DocumentCode :
2126965
Title :
A low power turbo decoder architecture
Author :
Elassal, Mahmoud ; Bayoumi, M.
Author_Institution :
Center for Adv. Comput. Studies, Univ. of Louisiana at Lafayette, LA, USA
fYear :
2003
fDate :
27-29 Aug. 2003
Firstpage :
105
Lastpage :
110
Abstract :
A new method is proposed to decrease the power consumption of turbo decoders. Turbo decoders iteratively decode the received sequence by exchanging extrinsic information. In the proposed method, the exchange of extrinsic information that exceeds a certain threshold is terminated and, instead, a predefined value is exchanged. This method enables reduction of memory accesses to both the interleaver memory and state metric memory, hence reducting power consumption. Simulation and synthesis results shows up to 25% reduction of power consumption at Eb/NO of 1.5 dB, compared to the conventional state parallel architecture.
Keywords :
convolutional codes; interleaved codes; iterative decoding; power consumption; turbo codes; extrinsic information; interleaved convolutional codes; interleaver memory; iterative decoding; low power turbo decoder architecture; memory accesses; power consumption; state metric memory; turbo codes; Computer architecture; Convolutional codes; Energy consumption; Iterative decoding; Maximum likelihood decoding; Parallel architectures; Performance analysis; Termination of employment; Transceivers; Turbo codes;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Signal Processing Systems, 2003. SIPS 2003. IEEE Workshop on
ISSN :
1520-6130
Print_ISBN :
0-7803-7795-8
Type :
conf
DOI :
10.1109/SIPS.2003.1235652
Filename :
1235652
Link To Document :
بازگشت