DocumentCode
2323920
Title
Joint Power Allocation for Nonregenerative MIMO-OFDM Relay Links
Author
Hammerström, Ingmar ; Wittneben, Armin
Author_Institution
Commun. Technol. Lab., ETH Zurich
Volume
4
fYear
2006
fDate
14-19 May 2006
Abstract
We consider a two-hop MIMO-OFDM communication scheme with a source, a relay, and a destination. The relay is assumed to be nonregenerative (or amplify-and-forward (AF)). We assume channel state information at each transmitter (CSIT), source and relay. We present a jointly optimized power allocation (PA) over the subchannels in space and frequency domain at source and relay with a joint transmit power constraint. The PA is based on a high SNR approximation of the SNR expression at the destination. In the low SNR regime, the performance is still very tight to the optimal rate. The presented PA can be computed with low computational complexity and achieves a considerable performance gain compared to a uniform PA at source and relay. To further enhance the performance of the considered scheme, all subchannels, in space and frequency domain, of the source to relay and relay to destination channel are paired according to their actual magnitude
Keywords
MIMO systems; OFDM modulation; computational complexity; radio links; wireless channels; SNR approximation; channel state information; computational complexity; joint power allocation; joint transmit power constraint; multiple input multiple output system; nonregenerative MIMO-OFDM relay links; orthogonal frequency division multiplexing; signal-to-noise ratio; Communications technology; Computational complexity; Constraint optimization; Frequency domain analysis; MIMO; OFDM; Power system modeling; Power system relaying; Relays; Throughput;
fLanguage
English
Publisher
ieee
Conference_Titel
Acoustics, Speech and Signal Processing, 2006. ICASSP 2006 Proceedings. 2006 IEEE International Conference on
Conference_Location
Toulouse
ISSN
1520-6149
Print_ISBN
1-4244-0469-X
Type
conf
DOI
10.1109/ICASSP.2006.1660902
Filename
1660902
Link To Document