A system for compressive sensing signal reconstruction

An architecture for hardware realization of a system for sparse signal reconstruction is presented in the paper. The threshold based reconstruction method is considered. Noise, appearing in the signal as consequence of the missing signal samples, is used to derive the threshold. The threshold is then used in order to separate signal and non-signal components, and it is dependent on the number of missing samples. The algorithm is modified in this paper with an aim to reduce the system complexity, and to provide easier hardware realization. Instead of using the partial random Fourier transform matrix, the minimization problem is reformulated using only the triangular R matrix from the QR decomposition. The triangular R matrix can be efficiently implemented in hardware without calculating the orthogonal Q matrix. A flexible and scalable realization of matrix R is proposed, such that the size of R changes with the number of available samples and sparsity level.