Research On Target Recognition Based On Radar High Resolution Range Profile

Radar high resolution range profile(HRRP)is the coherent summations of the complex returns from target scatterers onto the radar line-of-sight when the bandwidth of radar transmitted wave is wide.It contains the target structure signatures,which is very important for target recognition and classification.HRRP is the most promising radar target recognition technology in the field of engineering because of its advantage of easier acquisition and timeliness.In the process of HRRP recognition,target-aspect sensitivity,high dimensionality of HRRP and poor recognition stability of single classifier in low signal-to-noise condition are the important factors affecting the accuracy of recognition.In order to solve these problems and improve the accuracy and robustness of HRRP recognition,adaptive framing,low dimensional feature extraction and multi-classifier decision fusion are proposed in this paper.The main work of this paper is as follows:(1)In this paper,the causes of the sensitivity problem in HRRP recognition are studied.By analyzing the statistical characteristic of HRRP's frequency spectrum amplitudes,we propose a new adaptive framing method based on the factor analysis model which is combined the factor analysis model with Jensen-Shannon(JS)Divergence.This method can adaptively adjust the framing interval according to the change rate of the target attitude angle.The framing and recognition rate results based on the measured aircraft data prove that the proposed method can effectively reduce the impact of target-aspect sensitivity and provide a basis for subsequent recognition.(2)Aiming at solving the problem of recognition efficiency because of high dimensionality and redundant information of HRRP,multi-scale sparsity preserving projection and adaptive maximum margin sparsity preserving projection are proposed.Multi-scale sparsity preserving projection analyzes the signal characteristics from the perspective of multi-scale theory,which can effectively mine the inherent sparse structure information in HRRP multi-scale space and enrich the information of low-dimensional features.Adaptive maximum margin sparsity preserving projection incorporates the adaptive maximum margin criterion into the dimensionality reduction constraint.It can effectively fuse HRRP sparse structure information and sample label information and explore the signal intrinsic characteristics in the original scale space.The experimental results based on the measured data show that features extracted by the proposed methods have low dimensionality,good recognition performance and strong robustness.(3)Aiming at the low recognition accuracy and poor stability of a single classifier in complex environments such as low SNR,the HRRP multi-classifier decision fusion recognition is studied and an adaptive-classification-weight decision fusion method is proposed in this paper.Through the adaptive transformation of nearest neighbors,selection and weight between classifiers,we achieve an efficient fusion of multi-classifier recognition results.Experiments based on measured aircraft data prove the superiority of this method.