Block Adaptive Subspace Pursuit Method for Wall Clutter Mitigation

Abstract

Through-the-wall radar can detect and locate targets behind obstacles, which has been widely applied in various civil and military applications. However, wall reflections are usually stronger than those of the targets, making it very challenging to extract the target information. To tackle the problem, this paper proposes an adaptive wall clutter mitigation method. First, the discrete prolate spheroidal sequence basis is used to model wall clutter due to its superiority in modeling the spatially extended property of the walls. Then, by incorporating the compressed sensing technique, the subspace pursuit algorithm is combined with block and adaptive iteration to estimate wall clutter. In an adaptive manner, the proposed algorithm greatly eases the requirement on priors and thus improves the applicability in practice. Practical experiments commendably validate the effectiveness and superiority of the proposed method.