Smart noise convolution jamming suppression via subspace calibration and alternating inversion
Yuhang Gao, Zihao Liu, Huayu Fan, Shaoqiang Chang, Quanhua Liu, Lixiang Ren · Defence Technology · 2026
Smart noise convolution jamming (SNCJ) is an advanced active jamming technique enabled by digital radio frequency memory. By convolving an intercepted radar transmitted waveform with noise, SNCJ can simultaneously generate suppressive and deceptive jamming effects at the radar receiver, thereby severely degrading the target detection performance. The SNCJ suppression algorithm based on subspace calibration and alternating inversion is proposed for radar systems utilizing the interpulse code agile waveform. First, the discrete SNCJ signal model is established from the continuous-time formulation, and their equivalence in the sampled-observation sense is derived. Next, by exploiting the pulse-to-pulse spectral diversity introduced by code agility, a joint coarse estimate of the jamming delay and the intercepted waveform by the jammer is obtained to initialize the jamming subspace. Then, to address subspace mismatch induced by the delay estimation error, the delay error estimation is formulated as a projection-residual minimization problem. An optimization procedure based on gradient descent is adopted to refine the delay estimate and calibrate the jamming subspace, thereby reducing residual jamming leakage caused by grid mismatch. Finally, based on the calibrated subspace, an alternating inversion scheme is developed. It iteratively updates the jamming component via subspace projection and reconstructs the target component using orthogonal matching pursuit, thereby enabling separation and reconstruction of the jamming and target components. Simulation results demonstrate that, compared with the conventional fractional filtering algorithm, the proposed algorithm achieves more robust jamming suppression and superior target reconstruction performance under strong jamming conditions.