Capacity analysis of interlaced clustering in a distributed antenna system

Low signal strength and high interference lead to significantly reduced data-rates at the cell edge. With rising demand for spectrum and systems moving closer to universal frequency reuse, the problem has become more pronounced. Techniques like Fractional Frequency Reuse boost the edgeuser performance at the cost of the spectral efficiency and are therefore sub-optimal. In this paper we investigate interlaced clustering as a solution to the edge user problem for a general distributed cellular transmission system, and explore a multicell Distributed Antenna System as a particular example. In interlaced clustering, different parts of the spectrum use coverage patterns that are spatially shifted (by less than a cell size) replicas of each other. An information theoretic analysis is presented to characterize the proportional fairness boundary point of the achievable rate region. In the process, the joint resource allocation problem is formulated and shown to be convex. As opposed to using interior point methods which are relatively slower, the current paper proposes a novel gradientsearch algorithm to solve the resource allocation problem. It is demonstrated that fractional frequency reuse can in fact be represented as a special (albeit sub-optimal) case of interlaced clustering. Simulation results show that interlaced clustering can boost the edge-user rates by a factor of 2 with negligible degradation of rates in the cell interior. Results also show that interlaced clustering outperforms the edge-user rates achieved with fractional frequency reuse by a factor of 1∶5. The theoretical results are validated by comparing performance to a practical proportional fairness scheduler.