On the performance of subcarrier allocation techniques for multiuser OFDM cognitive networks with reconfigurable antennas

Abstract

Reconfigurable antennas (RA) have been viewed as a hardware-efficient alternative solution to multiple-input multiple-output systems whereby network users can vary the antenna radiation patterns using a single antenna element to maximize the received signal strength. In this paper, we study the potential benefits of employing RA in multiuser orthogonal frequency division multiple access cognitive heterogeneous networks (Het-Nets) in terms of the overall network capacity. In cognitive HetNets, a secondary (unlicensed) network is allowed to share the spectrum with the primary (licensed) network under the condition that the interference level at the primary network is below a predetermined value. To account for this interference constraint, the secondary user (SU) can limit their transmission power and thus reducing substantially its performance. Moreover, the large number of users expected for next generation network brings dense interference to the secondary network and thus even efficient interference mitigation and resource allocation techniques can fail in maintaining the required performance level. Therefore, in this paper, we consider utilizing an RA at the SUs that acts as an additional resource which can be optimized by selecting the best state that maximizes the signal strength among the SUs and limits the mutual interference between the secondary and primary network. In particular, we propose a game theoretical framework for selecting the subcarriers based on best allocation techniques as well as the random antenna state selection that maximizes the overall capacity of the network while obeying the interference level in the primary network. We use potential games which guarantee the Nash equilibrium existence. Our results show that by selecting the optimal RA state and the subcarriers for each user, the capacity of the secondary network increases substitutionally with limited hardware complexity.