Synchronization for cascaded distributed MIMO communications

In this paper, we propose a general decentralized synchronization scheme and a common error model for both frequency and time offsets in the cascaded distributed MIMO communications. By cascaded, we mean that Cluster A transmits to Cluster B, and Cluster B transmits to Cluster C, and so on, yet there is never any coordination between nodes within a cluster; for example, there is no signal redistributed within a cluster. The proposed scheme includes two parts: distributed pre-synchronization of the MIMO transmitters and the weighted combination at a receiver of estimates associated with different relays. The analysis shows that, with the power-weighted combination, although both the variance and covariance of pre-synch errors keep increasing from hop to hop, the expectation of sample variance will converge to a relatively small value after several hops. The convergence property of the expected sample variance makes the multi-hop transmission possible. Both simulation and experimental results verify our analysis.