Matched Design Method for Concatenated Trellis-Coded Modulation

概要

論文の詳細を見る
A new design method, which is referred to as the matched design method, for concatenated trellis-coded modulation (TCM) is presented. Most of the conventional concatenated TAM employs TCM designed to maximize the minimum squared Euclidean free distance, d^2_<free>. With the matched de-sign method, we maximize d^2_1(t) instead of d^2_<free>, where d^2_1(t) is the effective minimum squared Euclidean distance (MSED)when the outer code has a t-error correcting capability. The effective MSED is derived from the Euclidean/Hamming (E/H) joint weight distribution of terminated TCM. We here assume the concatenated TCM whose transmitted symbol corresponds to a symbol of outer code. The new classes of 2-dimensional(2D)and 4-dimensional(4D) codes are found by a computer search.Under the performance measures of the effective MSED or the effective multiplicity, these codes are superior to the conventional codes such as the Ungerboeck's 2D-codes when those are used as an inner code. We disclose an interesting fact that the new class of codes using rate- 1/2 encoder is superior to the class of codes using rate-2/3 encoder. This fact implies that the codes using rate- 1/2 encoder have two advantages: 1) better overall decoding performance and 2) less decoding complexity.
社団法人電子情報通信学会の論文
1996-11-25