sequence of symbols. This decoding operation simply involves the mapping of the received symbols into a set of values that estimate the transmitted signal. This mapping process typically imposes distortion on the original transmitted signal that may be considered as a form of noise.
The nature of wireless networks is predetermined by the mobility of communicating nodes. In other words, the distances among the source, relay and destination nodes are time-variant. Therefore, an adaptive relaying scheme should be employed, instead of a fixed strategy at each relay. Such an adaptive scheme was proposed in [125, 131], while the condition of switching between the AF and DF was further investigated in [23, 129].
2.3.3 Relaying Protocols
2.3.3.1 Traditional Relaying
As illustrated in Fig. 2.31, traditional relaying schemes require four communication phases: two for the UL and two for the DL. In the UL, the signal transmitted from the MS is broadcast to both the RS and the BS and then it is forwarded from the RS to the BS. The process is similar in the DL, but in the reverse-direction. The relaying protocol may be applied to all the above-mentioned relaying types.
BS MS
RS
Phase 1
BS MS
RS
Phase 2
BS MS
RS
Phase 3
BS MS
RS
Phase 4
Figure 2.31: Traditional four-phase relaying
The four-phase relaying protocol provides a high diversity gain as well as an improved link quality at the cost of halving the effective throughput. However, the employment of relays is capable of offering an improved link quality owing to the presence of LOS paths. Consequently, an improved bandwidth efficiency may be achievable by exploiting higher-order, higher-throughput modulation and high-rate FEC codes.
2.3.3.2 Successive Relaying
In order to reduce the number of slots required, the successive relaying technique [126, 132] was proposed, which relies on at least two relay stations. The DL communications of a two-relay aided network is illustrated in Fig. 2.32. According to the figure, in the first phase the BS broadcasts
2.3.3. Relaying Protocols 75 message m1 and RS1 listens. In the second phase, RS1 processes m1 and then forwards it to the MS. Meanwhile, the BS broadcasts the next message m2 while RS2 listens. In the third phase, the BS broadcasts message m3 and the RS1 listens again, while RS2 relays message m2 to the MS. The process continues in a manner that the two relays alternatively listen and transmit.
BS MS
Figure 2.32: Three-phase successive relaying, which requires at least an extra relay compared to the traditional four-phases relaying of Fig. 2.31.
It may be readily realized that to transmit N messages, (N + 1) slots are required. Therefore, the successive relaying philosophy may significantly reduce the throughput loss compared to the traditional four-phase relaying protocol, while the diversity order of two is still retained in case of two relays aided.
Moreover, a specific drawback of this protocol is the presence of interference among messages. For instance, message m2 is broadcast by the BS the same time as message m1, which is relayed from RS1. Consequently, the two messages interfere with each other, resulting in a degraded overall performance at the MS. Moreover, the extra relay required increases the overall infrastructure cost.
2.3.3.3 Network Coding Aided Three-Phase Relaying
In order to avoid the requirement of an extra relay as in the successive relaying protocol, the Network Coding (NC) aided relaying protocol was designed in [133, 134]. As shown in Fig. 2.33, this protocol requires three communication slots. In the first two time slots, the MS and BS transmits the codeword c1 and c2 to the RS, respectively. The two corresponding signals y1 and y2 received at the RS are decoded and then combined into a single stream. More particularly, the bit-wise exclusive OR (XOR) operation of the two decoded codewords is executed at the RS to obtain the composite codeword c3 = c1⊕ c2. Thereafter, the codeword c3 is sent to both the MS and the BS. Note that either zero padding or repetition coding has to be employed, when the lengths of c1 and c2 are different.
BS MS
Figure 2.33: Network coded three-phase relaying, which avoids the requirement of an extra relay, as in the three-phase successive relaying of Fig. 2.32.
At the destinations, the signal received is the combination of the UL and DL packets. To recover
2.3.3. Relaying Protocols 76 the desired packet, the a-priori knowledge of the transmit packet in the previous slot is employed at each destination node. For instance, the BS first computes the LLRs of the network-coded bits. To extract the soft information related to the UL packet, the sign of the network-coded bit LLR is flipped, when the corresponding DL bit was a logical 1 based on the XOR function, otherwise its magnitude is maintained. In order to achieve a diversity gain, the signal received during the third slot is often combined with the signal directly received from the source in the previous slots.
Due to requiring three time slots, the protocol obviously reduces the effective throughput by a third compared to the traditional direct communications. Similar to the four-phase relaying protocol, the high-order modulation and high-rate coding should be employed to improve the achievable throughput.
2.3.3.4 Two-Phase Relaying
To further reduce the number of communication time slots required, the two-phase relaying protocols, which invoke the AF relaying, Denoise-and-Forward (DNF) [132, 135] or DF relaying, were proposed.
BS RS MS
Phase 1
BS RS MS
Phase 2
Figure 2.34: Network coded two-phase relaying, which eliminates one of the transmission phases compared to the network coded three-phase relaying regime of Fig. 2.33, albeit this is achieved at the cost of degrading the signal quality.
Two-phase AF relaying
In first phase, the MS and the BS simultaneously transmit their signal to the RS. At the relay, the superposition of the two signals is amplified and adjusted to conform with a power constraint before being forwarded to both the MS and the BS. Since the MS and the BS have their own knowledge about the previous transmitted signal, they can subtract the so-called back-propagation self-interference prior to decoding, provided that the Channel Impulse Response (CIR) of all links is available at the MS’s and BS’s receivers. In order to relay the composite signal, the RS has to increase the transmit power. Furthermore, the noise and interference amplification is unavoidable in this relaying protocol.
Therefore, the overall BER performance of this protocol becomes significantly degraded compared to those of the above-mentioned protocols , which is the price paid for the reduction in the number of time slots employed.
NC aided two-phase relaying
Similar to the two-phase AF relaying, in the DNF relaying the BS and the MS simultaneously transmit their signal to the RS during the first phase of cooperating. Instead of amplifying the superposition based signal, the RS detects the modulated symbols received and generates a combined signal by employing the XOR-ed operation of the UL and the DL packet, which is the same as the corresponding operation of the NC aided three-phase relaying. This process of detecting the symbols is known as the denoising operation and hence the resultant relaying protocol is referred to as NC aided DNF relaying. As a benefit of the denoising process, the relaying protocol typically offers a
2.3.4. Performance of Wireless Cooperative Networks 77