A novel subspace pursuit of residual correlation step algorithm for distributed compressed sensing

Abstract

Multi-signal joint reconstruction is critical in distributed compressed sensing (DCS). We propose a joint reconstruction algorithm for subspace pursuit of residual correlation steps to balance reconstruction efficiency and data quality. The algorithm improves on two aspects. Firstly, the residuals are used as the step feedback condition to achieve dynamic step adjustment, reducing the iteration’s overall number. Secondly, based on the mixed-support set model (MSM), the residual non-decreasing surplus condition is set to reconstruct the common and the innovation components of the target signal in groups, which reduces the mixing error in joint reconstruction. This paper compares the performance of six algorithms of the same type under the conditions of Gaussian sparse signal and measured shock wave field sensor network data. The results show that the proposed algorithm can effectively reduce the number of measurements required for reconstruction and improve efficiency while maintaining accuracy.