Research On Automotive Millimeter Wave Radar Signal System And Super Resolution Angle Measurement Based On MIMO

In recent years,vehicles have become increasingly important in our daily life.But the increasing use of vehicles has caused a serious problem,that is traffic accidents.Human factors are the main cause of traffic accidents,accounting for about 70%~90%,most of which are inattention or lacking enough reaction time.Therefore,radar is used as an import auxiliary tool in automobile driving.Also,with the rocketing development of intelligent transportation system,radar system is applied to the automatic assistant driving system as an essential part.The millimeter-wave radar can meet the requirements of vehicles for adaptation to all-day climate because of the low atmospheric attenuation and loss in the transmission window and strong penetrability.Unlike cameras and lidars,they are vulnerable to severe weather such as rain,snow or haze.Coupled with the small size and light weight of millimeter-wave radar sensors,automotive millimeter-wave radars have increasingly become the research direction of scholars at home and abroad.Angular resolution is an important indicator to distinguish between good and bad performance of automotive millimeter wave radar.If the angular resolution is not small enough,the automobile millimeter-wave radar may not be able to distinguish between distant and closely spaced targets,such as vehicles in front and oncoming vehicles or pedestrians in front.For automotive millimeter-wave radar,this is very serious.In addition,high angular resolution is very important for radar imaging,so how to improve the automotive radar angular resolution has become an urgent problem to be solved.There are only two ways to improve angular resolution: virtual aperture and super-resolution algorithms.This paper combines the two methods to improve the angular resolution as much as possible.First,an improved genetic algorithm(GA)is adopted.By jointly optimizing the transmission and reception of the multiple-input multiple-output radar,and taking into account the main lobe width,the peak sidelobe level can be reduced as much as possible while the main lobe width remains in a suitable range.Then this paper introduces some classic super-resolution angle measurement algorithms,and a super-resolution algorithm has proposed,which is based on conjugate weighted subspace.It can be proved from simulation that this algorithm can obtain the angular resolution equivalent to the sparse recovery / compressed sensing DOA estimation algorithm under the condition of low signal-to-noise ratio,few snapshots or even single snapshots,but the calculation amount is greatly reduced.Based on the optimization of the MIMO radar array,some super-resolution algorithms are used.And it is verified by simulation that in the case of non-ideal factors in the project,this can meet the requirements of angular resolution.