The current cellular industry is undergoing a huge paradigm shift from an old homogeneous one-tier network structure to a new heterogeneous two-tier structure with joint deployment of traditional macrocell base stations along with a relatively new small cell base stations, widely known as femtocells. Femtocells are low-powered, low-cost, user-deployed base stations meant to improve poor network coverage and, thereby, increase overall system capacity. As more and more femtocells are deployed, their spectrum usage and resulting interference become non-negligible. While using different operating frequency for femtocells is indeed possible, a co-channel deploy- ment of these will increase spectral efficiency, a much sought design by cellular opera- tors. In this thesis, a femtocell-based scheme is considered as a prospective means to enhance the performance of the current cellular infrastructure. In the adopted frame- work, the femtocell access point is tasked with connecting local femtocell users to the network operator without creating undue interference to cellular users. As such, the femtocell is required to cease communication when a nearby cellular user is present to prevent interference. In the envisioned paradigm, an access point possesses little information about the parent cellular base station. For instance, it may not know the individual channel gains, user locations or frequency allocations. To achieve this goal, femtocell users collectively act as sensing devices and are used to acquire data about local signal strength. This work shows that, despite having little knowledge of the operation of the macro environment, a femtocell can take advantage of the data provided by the acquisition devices and agility of the re-configurable antenna to gain insight about proximate cellular devices. The proposed inference scheme leads to a significant performance gain over oblivious femtocells. Experimental results are provided to support this study and its conclusions.
