Probabilistic QAM Shaping for LDPC Spectral Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communication systems using LDPC-coded modulation with uniformly distributed QAM constellations suffer from a 1.53 dB gap in spectral efficiency due to non-optimal distribution of codebooks, leading to inefficiencies in error correction and energy consumption.
Innovation Solution
Implementing a probabilistic constellation shaping technique with an LDPC-coded modulation system, utilizing a shaping encoder to apply non-uniform probability distributions to QAM symbols, and adjusting LDPC code rates and parity bit management to achieve a shaping gain of 1.53 dB and maintain desired code rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If uniformly distributed QAM constellations are used with LDPC codes, then the encoding/decoding process is simple, but spectral efficiency is reduced due to non-optimal codebook distribution
Solution Approach 1:
The patent applies parameter changes by transitioning from uniform probability distribution to non-uniform probability distribution in QAM constellation mapping. Specifically, it uses probabilistic constellation shaping where different constellation points are assigned different probabilities based on their distance from the origin, optimizing the codebook distribution to match channel characteristics and achieve closer to Shannon limit spectral efficiency.
Solution Approach 2:
The patent implements local quality by applying different probability weights to different regions of the QAM constellation. Inner constellation points (closer to origin) are assigned higher probabilities while outer points are assigned lower probabilities, creating non-uniform local characteristics that optimize error correction performance and spectral efficiency simultaneously.
2Productivity
If non-uniform probability distributions are applied to QAM symbols through probabilistic constellation shaping, then spectral efficiency is improved by reducing the gap to Shannon limit, but device complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-computing and storing the probabilistic constellation shaping lookup tables and codebook configurations before actual communication. The shaping parameters and probability distributions are predetermined and stored in memory, allowing the encoder to simply retrieve and apply pre-configured mappings during transmission, significantly reducing real-time computational complexity.
Solution Approach 2:
The patent uses copying by creating standardized probabilistic shaping codebooks that can be replicated and stored. Instead of computing complex probability distributions in real-time, the system copies pre-generated optimal constellation mappings from storage memory, reducing encoder/decoder complexity while maintaining spectral efficiency improvements.
3Reliability
If probabilistic constellation shaping is implemented, then error correction capabilities are improved, but the complexity of generating and managing parity bits increases
Solution Approach 1:
The patent merges the probabilistic constellation shaping function with the LDPC encoding process by integrating the shaping operation into the code generation stage. The shaped constellation points are directly used as LDPC codebook entries, combining what would traditionally be separate operations (constellation shaping followed by encoding) into a unified process that reduces parity bit management complexity.
Solution Approach 2:
The patent applies universality by designing the LDPC code structure to work with both uniform and non-uniform constellation distributions. The same LDPC encoder can handle both conventional QAM and probabilistically shaped QAM by simply changing the input distribution, making the parity bit generation mechanism universally applicable without requiring separate complex generation processes for different modulation types.
Data Source
AI summary
An apparatus may include a transmitter and one or more processors configured to identify a code rate of a low-density parity-check (LDPC) code, receive, by an LDPC encoder, a set of information bits and encode, using the code rate, the set of information bits to generate a set of encoded bits and a set of parity bits, generate, from the set of encoded bits, a matrix of bit arrays, discard one or more parity bits from the set of parity bits to generate a parity bit array with a size equal to a number of column of the matrix, generate a bit array by concatenating (1) one or more bit arrays selected from the matrix of bit arrays and (2) one or more bits selected from the parity bit array corresponding to the one or more bit arrays, and modulate, by a modulator, the bit array to generate modulated data.