In this paper we first establish a general performance upper bound on the diversity-multiplexing tradeoff (DMT) for communication protocols for a half-duplex cooperative network where every node has a single antenna. This bound shows that the dynamic-decode-and-forward (DDF) protocol achieves optimal DMT among all variations of decode-and-forward based protocols when multiplexing gain r ≤ 1/2. Secondly, when r > 1/2 it shows that the diversity gain must be less than 1 for all protocols and for any number of relay nodes. Our bound also implies that the well-known cut-set-based DMT upper bound is not tight except for the single-relay network and that there are rooms to improve the performance of DDF protocol, thereby settling a long-time open question raised by Azarian et al.. In light of this breakthrough, a novel protocol based on a rate-split is proposed in this paper. Two variations, dynamic and non-dynamic, of the rate-split-based protocol are also presented. It is shown that the dynamic version achieves a DMT better than the DDF protocol, and in particular, meets the cut-set bound in the case of a single-relay network. The other variation, i.e., the non-dynamic one, has a much simpler implementation and reaches its optimal performance in only two channel uses. Moreover, it meets the cut-set bound in the case of a single-relay network.
[1]
Lizhong Zheng,et al.
Diversity and multiplexing: a fundamental tradeoff in multiple-antenna channels
,
2003,
IEEE Trans. Inf. Theory.
[2]
Michael Gastpar,et al.
Compute-and-Forward: Harnessing Interference Through Structured Codes
,
2009,
IEEE Transactions on Information Theory.
[3]
Philip Schniter,et al.
On the achievable diversity-multiplexing tradeoff in half-duplex cooperative channels
,
2005,
IEEE Transactions on Information Theory.
[4]
Gregory W. Wornell,et al.
Cooperative diversity in wireless networks: Efficient protocols and outage behavior
,
2004,
IEEE Transactions on Information Theory.