Cognitive femtocells: Breaking the spatial reuse barrier of cellular systems

We consider smart “cognitive” femtocells, that can decode the base station control channel and make decision on their transmission opportunities based on the scheduling and resource allocation of the overlaying macrocell. The operation of such femtocells is somehow reminiscent of “Type II” relays, as specified in the IEEE 802.16j standard. Under certain system assumptions, we consider the tradeoff region of the macrocell versus aggregate femtocell throughput. For the basic system architecture, we evaluate the impact of open versus closed access and interference cancellation at the femtocells. As possible improvements, we consider an application of the “downlink interference alignment” idea to our system, and the use of multiple antennas at the femtocell access points. We show that a careful combination of these techniques achieves a remarkably large aggregate femtocell throughput, at very low penalty in terms of the precious and scarce macrocell throughput. The unprecedented performance achievable by the proposed architecture advocates a system design that places femtocells at the core of future wireless networks, rather than treating them as an add-on afterthought.