Group-Wise Listen-Before-Talk Protocol for Dynamic Spectrum Sharing: A New Framework for Full Frequency Reuse

Abstract

In existing dynamic spectrum sharing (DSS) systems, each node arbitrates channel access independently based on carrier-sensing mechanisms such as the listen-before-talk (LBT) protocol. Owing to the uncoordinated channel access between all nodes, the channel is occupied in a random pattern. This makes it difficult to reduce the mismatch in channel quality indicator (CQI) while increasing the spatial reuse gain between the different nodes; therefore, the areal capacity gain obtained by adding nodes is much lower in existing DSS systems than in the cellular system. In this paper, we propose a different means of improving the areal capacity for downlink DSS systems. It is a group-wise DSS approach that ensures full frequency reuse in each group of base stations (BSs) by performing LBT only between the representative BSs, each of them selected by each group. Once a channel is secured by each group, all its member BSs share a channel simultaneously. This approach makes closed-loop feedback-based link adaptation practical while boosting spatial reuse gain even in the DSS environment. To implement the group-wise DSS, we propose a single unified framework that employs elementary algorithms for BS grouping and carrier-sensing threshold adjustment. Our system-level simulation results demonstrate that the proposed framework boosts the areal capacity gain by approximately 4.42 times as much as the conventional approach.