Trellis-coded modulation with redundant signal sets Part II: State of the art

Detection of coded and distorted QAM signals

“…The m' + 1 coded bits are used to select one of the zo'" subsets of the 2"" signal points. The remaining uncoded bits determine which of the zo,·' points in the subset is to be transmitted [68). This is further described by considering an example, where a 32-level QAM signal constellation is partitioned into eight subsets, each subset representing four signal points, giving m=4, and m' =2.…”

“…This can cause catastrophic errors in the decoding of the received data [4,68]. The problem is solved by employing differential encoding techniques that remove these rotations [4,5,63].…”

“…Many trellis coded schemes based on the Ungerboeck technique which use differential encoding, have been considered as international CCITT Standards for 9600 bitls transmission over the public switched telephone network [5,9,[68][69][70][71], and 14400 bitls transmission over private lines [7]. The codes used are [10,11,61].…”

“…- encoder is known as the G8 encoder [9], and is one of the trellis coded modulation schemes which have been considered as CCITT standards for V32 modems [9,68].…”

“…NonIinear elements (NOT and AND logic gates in Fig. 3.3) have been used in the encoder circuit to make the code invariant to 9<1' rotations, while maintaining its coding gain [5,9,68]. The differential encoder is used here to resolve the problem of the phase ambiguity [9].…”