A Throughput-Optimum Adaptive ALOHA MAC Scheme for Full-Duplex Wireless Networks

Abstract

This paper proposes an adaptive and distributed MAC scheme for wireless networks with full-duplex radios. Fullduplex (FD) radios can exchange data simultaneously using the same frequency band and potentially double the throughput. On the other hand, it will inevitably cause extra interference to the network to use FD transmission instead of half-duplex (HD) transmission. Hence, there is a tradeoff between interference and throughput for wireless networks with FD radios. In order to mitigate the interference and maximize the network throughput, we propose the Optimum Adaptive ALOHA (OA-ALOHA) MAC scheme based on the nodes' local information about their neighbors. We derive each node's optimal transmit probability that maximizes the throughput and achieves proportional fairness. Our numerical results show that if the network is using the proposed adaptive MAC scheme, significantly better network throughput can be achieved than with the non-adaptive MAC schemes. HD can achieve around 50% maximum throughput gain using adaptive ALOHA while FD more than 70%. Moreover, a Simplified Adaptive ALOHA (SA-ALOHA) MAC scheme is also included that is based on an approximation of the optimum transmit probability and achieves comparable performance.

A Stochastic Geometry Approach to Asynchronous Aloha Full-Duplex Networks

“…Let us initially assume for the unslotted network all data transmissions to be of the same duration. In this case, a direct comparison of the throughput expressions in (14) and (29) shows how imperfect self-interference cancellation besets the two scenarios in the same way, so that Ξ is in fact independent of β and captures the intrinsic differences between the two access policies beyond specific implementation aspects. The behaviour of the throughput ratio is reported in Fig.…”

“…On the other hand, even when the capabilities of deployed nodes cannot be changed, q may still represent a key design parameter to tweak the fraction of links that shall in fact resort to full-duplex to leverage the non-trivial tradeoff between spatial reuse and additional interference at its packet duration, D utmost. The question can be effectively addressed leaning on the simple structure of (14), and recalling that Ω hd and Ω fd do not depend on q. Setting then ∂T /∂q = 0, we obtain the optimal fraction of full-duplex clusters when varying D:…”

“…The solution to this problem follows from the path loss exponent, α 2.5 3 3.5 4 4. 5 5 peak throughput gain, χ simple dependence of T on the packet duration D reported in (14). Straightforward calculations allow to derive the optimal operating point D * = 1/ (λ ((1−q)Ω hd +q Ω fd )), 7 and the corresponding throughput density as…”

“…With this horizon in mind, information-theoretic tools helped in clarifying the potential of concurrent transmission and reception in toy topologies, exploring the achievable rate regions and the degrees of freedom in relation to the accuracy of self-interference cancellation and the number of available antennas [6]- [9]. Alongside with these efforts, a number of protocols at the medium access layer have been designed to bridge the gap with existing standards, showing by means of simulations and testbed results interesting improvements over a half-duplex configuration in centralised [10] and distributed ad hoc scenarios [11]- [14] based on IEEE 802.11.…”

In-Band <inline-formula> <tex-math notation="LaTeX">$\alpha $ </tex-math> </inline-formula>-Duplex Scheme for Cellular Networks: A Stochastic Geometry Approach

“…Several research efforts are exerted to study the effect of cross-mode interference on the FD performance for different types of large-scale wireless networks. In the context of ad-hoc networks, authors in [11]- [13] show that FD communications can improve the overall throughput despite the increased aggregate interference level. In the context of cellular networks, FD communications' gains are mainly quantified for the DL performance.…”