Distributed compressed sensing for the MIMO MAC with correlated sources

In this paper we consider the transmission of jointly sparse signals over a MIMO MAC composed of two transmitters and one receiver. Distributed compression is performed at the transmitters using a liner transformation and joint reconstruction at the receiver is enabled using the theory of distributed compressed sensing. The objective is to minimize the sum MSE between the uncompressed signals at the transmitter and the recovered signals at the receiver. We present a new theoretical framework to study this problem and provide an algorithm enabling to transmit correlated signals over a MIMO MAC, which performs provably close to optimal. To validate our approach we analyze the performance of our system for different noise conditions and varying compression factors. The results show that distributed compressed sensing can be performed reliably over a MIMO MAC.