Back to results
US 4,788,694 AGranted
Trellis coding with substrates
AssigneeAmerican Telephone and Telegraph Company, AT&T Bell Laboratories
InventorArthur R. Calderbank
Life of this patent
Estimating the term…
Loading patent…
Loading patent…
AssigneeAmerican Telephone and Telegraph Company, AT&T Bell Laboratories
InventorArthur R. Calderbank
Estimating the term…
A succession of (k+n)-bit input words are coded for transmission over a (1 - D) partial response channel by applying n of the bits of each word to a trellis coder which, in turn, selects one of a plurality of 2.sup.m cosets from a predefined lattice, where M>n. A signal point, or "seed," is then selected as a function of both (a) the values of the other k bits and (b) the value, referred to as the "channel state", of the last pulse in the previous signalling interval, in such a way as to ensure that the signal point presented to the receiver is a point in the selected cost. Viterbi decoding on the cosets is then used in the receiver to recover the most likely sequence of transmitted signal points and, in turn, the (k+n)-bit input words.
A data transmitter for generating and sequentially applying to a transmission channel during each of a succession of signalling intervals at least a pair of signal pulses representing a respective one of a succession of (k+n)-bit input words, said channel being such that at least one of the signal pulses applied thereto during each one of said signalling intervals is modified as a function of the value of at least one signal pulse applied to the channel in a prior signalling interval, said transmitter comprising,
A data transmitter for generating and sequentially applying to a transmission channel during each of a succession of signalling intervals at least a pair of signal pulses representing a respective one of a succession of (k+n)-bit input words, said channel having a curent channel state defined by the value of at least one signal pulse applied to said channel in a prior signalling interval and said channel being such that at least one of the signal pulses applied thereto during each one of said signalling intervals is modified as a function of said current channel state, said transmitter comprising,
A transmitter for generating a sequence of outgoing signals for transmission over a communications channel in response to a sequence of (k+n)-bit input words, said channel having a channel state which, at any point in time, is a function of the values of a particular one or more of the signals previously applied to said channel, said channel changing the value of each signal transmitted thereover as a function of the then current channel state, said transmitter comprising
A method for generating and sequentially applying to a transmission channel during each of a succession of signalling intervals at least a pair of signal pulses representing a respective one of a succession of (k+n)-bit input words, said channel being such that at least one of the signal pulses applied thereto during each one of said signalling intervals is modified as a function of the value of at least one signal pulse applied to the channel in a prior signalling interval, said method comprising the steps of,
A method for generating and sequentially applying to a transmission channel during each of a succession of signalling intervals at least a pair of signal pulses representing a respective one of a succession of (k+n)-bit input words, said channel having a current channel state defined by the value of at least one signal pulse applied to said channel in a prior signalling interval and said channel being such that at least one of the signal pulses applied thereto during each one of said signalling intervals is modified as a function of said current channel state, said method comprising the steps of,
A method for generating a sequence of outgoing signal points for transmission over a communications channel in response to a sequence of (k+n)-bit input words, said channel having a channel state which, at any point in time, is a function of at least a portion of a particular one or more of the signal points previously applied to said channel, said channel changing each signal point transmitted thereover as a function of the then current channel state, said method comprising the steps of
A method comprising the steps of
Dependencies are inferred from the claim text.
Drawings are temporarily unavailable. Open the PDF at USPTO
Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words:…
Arrangements at the receiver end: maximum-likelihood or sequential decoding, e.g. Viterbi, Fano, ZJ algorithms
Using code conversion at the transmitter; using predistortion; using insertion of idle bits for obtaining a desired frequency spectrum; using…:…
Using code conversion at the transmitter; using predistortion; using insertion of idle bits for obtaining a desired frequency spectrum; using…: by…
Term ended 20 Feb 2007 (estimated)
USPTO file history isn’t available for this patent, so only indexed dates are shown.
The present invention relates to data transmission systems and, in particular, to channel coding in such systems.
Much attention has been focused in recent years on channel codes which provide so-called coding gain. Prominent among these are the so-called "trellis" codes described in such papers as G. Ungerboeck, "Channel Coding With Multilevel/Phase Signals," IEEE Trans. Informaton Theory, IT-28, 1982, pages 55-67; A. R. Calderbank and N. J. A. Sloane, "A New Family of Codes for Dial-Up Voice Lines," Proc. IEEE Global Telecomm. Conf., November 1984, pages 20.2.1-20.2.4; A. R. Calderbank and N. J. A. Sloane, "Four-Dimensional Modulation With an Eight-State Trellis Code," AT&T Technical Journal, Vol. 64, No. 5, May-June 1985, pages 1005-1018; A. R. Calderbank and N. J. A. Sloane, "An Eight-Dimensional Trellis Code," Proc. IEEE, Vol. 74, No. 5, May 1986, pages 757-759; and L.-F. Wei, "Rotationally Invariant Convolutional Channel Coding With Expanded Signal Space-Part I: 180 Degrees and Part II: Nonlinear Codes," IEEE J.Select.Areas Commun., Vol. SAC-2, September 1984, pages 659-686, all of which are hereby incorporated by reference.
In typical known trellis coding arrangements, n of the bits of each of a succession of (k+n)-bit input words are applied to a so-called trellis coder which, in turn, selects one of a plurality of 2.sup.m.sub.k subsets of the signalling constellation, wherein m>n. Each subset contains 2.sup.k signalling points and the other k bits, referred to herein as the non-trellis-coded bits, are used to identify a particular one of the 2.sup.k points in the selected subset as the signal point to be communicated to the receiver. So-called Viterbi decoding is then used in the receiver to recover the most likely sequence of transmitted signal points and, in turn, the (k+n)-bit input words.
In the drawing,
FIG. 1 is a block diagram of a data transmitter embodying the principles of the invention;
FIG. 2 shows a portion of the lattice from which the signalling constellation used in the transmitter is chosen;
Encoder 10 in the transmitter of FIG. 1 receives data in the form of a serial bit stream from a data source 9. This bit stream is applied to serial-to-parallel (S/P) converter 11 within the encoder, which generates a sequence of (k+n)-bit words. Illustratively, both k=2; and n=2, and converter 11 generates these 4-bit words in parallel on its four output leads--two bits on leads 12a-b and the other two bits on leads 13a-b.
The task of the transmitter of FIG. 1 is to convey to a receiver (shown in FIG. 13) signals representing respective signal points of a constellation wherein the points of the constellation are points from a predetermined p-dimensional lattice. Illustratively, p=2. To this end, the bits on leads 13a-b, referred to herein as the "trellis-coded bits", are applied to a rate 2/3 trellis coder of conventional design. The output of coder 15 on leads 16 is a three-bit word identifying one of eight predetermined subsets of points of the aforementioned constellation. As discussed in further detail hereinbelow, the subsets, more particularly, are illustratively cosets of a predetermined sublattice of the lattice. The subset-, or coset-, identifying word, on leads 16 is applied to seed generator ROM 21, which also receives as inputs the two bits on leads 12a-b, referred to herein as the "non-trellis-coded bits". In accordance with the invention, seed generator ROM 21 receives as a further input a signal from leads 18 as will be described.
The job of seed generator ROM 21 is to generate a two-dimensional signal point (x.sub.i,y.sub.i) whose two components x.sub.i and y.sub.i are provided on leads 22 and 23, respectively. (The subscript "i" advances at the signalling rate, i.e., the rate that S/P converter 11 puts out words on its output leads. Thus a two-dimensional signal point is output by seed generator ROM 21 for each converter 11 output word.) The components x.sub.i and y.sub.i are applied to parallel-to-serial (P/S) converter 31. The latter provides on its output leads 32 a multi-bit word representing x.sub.i, followed by a multi-bit word representing y.sub.i. These are applied to a conventional baseband pulse generator 40 which, for each word applied thereto, generates a baseband pulse whose amplitude is given by the input word value. The output of generator 40 is applied to the transmission channel.