Research On Direct Position Determination Method Based On Sparse Representation

Passive location technology has always been an important research direction in the field of electronic reconnaissance.Compared with the traditional two-step location method,the direct position determination method does not estimate the location parameters,but directly processes the original sampling signal to obtain the location estimation of the emitter,which reduces the loss of information and has higher positioning accuracy and stronger robustness,so it has attracted more and more attention.In recent years,sparse reconstruction theory has been introduced into emitter location algorithm,which provides a new perspective for emitter location.Aiming at the problem of emitter location with partially unknown parameters of signal propagation model and off gird model,this thesis studies the emitter direct position determination method based on sparse representation.The main work is summarized as follows:(1)On the basis of constructing the direct position determination model,the single emitter direct position determination algorithm with known and unknown signal forms,and the multi emitter direct position determination algorithm based on multiple signal classification(MUSIC)and minimum variance distortionless response(MVDR)are stud-ied respectively.The cramero lower bound of the emitter direct position determination accuracy is analyzed,and the effectiveness of the emitter direct position determination method is verified by numerical simulation.(2)In the emitter location based on sparse representation,the of signal reconstruc-tion depends on the accurate sparse dictionary.When the parameters of the propagation model cannot be fully known,the constructed sparse dictionary is different from the real dictionary,which will lead to the decline of location performance.To solve this problem,a multi emitter direct position determination method based on multi dictionary joint and hierarchical block sparse Bayesian framework is proposed.The emitter location problem is transformed into recovering the signal with shared sparsity under multiple dictionar-ies,and the basis mismatch caused by channel attenuation is solved by multi dictionary combination.Simulation results show that the proposed method has better positioning per-formance than sparse Bayesian method and traditional direct positioning method under the conditions of low signal-to-noise ratio and less snapshots.(3)Grid mismatch and partial unknown parameters of propagation model will lead to the inaccuracy of the constructed sparse dictionary,which will lead to the mismatch of base matrix and the degradation of reconstruction performance.To solve this problem,a sparse Bayesian emitter direct position determination method based on parametric dic-tionary learning is proposed.By unifying the off-grid model and the unknown dictionary as the error between the constructed dictionary and the real dictionary,the dictionary pa-rameters are updated and learned,so that the constructed dictionary gradually approaches the real dictionary,so as to realize the direct position determination of the off-grid emitter under channel attenuation.Simulation results show that the proposed method can effec-tively solve the problem of off-grid emitter location,and has higher location accuracy than sparse Bayesian method and orthogonal matching tracking algorithm.