Comments on "Spectrally-efficient relay selection with limited feedback"
Li Sun, Taiyi Zhang · IEEE Journal on Selected Areas in Communications · 2009
In the above mentioned paper, the authors proposed an incremental transmission relay selection (ITRS) protocol, whose diversity-multiplexing-tradeoff performance was illustrated and compared with existing alternatives in Fig. 3. However, in our opinion, these comparisons are unfair and a curve depicted in Fig. 3 is wrong. First, in [ibid., vol. 26, no. 8, pp. 1419'1428, Oct. 2008], the authors assumed that the channel coefficients change in every time slot, whereas other protocols assumed that the channel coefficients remain constant during the whole frame transmission. This implies that an additional temporal diversity can be extracted by the ITRS protocol whereas other protocols (DDF, DSTC, opportunistic relaying) can only benefit from spatial diversity. So, the comparisons in the original paper are unfair because they are made under different conditions. Second, according to the original authors and references therein, the maximum diversity order of the ITRS protocol is M+2, and that of other protocols (DDF, DSTC, opportunistic relaying) are all equal to M+1, where M is the number of relay nodes (i.e. eight for Fig. 3). Thus, the DMT curve for DSTC (it is also the DMT curve for opportunistic relaying) is wrong in Fig. 3. Note that although the authors allowed the source to participate in the second phase of transmission when they considered DSTC and opportunistic relaying, the diversity order cannot be enhanced at all due to the fact that the channel coefficients remain constant during the two phases. And our declaration holds true all the same.