基于压缩感知的认知雷达参数估计

Parameter Estimation for Cognitive Radar Based on Compressed Sensing

为了改善基于压缩感知技术的认知雷达(CR)目标参数估计的性能,研究了CR闭环反馈中的感知矩阵更新和稀疏目标场景重构。首先,根据等角紧框架和复Gram矩阵,构造胖矩阵情况下感知矩阵更新的目标矩阵;然后,将测量矩阵优化设计的最小二乘问题展开,利用Kronecher积推导测量矩阵更新的一步迭代公式;最后,引入阈值收缩函数去除迭代重加权最小二乘估计值中的无关小量,进而去除重构场景中的伪峰。计算机仿真实验验证了该算法的有效性。

To enhance the performance of the parameter estimation for cognitive radar （CR） based on compressed sensing,the update of the sensing matrix and the recovery of the sparse target scene are studied in this paper.When the sensing matrix is broad,a novel target matrix is built based on the concatenation of the equiangular tight frame and the complex Gram matrix.Then,the iterative formulation for the measurement matrix update is deduced based on the Kronecher product through the optimization of the corresponding least square problem.Lastly,the threshold-shrinkage function is utilized to eliminate fake peaks in recovered target scene by eliminate the irrelevant epsilon in iteratively re-weighted least square estimation.The effectiveness is demonstrated by computer simulations.