Research On Method Of Improving Distance Resolution Based On Band Fusion

For traditional radar systems,the range resolution of the radar system is limited by the sweep frequency bandwidth of the transmitted signal.In order to obtain a high-resolution range image,a larger sweep frequency bandwidth needs to be sent.It is often accompanied by an increase in hardware cost and hardware complexity,and there are many technical bottlenecks.Radar multi-band fusion technology,as an emerging research direction,aims to fuse multiple sub-bands of the frequency spectrum into large-bandwidth signals by means of signal processing,thereby reducing hardware costs without losing radar image performance.Radar multi-band synthesis mainly includes two parts: band registration and band fusion.Among them,the reception echo of different radars shows coherent mismatch due to the differences in the start time,spatial position and transmission frequency of multiple radars.Coherent registration Quality directly determines the frequency band fusion performance.Radar target modeling is the basis for analyzing band fusion.In the second chapter of this article,starting from the signal model,we introduce the moving target modeling when the transmitted signal is a chirp signal and a step frequency signal.Because radar targets are mainly moving targets,before multi-band coherent registration,motion compensation must be performed on the received echoes to eliminate non-coherent elements introduced by target motion.Multi-radar coherent registration is the first step to achieve band fusion.Combined with the actual physical background,the causes of multi-band coherent mismatch and the effect of coherent mismatch target range profile are introduced.An adaptive iterative algorithm based on the combination of genetic algorithm and information entropy criterion is proposed to estimate the linear phase.The disadvantages of traditional cost function in estimating fixed phase are pointed out.A fixed phase estimation function based on pole information is proposed.The computer simulation results show that the algorithm in this paper is less affected by noise and the estimates are accurate.Band fusion is an important step to achieve multi-band synthesis.The distribution of multi-band echoes in the frequency spectrum is divided into three scenarios: band overlap,band tangent,and band separation.Among them,band overlap will cause radars to interfere with each other.Multi-band synthesis.For the scenario where the frequency bands are separated,since there are vacant frequency bands between different sub-bands,the blank frequency bands need to be estimated from the frequency band data after coherent registration.The fourth chapter of this paper first introduces the basic principle and algorithm flow of frequency band fusion,including the establishment of a full-band pole model,model parameter estimation,frequency band estimation,and multi-band splicing.Second,it introduces the use of MDL and AIC criteria to estimate the model order.The basic principle of the method is combined with high-resolution spectral estimation algorithm to estimate the model pole information;the linear least square method is used to estimate the amplitude information;finally,based on the full-band pole model,the estimation method of blank frequency bands and the implementation steps of band splicing are discussed.The effects of experimental parameter settings on the fusion performance are analyzed in detail through computer simulation results,and the reason for the degradation of the frequency band synthesis quality in a low signal-to-noise ratio scenario is introduced.Simulation results show that the distance resolution after band fusion is significantly higher than that of a single sub-band,and the effectiveness of the algorithm is verified.