Distributed MIMO Coordination for Next Generation Small-cell Networks

While a primary challenge addressed by earlier wireless standards was handling channelfading, the current challenge for 5G standardization comes from ever-growing numberof smartphone users. This issue has aroused considerable interest in the dense small-lcellnetworks to accommodate more spatially reusable spectrum. Promising theoreticalresults have been shown improved network sum-rate is possible. However, despite itstheoretical success, this system suffers from substantially increased inter-cellinterference.Advanced multiple-input multiple-output (MIMO) antennas techniques show that it is,nonetheless, possible to cancel out the interference, if all small-cells cooperatively designprecoders/combiners by sharing global channel state information (CSI). However, it wasvery quickly revealed that sharing global CSI becomes infeasible as the networkcomplexity increases. Moreover, there are many practical challenges that have to beaddressed before any of the promised gain could be harnessed. At the front of them arequestions about efficient distributed precoder/combiner adaptation and reliable local CSIestimation in the presence of interference. The distributed precoder/combiner adaptationemploys an iterative forward-backward (F-B) directional training, exploiting channelreciprocity. In each direction, the receiver adapts its filter to the local CSI, estimatedusing pilot signals. In particular, high quality precoder/combiner should be extractedbefore the channel coherence time expires. However, as the numbers of small-cells andusers in the network grow, the amount of interference accumulates. Consequently, therequired number of F-B iterations and the pilot overhead in each F-B iteration, shouldincrease, which can potentially overwhelms the channel coherence resource. Thissignificantly deteriorates the achievable sum-rate promised by the dense small-cellsystems because of less time left for data communications.Our goal is to tackle the above issue by providing efficient distributed precoder/combineradaptation techniques for the dense small-cell MIMO networks. The system isconstrained to only allow a fixed number of F-B iterations, while still improving thequality of filters. The novelty of the project mainly lies in introducing low-overhead, butreliable techniques that perform fully distributed rank adaptation and provide robustCSI estimates. It also exploits past pilot signals received over many coherence blocks toreduce the estimation error and provides initial filter alignment strategies that furtherenhance the performance. The adoption of the devised technique to the practical small-cellnetwork will result in an upgraded cellular coordination that greatly scales thenetwork sum-rate. Such a framework could serve as a reference for developing advancedsmall-cell systems that could potentially be evolved into the 5G wireless standards.?