Full-Duplex MAC Protocol Based on
Adaptive Contention Window for
Visible Light Communication
Zhong Wang, Yonghe Liu, Yaping Lin, and Shengye Huang
Abstract—As an alternative technology to radiofrequency-based communication, visible light communication (VLC) has attracted tremendous interest from both academia and industry. For its medium access control (MAC) protocol, most existing design has focused on the half-duplex mode, which can be inefficient for two-way communications. In this paper, we propose a novel fullduplex MAC protocol for VLC networks. Different from half-duplex operation, full-duplex operation allows concurrent sending and receiving of data frames between the central node and terminal nodes. We show that under certain scenarios, it can achieve almost threefold downlink throughput from the central node to terminal nodes as compared with half-duplex operation. Unfortunately, owing to asymmetric traffic demand on the uplink and downlink, full-duplex transmission cannot always be realized. To increase the probability of full-duplex transmission, we further propose a self-adaptive minimum contention window full-duplex MAC protocol that can significantly increase the probability of duplex operation. Our experimental study shows that the proposed MAC protocol can effectively mitigate channel collisions and achieve significantly higher system throughput.
Index Terms—Contention window; Medium access control (MAC) protocols; Visible light communication.
R ecently, visible light communication (VLC) based onlight-emitting diodes (LEDs) has attracted significant interest from both industry and academia. Compared to the ever-popular radio-based wireless communication,
VLC presents an alternate with potential for high bandwidth and data rates, high transmit power, harmlessness to the human body, high security, and minimal electromagnetic interference with RF systems [1–4].
VLC technology was standardized in IEEE 802.15.7 in 2011 . An extensive set of work has since focused on improving its performance by increasing its transmission rate and avoiding multinode transmission interference. Owing to the nature of physical (PHY) layer transmission, there has been little study on the higher link layer medium access control (MAC) protocol design. Indeed, the MAC protocol defined in IEEE standard 802.15.7 and various other
MAC protocols work only in half-duplex mode [6–10]. Generally speaking, there are two disadvantages for operating in half-duplex mode. First, it will incur low bandwidth utilization when bidirectional transmissions are needed, especially for short messages. Second, for transmission delay, half-duplex when facing a command frame can take a long period of time as well. In contrast, full-duplex mode, with an effective MAC protocol, can significantly improve channel efficiency.
In this paper, we propose a novel full-duplex MAC protocol for VLC networks. Different from half-duplex operation, full-duplex operation allows concurrent sending and receiving of data frames between the central node and terminal nodes. It can effectively mitigate the hidden terminal problem and improve fairness among all nodes.
We show that under certain scenarios, it can achieve almost threefold increase in downlink throughput from central node to terminal nodes as compared with half-duplex operation. Unfortunately, under certain circumstances, owing to asymmetric traffic demand on the uplink and downlink, full-duplex transmission cannot always be realized. To increase the probability of full-duplex transmission, we further propose a self-adaptive minimum contention window (SACW) full-duplex MAC protocol that can significantly increase the probability of duplex operation.
The key idea here is to allow the central node to monitor the data traffic in real time and control the channel access probability by changing the contention window adaptively.
This way, the probability of successful full-duplex transmission can be controlled and the downlink can achieve higher throughput if needed. Our experimental study shows that the proposed MAC protocol can effectively mitigate channel collisions and achieve significantly higher system throughput.
The remainder of this paper is organized as follows. In
Section II, we briefly introduce related works, followed by the proposed MAC protocol and SACW algorithm in
Section III. We present our simulation study and performance analysis in Section IV. Finally, concluding remarks are given in Section V.http://dx.doi.org/10.1364/JOCN.7.000164
Manuscript received September 24, 2014; revised December 25, 2014; accepted December 30, 2014; published February 25, 2015 (Doc. ID 223497).
Z. Wang, Y. Liu, Y. Lin (e-mail: email@example.com), and S. Huang are with the College of Computer Science and Electronic Engineering,
Hunan University, No. 252, Lushan South Road, Changsha, Hunan 410082, China.
Y. Liu is also with the Department of Computer Science and Engineering, University of Texas at Arlington, Arlington, Texas 76019, USA. 164 J. OPT. COMMUN. NETW./VOL. 7, NO. 3/MARCH 2015 Wang et al. 1943-0620/15/030164-08$15.00/0 © 2015 Optical Society of America
II. RELATED WORK
The MAC protocol for VLC networks is officially specified in IEEE 802.15.7, which details various aspects including addressing, framing, error recovery, and data flow controls [11,12].
As shown in Fig. 1, both the MAC sublayer and PHY layer conceptually include management entities, termed
MAC sublayer management entity (MLME) and PHY layer management entity (PLME). These entities provide the layer management service interfaces for the layer management functions .
A MAC superframe structure can be used by a VLC coordinator to bound its channel time optionally. As shown in
Fig. 2, there are two parts of a superframe, an active part and an inactive part. The active part has 16 equal sized slots called backoff periods and is divided into three parts, a contention access period (CAP), a contention-free period (CFP), and a beacon. A slot refers to the minimum time to transmit the smallest data to a node. The inactive part is the period which the VLC coordinator may enter a lowpower mode to save power.