DocumentCode :
2647506
Title :
VLSI Implementation of Area-efficient List Sphere Decoder
Author :
Lee, Seungbeom ; Lee, Jin ; Sang-ho Seo ; Park, Sin-Chong
fYear :
2006
fDate :
12-15 Dec. 2006
Firstpage :
610
Lastpage :
613
Abstract :
Generating soft information in list sphere decoder (LSD) increases the computational complexity to select a specific number of candidate lattice points. Among the techniques to reduce the computational complexity, we find that real-valued operation is able to reduce the per-node complexity, ordering process, and storage elements significantly with some tradeoff in speed degradation compared to complex-valued operation. This is useful in systems using large number of antennas and high-order modulation. This paper presents a real-valued LSD architecture for a mode of 4x4 64QAM and compares the hardware complexity of the proposed architecture to that of complex-valued LSD. Although the processing cycle of the proposed architecture is twice that of complex-valued LSD, the hardware complexity of the proposed architecture is less than a fourth of that of the complex-valued LSD. The proposed architecture is also implemented using 0.25um technology and attains an average throughput of 430kvectors/sec at a signal-to-noise ratio (SNR) of 20dB.
Keywords :
MIMO communication; VLSI; computational complexity; decoding; quadrature amplitude modulation; 64QAM; VLSI; computational complexity; list sphere decoder; Computational complexity; Computer architecture; Degradation; Detectors; Hardware; Iterative decoding; MIMO; Maximum likelihood decoding; Maximum likelihood detection; Very large scale integration;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Signal Processing and Communications, 2006. ISPACS '06. International Symposium on
Conference_Location :
Tottori
Print_ISBN :
0-7803-9732-0
Electronic_ISBN :
0-7803-9733-9
Type :
conf
DOI :
10.1109/ISPACS.2006.364730
Filename :
4212348
Link To Document :
بازگشت