Article Preview
TopIntroduction
One of the major properties of wireless mesh networks (WMNs) consists in the possibility of breaking long distances into a series of shorter hops. Apart from increasing the signal quality of the links, the mesh architecture allows the cooperative forwarding of data packets through intermediate terminals in the network. Cooperative communications provide an interesting contribution in this context. More precisely, they enable data transmission between two terminals through an alternate path when the direct wireless link is experiencing a deep fade. Cooperative communications can be envisioned at several network layers. However, implementing the forwarding scheme at the lowest layers renders the protocol more reactive to network conditions and minimizes the transmission delay since each layer adds its own processing time and hence includes its own latency. Cooperative protocols are mainly implemented in two layers: cooperative transmissions are managed at the physical (PHY) layer whereas the set up of the cooperation is done at the medium access control (MAC) layer. At the PHY layer, cooperative communications increase the wireless link reliability. In a cooperative scenario, a source terminal S sends data to a destination terminal D through a direct path. One or several relay terminals help the transmission by receiving the source message and forwarding it to D through a relaying path (Figure 1). Hence the direct path is rendered more robust (Laneman & Wornell, 2000, 2003; Sendorais, Erkip, & Aazhang, 2003; Hunter & Nosratinia, 2006). However, this comes at the price of bandwidth consumption so that the system operates at diminished capacity1. One common way to compare cooperative transmission techniques is to compute the diversity-multiplexing tradeoff (DMT) (Zheng & Tse, 2002). The DMT analysis of a transmission scheme yields the diversity gain achievable for a spatial multiplexing gain . The diversity gain helps in quantifying the robustness of the S-D link and the multiplexing gain gives an hint on the capacity of the link. Both indicators should be maximized in order to get an optimal DMT curve. When relay candidates are involved in a cooperative scenario, the optimal DMT curve is achievable by protocols that implement both on-demand relaying and a selection of the best relay (Bletsas, Khisti, & Win, 2008; Escrig, 2010): for . In an on-demand relaying scenario (Laneman, Tse, & Wornell, 2004; Gomez, Alonso-Zarate, Verikoukis, Perez-Neira, & Alonso, 2007), the relay terminal transmits only when D fails in decoding the data transmitted by S. Thus, the bandwidth consumption due to cooperative transmissions is minimized. Moreover, when cooperation is needed, only the best relay terminal retransmits the source message (Bletsas, Khisti, Reed, & Lippman, 2006). This optimizes the robustness of the wireless link between the source terminal and the destination terminal through the property of spatial diversity while minimizing the resource consumption compared to the case of multiple relays. Hence, an optimal tradeoff between link robustness and bandwidth consumption is reached. This optimal tradeoff has been demonstrated by computing the DMT curve of the transmission scheme. Moreover, minimizing the number of relays also reduces the impact of cooperative communications on the rest of the network. Indeed, reducing the number of relays diminishes the contention area due to cooperative transmissions compared to the case of multiple relay transmissions. Note however that the DMT criterion fails in providing indications on the amount of bandwidth that is used at the MAC layer in order to implement the cooperative network. For instance in Laneman et al. (2003), the overhead induced by the allocation of space-time codes to relay terminals has not been taken into account. Practically, further optimization is required at the MAC layer in order to reduce the overhead due to the implementation of the cooperative network. In particular, the fast selection of appropriate relay terminals is a main issue the design of cooperative MAC protocols.