Measurement Matrix Design For Compressive Sensing Radar

The compressive sensing(CS)theory exploits the sparsity of signal and samples signal in a specific transform domain.The number of targets is always far less than the number of resolution cells in the target scene of radar system,that is,the target scene is sparse.Radar researchers proposed the concept of CS radar,detecting the link between the CS theory and the sparsity of radar system.CS radar provides new methods to improve the performance of radar system,reducing the amount of data to be transmitted and processed.Nevertheless,there are still some difficult problem for further development of CS radar.One of these problems is that implementation of measurement matrix is complex.Meanwhile,since the performance of CS theory decreases rapidly in the presence of strong noise and interference,it is another important problem to be solved for CS radar.At last,when CS theory is applied to multiple-in-multiple-out(MIMO)radar system,the sparse representation of multiple channel signal and the high computation complexity of recovery of big range target scene is also a problem to be solved.Hence,this paper focus on the aforementioned problems and study the measurement matrix based methods as follows: 1.Sub Nyquist sampling scheme based on white noise random digital filters in CS radar.Since the existed analog-to-information convertors(AICs)always have high rate pseudo random sequence which has high complexity of hardware implementation,a newly sub-Nyquist sampling scheme based on digital filters is proposed.Firstly,a random frequency hopping signal is analyzed and used as the transmitted signal of CS radar.Secondly,a sub-Nyquist sampling scheme for Cs radar is proposed and an experiment simulation platform is also used to verify the feasibility of the proposed sampling scheme.Finally,a white noise random filter is proposed for measurement matrix design.Compared with the existed AIC schemes,the proposed method reduce the complexity of hardware implementation by avoiding the high rate random sequence.Employing the white noise random filter,the coherence of measurement matrix is reduced and the performance of CS radar in the presence of strong noise is improved.2.Pulse accumulation measurement matrix for CS radarA measurement matrix based pulse accumulation method for CS radar is proposed.Firstly,in order to improve the detection performance with low level signal-to-noise ratio(SNR),a grouping pulse accumulation method which elimates the range migration of targets is proposed and the corresponding measurement matrix is designed.Secondly,the pulse accumulation method is expanded to a two dimensional measurement method to further reducing the amount of data.Finally,a mixed pulse accumulation measurement matrix is proposed for compensating the range migration.Compared with the exsited multiple pluses processing methods,the proposed pulse accumulation method significantly improve the performance of CS radar in the presence of strong noise.3.Measurement matrix based aperture completion method and spatial filter measurement matrix for CS-MIMO radar.Measurement matrices for improving the performance of CS-MIMO radar are proposed.Firstly,in order to improve the detection performance with limited antennas,an aperture completion method that employing structured measurement matrix for CS-MIMO radar with co-prime receive array is proposed.Then,in order to improve the performance of CS-MIMO radar with strong interference,a spatial filter measurement matrix based on band beamforming is proposed.Employing the proposed measurement matrix,with the rough prior information of targets,the out-section interference can be suppressed and the targets' angle can be accurately estimated.These two proposed methods both use structured measurement matrix to replace the random matrix.Aperture completion and interference suppression are achieved while the received data is compressed,improving the detection performance of CS-MIMO radar.4.Targets' parameters estimation for distributed compressive sensing MIMO radar(DCS-MIMO)and the guaranteed stability of DCS-MIMO.Targets detection method for DCS-MIMO radar and the guaranteed stability are studied.Firstly,the joint sparse modeling and the joint measurement matrix for DCS-MIMO radar is proposed,a joint OMP algorithm is also devised.Secondly,the guaranteed stability based on the average mutal coherence of sensing matrix is proposed to show the relation between the sparsity and the lower limit of compressed data.Finally,an optimized detection method based on secondary information cell recovery is proposed.The detection accuracy of big target scene is increased while the computation time is decreased.These methods provide theory evidence for DCS-MIMO radar system design and signal processing.