# US 2016/0233979 A1 · Method and System for Reliable Data Communications with Adaptive Multi-Dimensional Modulations for Variable-Iteration Decoding

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2016/0233979 A1
- Document ID: US20160233979A1
- Status: Published application
- Kind: A1
- Application number: 14/619,392
- Filed: 2015-02-11
- Published: 2016-08-11
- Earliest claimed: 2015-02-11
- Assignees: Mitsubishi Electric Research Laboratories, Inc.
- Applicants: Mitsubishi Electric Research Laboratories, Inc.
- Inventors: Toshiaki Koike-Akino
- Web page: https://patentgrep.com/patent/US20160233979A1

## CPC

- H03M 13/116: {Quasi-cyclic LDPC [QC-LDPC] codes, i.e. the parity-check matrix being composed of permutation or circulant sub-matrices}
- H03M 13/036: {Heuristic code construction methods, i.e. code construction or code search based on using trial-and-error}
- H03M 13/1102: {Codes on graphs and decoding on graphs, e.g. low-density parity check [LDPC] codes}
- H03M 13/1111: {Soft-decision decoding, e.g. by means of message passing or belief propagation algorithms}
- H03M 13/1154: {Low-density parity-check convolutional codes [LDPC-CC]}
- H03M 13/1191: {Codes on graphs other than LDPC codes}
- H03M 13/21: Non-linear codes, e.g. m-bit data word to n-bit code word [mBnB] conversion with error detection or error correction
- H03M 13/251: {with block coding}
- H03M 13/255: {with Low Density Parity Check [LDPC] codes}
- H03M 13/353: {Adaptation to the channel}
- H03M 13/616: {Matrix operations, especially for generator matrices or check matrices, e.g. column or row permutations}
- H03M 13/6325: {Error control coding in combination with demodulation}
- H03M 13/6516: {Support of multiple code parameters, e.g. generalized Reed-Solomon decoder for a variety of generator polynomials or Galois fields}
- H04L 1/0026: {Transmission of channel quality indication}
- H04L 1/0035: {evaluation of received explicit signalling}
- H04L 1/1829: Arrangements specially adapted for the receiver end
- H03M 13/136: {Reed-Muller [RM] codes}
- H03M 13/138: {Codes linear in a ring, e.g. Z4-linear codes or Nordstrom-Robinson codes}
- H03M 13/1505: {Golay Codes}
- H03M 13/19: Single error correction without using particular properties of the cyclic codes, e.g. Hamming codes, extended or generalised Hamming codes
- H04L 1/0003: {by switching between different modulation schemes}
- H04L 1/0009: {by adapting the channel coding (H04L1/1812 takes precedence)}
- H03M: CODING; DECODING; CODE CONVERSION IN GENERAL (using fluidic means F15C4/00; optical analogue/digital converters G02F7/00; coding, decoding or code conversion, specially adapted for particular applications, see the relevant subclasses, e.g. G01D, G01R, G06F, G06T, G09G, G10L, G11B, G11C, H04B, H04L, H04M, H04N; ciphering or deciphering for cryptography or other purposes involving the need for secrecy G09C)
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

In an advanced adaptive modulation and coding (AMC) scheme, the code rate and the parity-check matrix (PCM) for low-density parity-check (LDPC) codes are adapted according to modulation formats and variable-iteration receivers. The degree distribution for the PCM adaptation is designed by heuristic optimization to minimize the required SNR via an extrinsic information transfer (EXIT) trajectory analysis for finite-iteration decoding. The method uses dynamic window decoding by generating spatially coupled PCM for quasi-cyclic LDPC convolutional coding. The method also provides a way to jointly optimize labeling and decoding complexity for high-order and high-dimensional modulations. The problem to use a large number of different LDPC codes for various modulation formats and variable-iteration decoding is also dealt with by linearly dependent PCM adaptation across iteration count to keep using a common generator matrix. This PCM adaptation can improve a convergence speed of belief propagation decoding and mitigate an error floor issue.

## Claims

### Claim 1

A method for adaptive modulation and coding (AMC) in a communications network, comprising steps:
coding and decoding data using a set of codes having variable rate and variable parity-check matrix (PCM); and
modulating and demodulating the data using a set of modulation formats having variable order and variable dimension, wherein the coding and the modulating are performed in a transmitter and the demodulating and the decoding are performed in a receiver.

### Claim 2 (depends on 1)

The method of claim 1, further comprising a step:
selecting a best combination of one of the codes and one of the modulation formats, depending on a signal-to-noise ratio (SNR) and iteration counts for the demodulating and the decoding, according to a network requirement of power consumption, latency, bit error rate (BER), data rate, and complexity of the coding, modulating, demodulating and decoding.

### Claim 3 (depends on 1)

The method of claim 1, further comprising steps:
encoding the data using a selected generator matrix to produce encoded data;
modulating the encoded data using the selected modulation format to produce modulated data;
transmitting the modulated data from the transmitter to the receiver over a channel;
receiving output of the channel as noisy data;
demodulating the noisy data using a soft-input soft-output maximum a posteriori probability (MAP) algorithm to calculate a log-likelihood ratio (LLR) to produce demodulated data; and
decoding the demodulated data using a selected PCM, wherein a number of decoding iterations and a number of demodulating iterations are variable.

### Claim 4 (depends on 1)

The method of claim 1, wherein the decoding uses a set of low-density parity-check (LDPC) codes having variable rate and variable PCM, wherein multiple PCMs have different degree distributions to achieve a minimum required SNR designed for different modulation parameters, different receiver parameters, and different coding parameters.

### Claim 5 (depends on 1)

The method of claim 1, wherein the decoding uses a series of multiple PCMs for different iteration counts, wherein the i-th PCM has a different degree distribution designed for the i-th decoding iteration, wherein one or more of the multiple PCMs are linearly dependent and orthogonal to one common generator matrix used at the transmitter.

### Claim 6 (depends on 4)

The method of claim 4, wherein designing the multiple PCMs comprises steps;
optimizing a degree distribution for variable nodes and check nodes in a bipartite graph for a given combination of the coding parameters, the modulation parameters, and the receiver parameters;
optimizing a girth according to the optimized degree distribution; and
adapting the optimized PCM to refine for different combinations of the modulation parameters and the receiver parameters.

### Claim 7 (depends on 6)

The method of claim 6, wherein optimizing the degree distribution comprises steps;
setting up the degree distribution with average and maximum degree constraints;
modifying a bit labeling for a high-order and high-dimensional modulation;
analyzing mutual information updates across decoding iterations given the degree distribution by using an extrinsic information transfer (EXIT) trajectory analysis;
searching for a required SNR achieving the mutual information of one by using a line search algorithm;
iterating from the setting to the searching steps using heuristic optimization methods until a maximum iteration of optimizations reaches a pre-defined number; and
outputting the best degree distribution and the corresponding required SNR.

### Claim 8 (depends on 6)

The method of claim 6, wherein the degree distribution is obtained by a protograph base matrix, in which non-zero elements are confined in a band diagonal matrix so that quasi-cyclic (QC) LDPC convolutional code is designed for low-complexity encoding and low-latency decoding with a dynamic window decoding, wherein QC-LDPC codes use a Galois field or a Lie ring.

### Claim 9 (depends on 7)

The method of claim 7, wherein heuristic optimization methods use a multi-objective function to search for a set of Pareto optimal solutions, wherein the multi-objective function comprises joint minimizations of the required SNR, BER, encoding complexity, decoding complexity, average degree, maximum degree, decoding latency, and circuit size, by using a multi-objective variant of differential evolution, evolutionary strategy, simulated annealing, genetic algorithm, or swarm optimization.

### Claim 10 (depends on 6)

The method of claim 6, wherein optimizing the girth comprises steps;
generating the PCM having an optimized degree distribution by using a progressive edge growth (PEG) or a greedy linear transform;
calculating a corresponding generator matrix by using a Gaussian elimination of the optimized PCM; and
outputting the optimized PCM and generator matrix.

### Claim 11 (depends on 7)

The method of claim 7, wherein the EXIT trajectory analysis comprises steps:
emulating a communications network for a target SNR via Monte-Carlo runs;
calculating LLR via the MAP algorithm, wherein the LLR is algebraically inverted by a precoded coset leader; and
analyzing mutual information updates across iterations for finite-iteration decoding and demodulating.

### Claim 12 (depends on 11)

The method of claim 11, wherein the mutual information updates account for a standard deviation loss due to finite-precision computations and finite-length codes, wherein the deviation loss is empirically modeled as a function of input mutual information, the variable node degree, the check node degree, the code length, and precision digits through Monte-Carlo runs.

### Claim 13 (depends on 1)

The method of claim 1, wherein the PCM is variable depending on the modulation format.

### Claim 14 (depends on 1)

The method of claim 1, wherein the set of modulation formats have different orders and different dimensions, further comprising;
block-coded high-dimensional modulations, comprising 4-dimensional (4D) simplex modulation, 8D modulation based on an extended Hamming code, 16D modulation based on a Nordstrom-Robinson nonlinear code, and 24D modulation based on an extended Golay code;
sphere-cut lattice-packed modulations, comprising 4D checker board lattice, 6D diamond lattice, 12D Coxeter-Todd lattice, 16D Barnes-Wall lattice, and 24D Leech lattice;
space-time modulations, comprising Cayley-transformed unitary modulation, Alamouti modulation, and Reed-Muller operator modulation; and
quadrature-amplitude modulations (QAM), comprising phase shift keying (PSK), and pulse-amplitude modulation (PAM).

### Claim 15

An adaptive modulation and coding (AMC) controller, comprising:
a memory storing
a set of codes having variable rate and variable PCM for encoding and decoding data; and
a set of modulation formats having variable order and variable dimension for modulating and demodulating the data; and
an AMC selector for selecting a best combination of one of the codes and one of the modulation formats for coding and modulating in a transmitter, and demodulation and decoding in a receiver.
