基于稀疏贝叶斯学习的稳健空时自适应处理研究

Sparse-recovery space-time adaptive processing (STAP) has considerable potential advantages on moving target detection in nonhomogenious and nonstationary environments for air/space-borne radar systems. The advantages of this technology rely on the accuracy of the clutter sparse model. However, in real systems and environments, array errors, the intrinsic clutter motion and the space-time steering dictionary mismatch lead to the model mismatch compared with nowadays signal model assumptions, which results in the performance degradation of the sparse-recovery STAP. Thus, we first study the intrinsic clutter sparsity and sparse model in non-ideal conditions, including the quantificational definition of the clutter sparsity, the designing of the space-time steering dictionary and the effects of the model mismatch on the clutter sparsity. Then, combing the sparse model and incorporating the sparse Bayesian learning framework, we exploit the structural priori of the space-time power spectrum, jointly estimate the space-time power spectrum and the model mismatch parameters, and build the theory for the robust STAP methods based on sparse Bayesian learning. All in all, this research will enrich the sparse-recovery STAP theory and methods, reduce the distance between the sparse-recovery STAP and the real application, and promote the realization of the sparse-recovery STAP.

稀疏恢复空时自适应处理（STAP）技术在非均匀、非平稳杂波环境机/星载雷达运动目标检测方面具有极大的潜力优势。该技术的优越性能有赖于准确的杂波稀疏模型，然而，实际系统和环境中存在的阵列误差、杂波起伏、空时导向词典失配等非理想情况使得现采用的模型失配，导致稀疏恢复STAP性能下降。为此，本项目首先研究非理想情况下的杂波稀疏机理与模型，包括：杂波稀疏性度量、空时导向词典设计以及模型失配因素对杂波稀疏性影响分析。然后结合杂波稀疏模型，在稀疏贝叶斯学习框架下，挖掘和利用杂波空时功率谱结构先验，联合实现失配参数估计与信号稀疏恢复，形成基于稀疏贝叶斯学习的稳健空时自适应处理理论和方法。研究成果将丰富稀疏恢复STAP理论与方法，缩小稀疏恢复STAP与实际应用的差距，推动稀疏恢复STAP走向实用化。