TY - JOUR
T1 - Bounds and protocols for a rate-constrained relay channel
AU - Sabharwal, Ashutosh
AU - Mitra, Urbashi
N1 - Funding Information:
Manuscript received June 11, 2005; revised December 3, 2006. The work of A. Sabharwari was supported in part by the National Science Foundation under Grants CNS-0325971, CNS-0551692, and CNS-0619767. The work of U. Mitra was supported in part by the Office of Naval Research under Grant N-000140410273, by the National Science Foundation ITR CCF-0313392, by a Texas Instruments Visiting Faculty Fellowship, and by the University of Southern California.
PY - 2007/7
Y1 - 2007/7
N2 - In this correspondence, a relay rate-constrained variation of the three-node relay channel is introduced and studied. In the rate-constrained relay channel, the relay cannot reliably decode or encode beyond a rate constraint ̄R. A cut-set upper bound, and rate-constrained variations of decode-and-forward and estimate-and-forward protocol are derived. It is observed that relative performance of the unconstrained protocols does not predict relative performance of the constrained protocols; that is, estimate-and-forward almost always offers superior performance for small rate constraints. A new metric, denoted as the relay efficiency, measures the slope at which the relay channel rate increases over the relay-less channel. While the cut-set upper bound has a relay efficiency of one for the Gaussian relay channel at ̄R = 0, the two new protocols do not. Finally, it is shown that an effective relay rate constraint can be computed for general relay protocols. To demonstrate this concept, a Markovian amplify-and-forward protocol, in which the relay does not perform any explicit encoding or decoding, is examined and compared to the introduced rate-constrained protocols.
AB - In this correspondence, a relay rate-constrained variation of the three-node relay channel is introduced and studied. In the rate-constrained relay channel, the relay cannot reliably decode or encode beyond a rate constraint ̄R. A cut-set upper bound, and rate-constrained variations of decode-and-forward and estimate-and-forward protocol are derived. It is observed that relative performance of the unconstrained protocols does not predict relative performance of the constrained protocols; that is, estimate-and-forward almost always offers superior performance for small rate constraints. A new metric, denoted as the relay efficiency, measures the slope at which the relay channel rate increases over the relay-less channel. While the cut-set upper bound has a relay efficiency of one for the Gaussian relay channel at ̄R = 0, the two new protocols do not. Finally, it is shown that an effective relay rate constraint can be computed for general relay protocols. To demonstrate this concept, a Markovian amplify-and-forward protocol, in which the relay does not perform any explicit encoding or decoding, is examined and compared to the introduced rate-constrained protocols.
KW - Amplify-and-forward
KW - Decode-and-forward
KW - Estimate- and-forward
KW - Rate-constrained channels
KW - Relay channel
UR - http://www.scopus.com/inward/record.url?scp=34447302031&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=34447302031&partnerID=8YFLogxK
U2 - 10.1109/TIT.2007.899527
DO - 10.1109/TIT.2007.899527
M3 - Article
AN - SCOPUS:34447302031
SN - 0018-9448
VL - 53
SP - 2616
EP - 2624
JO - IEEE Transactions on Information Theory
JF - IEEE Transactions on Information Theory
IS - 7
ER -