Constellation Mapping for Non-Integer Information Rates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional digital modulation formats in communication systems, such as PAM and QAM, often fail to provide non-integer information rates and equiprobable distribution of constellation symbols, which are essential for efficient data transmission in applications like medium-reach fiber-optic links.
Innovation Solution
The proposed solution involves an electronic controller that employs Gray-coded mapping and Constant Composition Distribution Matching to select and exclude specific constellation symbols, achieving a smaller gap to the Shannon limit by generating bitstreams that carry labels of a subset of constellation symbols, thereby optimizing symbol distribution and error correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional digital modulation formats (PAM, QAM) are used, then the system is simple and easy to implement, but it fails to provide non-integer information rates and equiprobable distribution of constellation symbols
Solution Approach 1:
The patent segments the constellation symbols into two sets: a first set of M1 symbols with first amplitude levels and a second set of M2 symbols with second amplitude levels. This segmentation enables independent control of symbol probabilities and amplitudes, allowing non-integer information rates while maintaining systematic structure. The segmentation is implemented through separate encoding processes for each symbol set.
Solution Approach 2:
The patent applies different amplitude levels and probability distributions to different segments of the constellation. The first set of symbols uses first amplitude levels with first probabilities, while the second set uses second amplitude levels with second probabilities. This local differentiation enables equiprobable distribution within each segment while achieving non-integer overall information rate.
2Reliability
If conventional constellation mapping is used, then the implementation is straightforward, but it cannot achieve equiprobable distribution of constellation symbols
Solution Approach 1:
The patent performs preliminary action by pre-defining two distinct sets of constellation symbols with specific amplitude levels and assigning them specific probabilities before the actual modulation process. The first set of M1 symbols is prepared with first amplitude levels and first probabilities, while the second set of M2 symbols is prepared with second amplitude levels and second probabilities. This preliminary configuration enables equiprobable distribution to be achieved systematically during transmission.
3Loss of energy
If standard modulation schemes are used, then the system is well-established and reliable, but it creates a larger gap to the Shannon limit
Solution Approach 1:
The patent introduces dynamic probability assignment where the first set of symbols and second set of symbols are transmitted with different probabilities (first probability and second probability respectively). This dynamic adjustment of symbol transmission probabilities enables the system to optimize energy efficiency by adapting the constellation usage to match channel conditions, thereby reducing the gap to the Shannon limit compared to static uniform distribution schemes.
Data Source
AI summary
Architectures for inter-converting bitstreams and symbol streams of PAM and/or QAM constellations of different sizes that are not base-2 integers. Some of such constellations may be Gray-coded, and the constellation mapping may be performed to achieve an equiprobable distribution of different constellation symbols. Some embodiments may be compatible with FEC schemes. In an example embodiment, a transmitter DSP may employ a conventional constellation mapper preceded by an electronic encoder programmed to exclude some constellation-symbol labels from the bitstream applied to the mapper. In different embodiments, the electronic encoder may employ a CCDM and/or a long-division operation to select some amplitudes of the constellation and to exclude others. At least some embodiments are beneficially capable of achieving a smaller gap to the Shannon limit than comparable conventional solutions.