Interferer Identification in HetNets using Compressive Sensing Framework

We consider heterogeneous cellular networks (HetNet) where each base station (BS) sends unique training signal based on its physical layer cell identity. Received signal at mobile terminal (MT) is superposition of training signals from different BS. Neither BS identities nor their channel response is known apriori at MT. For this scenario, we consider the problem of finding constituent BS identities from superimposed components in received signal. Though number of BS with unique identities can be quite large in a HetNet, in any given scenario, actual number of BS interfering at MT is relatively few. By exploiting this sparseness, we show that our problem can be solved using block sparse signal reconstruction algorithms under compressive sensing framework where sensing matrix is block-matrix with circulant-blocks (BCB). We apply convex programming based ell_2 / ell_1 mixed norm minimization and greed based subspace matching pursuit approaches to recover interfering BS identities. We characterize block restricted isometry property, mutual subspace incoherence of BCB matrices with i.i.d. Rademacher distributed entries and establish certain recovery guarantees. Our proposed approaches give significant improvements over conventional successive interference cancellation approach over both randomly generated and 3GPP-LTE training signals.