Research On Radar Signal Detection Method Based On Compressed Sensing Theory

Signal detection is an important issue in radar system.In military or civilian fields,signal detection is applied widely,such as radar detection,ship navigation safety,so signal detection has attracted the attention of many scholars.In order to obtain better anti-jamming and higher resolution,the radar system often uses the large time bandwidth signal as the transmitting signal.However,in the traditional Nyquist sampling framework,the large bandwidth will inevitably bring about the problem of large data acquisition,transmission,storage and processing.The appearance of compressed sensing theory is an effective tool to solve this problem.Polynomial phase signal(PPS)is a kind of commonly used radar wideband signal,so this thesis will study the problem of PPS detection.Combined with compressed sensing theory,this thesis will also study and implement the PPS signal detection algorithm.First of all,the approach of this thesis studies the sparse representation of PPS signals,and then constructs different sparse dictionaries for different-order PPS signals.For the second-order PPS signal,LFM signal waveform,references constructed delay dictionary and FRFT orthogonal basis for the dictionary.And through experimental verification,the LFM signal can be expressed by these two dictionaries sparsely and the FRFT Dictionary performs better on anti interference of white noise.The third-order PPS signal is is constructed by the waveform matching dictionary.Secondly,the design and implementation of the compression detection algorithm is based on the establishment of the compression sensing detection model and the detection model are studied.Firstly,a detection model under Gaussian white noise channel is established.The existing detection methods is based on the sparse coefficient position,but its detection effect is bad under low SNR condition,and it can only detect known signal.The normalized residual is then introduced into the detection algorithm.According to the different applications of LFM signal and third-order PPS signal in radar system,a normalized residual detection algorithm is proposed for LFM signal,and the effectiveness of the algorithm is verified.Then,a multi-pulse detection algorithm is introduced into the LFM signal detection.A multiple detection algorithm and an accumulation detection algorithm are proposed.The simulation results showed that the two multi-pulse detection algorithms are different from the single-pulse normalized residual detection algorithm These two multi-pulse detection both improved the detection performance,and accumulation of detection algorithms performs better.Aiming at the third-order PPS signal,the detection algorithm of normalized residual based on multi-component model is studied emphatically.Based on the characteristics of normalized residual slope,an algorithm is proposed to estimate the number of sources.Finally,the algorithm of this thesis has verified the validity of the estimation of the number of information sources.