Analysis of Clustered Licensed-Assisted Access in Unlicensed Spectrum

Faced with explosive growth of data traffic and shortage of licensed spectrum, licensed-assisted access (LAA) to unlicensed spectrum has been proposed to boost system capacity. To ensure fair coexistence with WiFi systems, listen-before-talk mechanism has been standardized under LAA framework. However, in densely deployed urban networks, the system performance could severely deteriorate due to high collision probability. In this paper, we propose cooperative LAA (CLAA), where multiple LAA small base stations form a cluster and construct a virtual multiuser multiple-input single-output (MISO)/multiple-input and multiple-output (MIMO) system to transmit data cooperatively. CLAA can effectively reduce the number of contending nodes, thereby alleviating transmission collisions. A closed-form expression for the upper bound sum rate of the cluster is derived. Markov analysis is employed to derive the system collision probability and throughput for WiFi and LTE. Our analytical results point to an adequate cluster sizes, where the highest system throughput can be achieved. Extensive simulations confirm the validity of the proposed approach, and demonstrate that CLAA can increase by up to 27% the overall system throughput, and improve by 30% in fairness.