Title :
Soft decoding for vector quantization in combination with block channel coding
Author :
Skoglund, Mikael ; Hedelin, Per
Author_Institution :
Dept. of Inf. Theory, Chalmers Univ. of Technol., Goteborg, Sweden
Abstract :
According to the two-step source/channel coding procedure introduced by Shannon, the source and the channel codes are designed and used separately. Recent research has striven to find efficient combined approaches for source/channel coding. Much of this research has considered vector quantization (VQ) for noisy channels. We present a method for joint decoding of the combination of a vector quantizer and a channel code. Our decoder is soft in the sense that no decisions are involved in the decoding, and the unquantized channel outputs are utilized. We depart from the traditional way of decoding, in that we make the decoding into a one-step procedure, without any intermediate channel decoding. We also demonstrate that estimates of the transmitted binary data can be efficiently obtained in our framework
Keywords :
block codes; channel coding; data communication; decoding; source coding; telecommunication channels; vector quantisation; VQ; binary data transmission; block channel coding; channel codes; noisy channels; one-step decoding; soft decoding; source codes; two-step source/channel coding; unquantized channel outputs; vector quantization; AWGN; Channel coding; Covariance matrix; Decoding; Information theory; Matched filters; Vector quantization;
Conference_Titel :
Information Theory, 1995. Proceedings., 1995 IEEE International Symposium on
Conference_Location :
Whistler, BC
Print_ISBN :
0-7803-2453-6
DOI :
10.1109/ISIT.1995.550423