Research On Narrowband Interference Suppression Technology Based On Compressed Sensing

With the rapid development of modern wireless communication technology, space electromagnetic environment has been becoming more and more complex and causing severe electromagnetic interference. Interference and anti-interference become an important confrontation form of electronic warfare and information warfare. How to suppress interference and ensure the validity and reliability of communications in harsh environments has become an important issue in today’s communications technology research. Compressed sensing(CS) theory is a new information acquisition theory. If a signal is sparse or compressible in the transform domain, it can be sampled lower than the Nyquist sampling rate and projected into lower-dimensional space under the effect of a measurement matrix which is not associated with the sparse matrix at the same time, then restored with high probability by reconstruction algorithm.This paper combines anti-interference technology with compressed sensing theory and studies a narrow band interference suppression technology based on compressed sensing. The main research content is reflected in the following aspects:Firstly, this paper studies traditional interference suppression algorithms include time-domain interference suppression, transform-domain interference suppression, code-aided interference suppression. Among them, the frequency domain transform interference suppression technology is introduced in detail, including the choice of the interference threshold and the processing method of the interference.Secondly, This paper studies the three main contents of the compressed sensing framework, namely the sparse representation of the signal, the design of the observation matrix and the reconstruction of the signal. This paper illustrates sparse transformation effect of Fourier transform base, discrete cosine transform base and wavelet transform base. The design principles of observation matrix include the restricted isometry property and uncorrelated property. Several common observation matrix are studied, and their reconstruction performance for sparse signal under different observation compression ratio are tested. Gaussian random matrix has low correlation with most sparse matrix, so it is theoretically optimal observation matrix. The typical algorithms of three kinds of reconstruction algorithms are studied and tested. Orthogonal matching pursuit algorithm is most widely used since it’s hardware implementation is easy to achieve and reconstruction works well.Thirdly, in this paper, a method of interference suppression based on compressed sensing is proposed. Since the compressed sensing theory is able to reconstruct interference, so time domain cancellation can be a way to suppress interference. Due to the presence of leakage phenomenon after narrowband interference frequency domain transform, it need to be suppressed by windowing. Meanwhile, SNR loss caused by window function can be compensated by overlap windowing method. In order to process mixed signal that contains the desired signal, interference and noise with compressed sensing algorithm, interference threshold algorithm is used to determine the degree of sparsity of interference, and its feasibility is verified. The tests of narrowband interference in a BPSK transmission systems proves that interference suppression based on compressed sensing technology can really achieve a fair performance. But it is not so good comparing with the traditional method of frequency domain zero-set method. Through the analysis, it is found that the occasional judging error of interference frequency point’s location caused by the interference suppression algorithm based on compressed sensing is the main reason for the performance difference.Lastly, this paper improves the algorithm and proposes a narrow band interference iterate suppression technology based on compressed sensing. The key of interference suppression technology based on compressed sensing is the accuracy of reconstructed interference, so, the more accurate reconstructed interference is, the better performance transmission system will has. Thus, this paper proposes an iteration manner to achieve a more accurate reconstruction of interference. Single-tone, multi-tone and narrow-band interference are suppressed respectively in BPSK transmission system under the Gaussian and Rayleigh fading channel. The simulation results show that, the accurate estimation of interference and the performance of signal demodulation can achieve a virtuous cycle to suppress the interference and improve the transmission performance.