Research On Robust Beamforming Of Low Sampling Number

Beam forming is a key technology in the array signal processing field, which is widely used in national or military field such as sonar, radar, communications, acoustics, astronomy, seismic, medical imaging and so on. The technology achieved good results after several decades of development, and gradually moving into practical use. However, these traditional algorithms'beam occurred distortion, and the output SINR reduced in low sampling number, which is need for a large number of snapshots to obtain a good space filter effects. But this will result in high computational complexity, hardware implementation difficult, poor in real-time performance of communication and so on. In this paper, we will research on robust beam forming of low sampling number to solve this problem, and the main work includes:First, the principle of beam-forming technology is introduced briefly, and the classical algorithms'simulations are given, then analyzed the problems of classical beam-forming algorithms on low sampling number.Second, the particle swarm optimization theory is introduced, and the problem of running into local optimization is analyzed, then track the particles do simulation proving to have a good effect.Third, in this paper, a new method based on improved particle swarm algorithm for beam forming is presented. This method used an group intelligent iterative approximation algorithm for getting the optimal weights, to avoid the process of using sampling data estimate the signal covariance matrix, so it has a high robustness, low computational complexity, good real-time performance of communication in low sampling number. And the computer simulation proved that it can achieve a good spatial filter purpose and output SINR effect.Fourthly, compressed sensing theory is analyzed. The sampling method under this theory is sampling and appropriate compressing for the signal data at the same time. Therefore, a small amount of signal sampling value can contain all of the information, and can restore accurately by reconstruction algorithm. In this paper, for sparse or compressible signals, a new method based on the theory of compressed sensing for beam-forming is proposed. This algorithm applied the sampling methods of compressed sensing theory, according to the small amount of sampling values to estimate the covariance matrix, then obtain the optimal weight vector. It not only enhances robust performance and reduce computational complexity, but also avoid the time consumption which iterative algorithm brought. The computer simulation proved the algorithm is feasibility.