Target localization via compressed sensing based on SVD

To ensure localization accuracy, a novel target localization algorithm via compressed sensing based on singular value decomposition(SVD) was proposed to make the measurement matrix satisfy the isometry property. By using gridding method for sensing area, the new algorithm converts target localization to compressive sensing issue, the measurement matrix obtained can effectively satisfy the restricted isometry property, and the preprocessing does not change the sparsity of the original signal, which effectively ensures the reconstruction performance and improves the localization accuracy. The experimental results show that compared with the localization algorithm of sparse targets based on Orth, the new target localization algorithm via compressed sensing based on SVD which is insensitive to noise has a much better performance and lower computation complexity.