number of target resolutions can reach 0.66 times the time sampling frequency.Finally,this thesis studied the joint distance-angle super-resolution estimation technology of millimeter-wave radar,and proposed a distance-angle joint estimation method based on twodimensional coherent single-shot signal and a fast distance-angle joint super-resolution estimation method based on region partitioning.Simulation results showed that the distance-angle joint estimation method based on the two-dimensional coherent single-shot signal can effectively achieve joint estimation of distance and angle.When the target distance difference is 0.5m and the angle difference is 5?,using the two-dimensional coherent algorithm can achieve a detection success rate of over 95% when the signal-to-noise ratio is better than-20 d B.The fast distance-angle joint superresolution estimation method based on region partitioning can achieve the same estimation effect while reducing the time complexity by PQ times.