Research On Fast Estimation Algorithms Of Ship HRRPs In High-resolution Maritime Radars

Detection,recognition and classification of large targets on the sea surface such as ships and warships is always one of the important tasks of high-resolution maritime radars.Due to limited peak power of radar transmitters,high-resolution maritime radars often transmit pulses of large time-bandwidth product such as wide linear/nonlinear frequency modulated pulses and detect distant targets by the pulse compression technique.In this case,the signalto-clutter ratios of large targets are usually higher than the range side lobe level of the pulse compression.As a result,the range side lobe effect occurs.In other words,the radar returns of large leaked from the range side lobes of the pulse compression is still larger than the decision threshold of target detection after coherent integration,which results in the severe expansion of the large targets.Besides,the range expansions of large targets make that small targets in the range expansions of the large targets are difficult to be detected by traditional adaptive detection methods.It is worthy to note that the large targets such as ships and warships often have sparse high-resolution range profiles(HRRPs).Therefore,sparse recovery becomes an effective method to solve this problem of masked small target detection in high-resolution maritime radars.Firstly,the radar returns of a large target on the sea surface are reconstructed in high precision,then the reconstructed radar returns are removed from the received radar returns,and at last masked small targets can be detected from the residual signals.It is of great significance for practical applications to deeply understand and analyze the high-precision reconstruction methods of the radar returns of large targets from the precision and computational burdens.In this thesis,several sparse recovery algorithms of the HRRPs of large targets are reviewed and analyzed.Their performance is compared by experiments.Based on the experiments results,a refined processing scheme to improve computational efficiency are presented.And this thesis proposes a refined linear programming(LP)algorithm,a HRRP estimation algorithm based on subspace matching pursuit and an accelerated LP algorithm.The main content of the thesis is summarized as follows:In the first chapter,we introduce the research background and significance of the problems to be investigated in this thesis and the organization of this thesis.In the second chapter,we introduce the basic principle of sparse recovery methods of HRRPs of large targets in high-resolution maritime radars.And the existing iterative minimization sparse learning algorithm,linear programming algorithm,iterative minimization sparse recovery algorithms are reviewed.Moreover,the differences of the three sparse recovery algorithms are analyzed and discussed.Finally,through simulation data experiment,the performance of these three algorithms is compared and analyzedIn the third chapter,the concept of HRRPs and the statistical model of sea clutter in highresolution maritime radars are introduced.Secondly,a fast recursive algorithm to estimate ship HRRPs is presented that is based upon the subspace matching pursuit idea.Moreover,a refined LP-based algorithm is proposed to improve the ship HRRP estimation accuracy of the LP-based algorithm.The fourth chapter focuses on how to accelerate the LP-based algorithm.Firstly,the process of restoring ship HRRP by the LP-based algorithm is introduced,and then the factors to affect the efficiency of the LP-based algorithm are analyzed.Then,the principle of the simplex method and its transformation in GPU are reviewed.Finally,a GPU processor-based method to accelerate the LP-based algorithm for HRRP recovery is proposed.In the fifth chapter,we summarize the works in this thesis and discuss the further investigations.