Frequency Agile Radar Using Atomic Norm Soft Thresholding with Modulations

Abstract

Frequency agile radar (FAR) is known to have excellent performance in electronic counter-countermeasures (ECCM). Recently, compressed sensing (CS) algorithms have been applied in the joint estimation of velocity and range for FAR. Grid based CS algorithms suffer from the off-the-grid problem. Atomic norm minimization algorithms solve the off-the-grid problem by estimating parameters in the continuous domain, however, are not compatible with the cross term in the FAR signal model. To handle the cross term, we divide the value space of velocity into grids, and model the phase term related to velocity and the cross term as modulations. Based on atomic norm soft thresholding with modulations, range and velocity are estimated in the continuous and discrete domain, respectively. The proposed algorithm is compatible with the cross term and alleviates the influence of mismatch. The performance of the novel approach is verified by numerical results.