Performance Analysis and Optimal Cooperative Cluster Size for Randomly Distributed Small Cells Under Cloud RAN
Abstract : One major advantage of cloud/centralized radio access network (C-RAN) is the ease of implementation of multicell coordination mechanisms to improve the system spectrum efficiency (SE). Theoretically, large number of cooperative cells lead to a higher SE, however, it may also cause significant delay due to extra channel state information (CSI) feedback and joint processing computational needs at the cloud data center, which is likely to result in performance degradation. In order to investigate the delay impact on the throughput gains, we divide the network into multiple clusters of cooperative small cells and formulate a throughput optimization problem. We model various delay factors and the sum-rate of the network as a function of cluster size, treating it as the main optimization variable. For our analysis, we consider both base stations’ as well as users’ geometric locations as random variables for both linear and planar network deployments. The output SINR (signal-tointerference-plus-noise ratio) and ergodic sum-rate are derived based on the homogenous Poisson point processing (PPP) model. The sum-rate optimization problem in terms of the cluster size is formulated and solved. Simulation results show that the proposed analytical framework can be utilized to accurately evaluate the performance of practical cloud-based small cell networks employing clustered cooperation.
These challenges, new and existing technologies need to coexist and cooperate in order for the emerging networks to fulfill the users’ needs for high Quality of Experience (QoE).
Moreover, another factor that contributed to this throughput reduction is the intrabeam interference which increases for the previously existing UEs in the NOMA clusters when new edge users are added to them.
The existence of singleton coalitions in the game JT-CoMP case, when the conditions of the proposed coalition formation algorithm are not satisfied, resulted in a small percentage of users having the same average throughput as in the case of no JT-CoMP.
All of the parameters are converted into a function of cluster size to formulate the optimization problem.
The geographic area of a network is considered to be divided into separate clusters and an optimization problem is formulated by expressing ergodic sum-rate in terms of the cooperative cluster size.
To investigate the proposed clustering optimization problem with ZF and MRT precoding algorithms for linear and planar dense small cell deployments.
One of the relatively recently proposed concepts, aiming to fulfill the needs of mobile users in 5G networks, is the Non Orthogonal Multiple Access (NOMA) technique.
In, base station coordination with dirty paper coding was initially proposed with single-antenna transmitters and receivers in each cell.
The schemes proposed therein include a dirty paper coding approach with perfect data and power cooperation among base stations with a pooled power constraint and several sub-optimal joint transmission schemes with per-base power constraints.
We conjecture that there must be an optimal cluster size, large enough to mitigate interference into a reasonable level yet small enough to save the performance loss due to the delay-caused channel mismatch.
There must be an optimal cluster size to trade off delay and interference for maximizing system performance.
The SINR expression in is very difficult to analyze theoretically since it is a compound function of multiple variables including large-scale and fast fading of the channels, delay and precoding coefficients as well as multiple random deployed RRHs and users.
|