Autonomous Component Carrier Selection for 4G Femtocells - A Fresh Look at an Old Problem

This paper addresses the interference management problem in the context of LTE-Advanced femtocells. Due to the expected large number of user-deployed cells, centralized network planning becomes increasingly less viable. Consequently, we consider an architecture of autonomous decision makers. Our main contribution in this paper, denominated Generalized Autonomous Component Carrier Selection (G-ACCS), is a distributed carrier-based inter-cell interference coordination scheme that represents one step towards cognitive radio networks. The algorithm relies on expected rather than sensed interference levels. This approach facilitates scheduler-independent decisions, however, it can lead to overestimation of the interference coupling among cells when the resources are not fully utilized. Acknowledging this fact, G-ACCS leverages the power domain to circumvent the restrictive nature of expected interference coupling. This work focuses on the downlink and also provides an extensive characterization of the network performance as a function of the topology as well as the often overlooked temporal traits of traffic. We compare G-ACCS with other carrier-based solutions, including the simplest universal reuse strategy. The encouraging simulation results demonstrate that G-ACCS achieves an efficient and fair distribution of resources in all considered traffic and deployment conditions. More importantly, this is attained in a truly autonomous fashion, without any explicit parametrization.