Research On Long-time Coherent Integration Method For Complex Moving Target Detection

Radar transmits the electromagnetic wave to realize the target detection,parameter estimation,identification,relocation,etc.,which can provide the advantages of all-time,all-weather,wide coverage,long detection range,and is free of environment restriction and can effectively remedy the defects of information sensing capability of moving targets and surrounding scene compared with the optics systems including the visible light and the infrared ray.As one of the important tasks,ground moving target indication(GMTI)plays a very important role in airborne/spaceborne warning surveillance system.Combine GMTI with synthetic aperture radar(SAR)technology,we can not only obtain the high-resolution ground scene image,but also realize the moving target detection,which has the practical value in civil and military applications.However,in a practical SAR-GMTI system,if we apply the matched filtering function related to the static scene parameters to image the ground moving targets,they will be severely smeared along range and azimuth directions due to the target motions,especially for the ground maneuvering targets,which are not beneficial for the subsequent target relocation,identification,classification,etc.Additionally,with the development of stealth technology,many highly maneuvering weak air targets make the traditional radar suffer from the severe challenges.These targets can be characterized as “long range,low observation,and highly maneuvering”,and thus these moving targets usually have low SCNRs.Increasing the observation time can effectively increase the output SCNR and thus improve the target detection ability.However,during in a long coherent integration time,the complex motions of a maneuvering target,e.g.,high velocity,acceleration and acceleration rate,involve linear range walk,quadratic range curvature,cubic range curvature,and high-order Doppler frequency migration within a coherent integration interval,which may make it difficult to detect due to the complex coupling effects between range and azimuth.To deal with the energy dispersal issues of the complex ground and air moving targets,in this dissertation,we accomplish the target motion compensation and parameter estimation based on polynomial transform,time-frequency transform,and Keystone transform combined with matched filtering processing.The main works can be summarized as follows:1?To image a ground uniformly moving target in a SAR-GTIM system,we propose a fast SAR imaging method for moving target without target motion parameter estimation.After performing the range migration compensation by using Keystone transform and matched filtering function related to the platform parameters,we model the target echoes as an LFM signal with the Doppler center and chirp rate related to target cross-and along-tack velocities.Then,according to the equal interval sampling of the azimuth slow time,we propose the time reversal transform(TRT)to eliminate the Doppler frequency migration.Finally,the signal energy will be well accumulated in the range-Doppler domain,and thus the moving targets can be efficiently recognized in the focused image.The major advantage of the proposed method is that it can obtain well-focused images of all targets in one processing step without target motion parameter estimation,and thus it is computationally efficient and suitable for real-time target imaging.Additionally,compared with the conventional SAR imaging methods which are also free of search,the proposed method can precisely compensate the Doppler frequency migration caused by target along-track velocity,and thus can obtain a better target imaging quality.Both simulated and real data processing results are used to validate the effectiveness of the proposed method.2?To fast estimate the motion parameters of an uniformly moving target in a SAR-GMTI system,we propose a new time-frequency analysis method,i.e.,second-order WVD(So WVD).After range migration compensation,we still model the received signal as an LFM signal.Then,we apply the So WVD to transform this LFM signal into time and delay-time plane and realize the parameter estimation based on 2-D FFT with respect to time and delay-time variables.According to the estimated motion parameters,a well-focused target image can be obtained.Compared with the conventional Hough-WVD method,the proposed So WVD can directly eliminate the linearly coupling effects between time and delay-time variables,and thus has low computational complexity due to the FFT operations without grid search.Both simulated and real data processing results indicate that the proposed So WVD is an effective motion parameter estimation method.3 ? As for a ground maneuvering target in a SAR-GMTI system,we propose the second-order Keystone transform and generalized Hough-HAF(GHHAF)method to accomplish the motion parameter estimation.Firstly,we build the third-order motion model between a ground maneuvering target and the SAR platform,and then we analysis the impacts of target long-track velocity,acceleration,and cross-track velocity,acceleration on target imaging quality.After range migration compensation based on second-order Keystone transform and Hough transform,we apply the GHHAF to accomplish the parameter estimation.Finally,a well-focused result of a maneuvering target is obtained according to the estimated parameters.Compared with the HAF,the proposed GHHAF can obtain a higher integration gain,parameter estimation performance,and cross-term rejection performance.Both simulated and real data processing results are used to validate the effectiveness of the proposed method.4?Considering that the previously proposed GHHAF is still computationally burdensome due to the 2-D joint searching in Hough domain,we propose a modified GHAF transform,i.e.,scaled GHAF,to estimate the motion parameters of a ground maneuvering target without parameter searching.After transforming a QFM signal into GHAF domain,we apply the generalized Keystone transform to eliminate the coupling effect between time and delay-time variables,and then the parameter estimation is realized according to the peak position in the 2-D frequency domain with respect to time and delay-time variables.Compared with GHHAF,the proposed SGHAF can obtain a higher integration gain,a better parameter estimation precision,a better cross-term rejection performance,and a lower computational complexity.Both simulated and real data processing results show that the proposed SGHAF can be seen as a natural,typical,and intuitionistic time-frequency representation of a QFM signal.5?As for a highly maneuvering weak air target during a long observation time,the target complex motions,e.g.,high velocity,acceleration and acceleration rate,involve linear range walk,quadratic range curvature,cubic range curvature,high-order Doppler frequency migration,and Doppler ambiguity,which may make it difficult to detect due to the energy dispersal along range and azimuth directions.To deal with these problems,we propose a long-time coherent integration method based on Keystone transform and matched filtering function,which can effectively deal with the complex coupling effects between range and azimuth caused by target velocity,acceleration,and acceleration rate.Compared with the conventional long-time coherent integration methods,the proposed method can realize a longer integration time,a higher parameter estimation precision,and a better target detection performance.Simulated processing results show that the proposed method can effectively improve the detection performance of a highly maneuvering weak air target.