Research On Sparse Reconstruction And Recognition Of Undersampling Radar Signal

As the forerunner and premise of modern electronic warfare,electronic reconnaissance provides intelligence basis for protecting our equipment and conducting electronic attacks on the enemy by intercepting enemy signals and extracting information,and thus plays an increasingly prominent role in modern warfare.In the actual reconnaissance receiving of radar signals,it is often faced with the receiving of large bandwidth signals.The traditional signal receiving based on Nyquist sampling law poses great challenges to the performance of the reconnaissance receiver and the subsequent data transmission and storage.The theory of Compressed Sensing(CS)provides a theoretical basis for undersampling of signals at a low speed.Therefore,it is of great significance to study receiving radar signal in the way of undersampling,reducing data rate,and realizing reconstruction and recognition.Focusing on the problem of the interpretation of undersampling radar signals,this thesis constructs an undersampling receiving framework of radar signals based on compressed sensing theory,and focuses on the reconstruction and recognition of undersampling radar signals.Through the improvement of sparse representation and reconstruction method,the reconstruction effect and recognition performance are improved.The main research contents of this thesis include:1.Aiming at the problem that sparse representation of multiple modulation signals cannot be realized in fixed orthogonal basis,a coherence constraint based dictionary learning method is proposed.The sparse representation dictionary of signals is obtained based on signal samples,which improves the problem that the current sparse representation method has high mutual coherence with the dictionary obtained by radar signal,reduces the coherence of representation dictionary,and achieves better sparse representation effect.2.Aiming at the problem that the reconstruction effect of undersampling signal is poor when the estimation of sparsity prior information is not accurate or in the presence of noise,a covariance estimation based sparse reconstruction method is adopted to realize the signal reconstruction without estimating the sparsity of the signal,which improves the reconstruction accuracy and the adaptability of the reconstruction process.3.Aiming at the problem of insufficient discrimination performance and low recognition efficiency of current sparse representation based recognition methods,this thesis proposes a sparse projection dictionary learning based recognition method.By improving the objective function of dictionary learning,the time complexity of the recognition process is reduced on the basis of guaranteeing the recognition accuracy,and the recognition of reconstructed undersampling signals is realized.Finally,the feasibility of the proposed reconstruction method and recognition method is verified through the simulation experiments,which can realize the effective reconstruction and recognition of the undersampling radar signal.