适用于压缩感知估计角度的测量矩阵研究

发布时间：2018-08-08 17:48

【摘要】：针对接收数据压缩投影后导致到达角(Direction-Of-Arrival,DOA)估计精度不高的问题,提出一种高精度的全局信息压缩投影到达角估计算法。该算法首先提出更适应角度估计的等正弦空间稀疏化范德蒙矩阵作为测量矩阵,然后对由其组成的Gram矩阵的非对角元素进行压缩处理得到目标矩阵,接着利用步长符合沃尔夫条件的梯度下降法优化Gram矩阵,得到当Gram矩阵与目标矩阵最接近时所对应的可以保留更多全局信息的测量矩阵,最后利用此矩阵压缩接收数据,将接收数据投影到测量矩阵空间,进行稀疏重构得到角度估计结果。仿真实验表明,所提算法角度估计精度远优于同等条件下辐射源信号直接重构的角度估计结果,且在信噪比大于-6 dB时数据压缩投影后角度估计的成功率达到100%,性能优越。
[Abstract]:In order to solve the problem that the DOA estimation accuracy of Direction-Of-Arrival-DOA is not high after the compressed projection of data is received, a high-precision global information compression projection DOA estimation algorithm is proposed. In this algorithm, the Vandemont matrix, which is more suitable for angle estimation, is sparse in the equal-sinusoidal space as the measurement matrix, and then the non-diagonal elements of the Gram matrix are compressed to obtain the target matrix. Then the gradient descent method with step size matching the Wolff condition is used to optimize the Gram matrix. When the Gram matrix and the target matrix are closest to each other, the measurement matrix which can retain more global information is obtained. Finally, the received data are compressed by this matrix. The received data are projected into the measurement matrix space and the angle estimation results are obtained by sparse reconstruction. Simulation results show that the accuracy of the proposed algorithm is much better than that of the direct reconstruction of emitter signals under the same conditions, and the successful rate of angle estimation after data compression projection is 100 when the SNR is greater than 6 dB, and the performance is superior.
【作者单位】：解放军信息工程大学信息系统工程学院;
【基金】：国家高技术研究发展计划(863计划)(2012AA01A502,2012AA01A505) 国家自然科学基金(61401513)
【分类号】：TN911.7

【参考文献】