Compressed Sensing Based Azimuth Super-Resolution In Passive Radar And GPU Implementation

Passive radar is a special bistatic and multistatic radar system.The passive radar itself doesn't transmit electromagnetic wave,but detects and tracks target relying on the non-cooperative third party illuminators which have already existed in the environment.Since no electromagnetic waves emitted,this radar can be deployed to densely populated regions such as cities.In addition,passive radar has a strong anti-interference function and can detect stealth targets.Therefore,it has been widely recognized in radar research.However,the low carrier frequency and narrow effective bandwidth of the illuminations lead to a poor azimuth resolution of passive radar.For example,when there are more than two targets in the same distance-Doppler unit of one beam,traditional signal processing methods cannot distinguish the direction of targets.Super-resolution technology is an effective means to improve the azimuth resolution of targets in passive radar.Since the direct wave and multipath interferences in the signals received by passive radar seriously affect the result of super-resolution,the clutter cancellation method is first used to suppress direct wave and multi-path interference.In addition,passive radar needs long-term signal accumulation in order to improve the target SNR,but the traditional super-resolution technology is easily limited by the detection conditions such as the number of snapshots.Compressive sensing is an effective way to make up for this defect.Based on the background of a practical project,this thesis focuses on the research on the clutter cancellation algorithm of passive radar,the azimuth super-resolution method based on compressive sensing theory,and the GPU implementation.The main contents are as follows:1.The clutter cancellation algorithm and an improved ECA of passive radar are studied.In order to accurately achieve azimuth super-resolution for the target,the clutter cancellation algorithm is used to suppress direct wave and multi-path interferences in the signals received by the array element of passive radar.First,three kinds of clutter cancellation algorithms are studied,and the performance of the algorithms is analyzed.Then,aiming at the shortcomings of the traditional ECA algorithm,the corresponding optimization and improvement method is proposed.The improved ECA algorithm is simulated and analyzed by the collected audio signals of analog television,and the effectiveness of the improved algorithm is demonstrated.2.The azimuth super-resolution technology of passive radar based on compressive sensing is studied.The key technology of azimuth super-resolution based on compressive sensing is signal reconstruction.Firstly,the signal model of azimuth super-resolution is established.Then the sparse reconstruction model of azimuth super-resolution is established by compressive sensing theory,and the sparse reconstruction algorithm is determined.Finally,the performance of azimuth super-resolution based on compressive sensing is analyzed through experimental simulation.3.The GPU implementation technology for the azimuth super-resolution of passive radar is studied.Firstly,the improved ECA clutter cancellation algorithm is implemented by GPU.And the performance of signal processing by CPU and GPU is compared through large-scale matrix multiplication and matrix inversion module.Then the orthogonal match tracking algorithm(OMP)based on greedy iterative idea is implemented by GPU.The performance of the GPU real-time processing is analyzed by the OMP signal reconstruction algorithm,and the performance of GPU real-time signal processing is verified.