Wyner-Ziv coding for distributed compressive sensing

Distributed compressive sensing (CS) is emerging as a powerful technique of distributed signal processing in various applications such as sensor networks due to its capability of simultaneous sensing and compression. However, since distributed CS is an analog technique, a fundamental open question is to find the best source coding scheme for the distributed CS samples. This paper applies nested-lattice Wyner-Ziv coding to the CS data by exploiting the correlation among the CS samples at different sensors. The proposed coder consists of CS with Toeplitz/circulant sensing matrices and practical Wyner-Ziv coding. Simulation results shows this is a fast, energy-saving system and recovers good quality image sources with low distortion and high SNR.