Turbo Decoder Lookup Table for Lower-Complexity MAP Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The computational complexity of implementing the Maximum A Posteriori (MAP) algorithm for turbo code decoding in digital communication systems is high, making it difficult to achieve efficient error correction in noisy environments.
Innovation Solution
The method involves defining a transcendent function and a quantized approximating function, generating a look-up table with non-null and null values, and using this table to retrieve values for computing quantities related to the transcendent function, thereby simplifying the decoding process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the Maximum A Posteriori (MAP) algorithm is implemented for turbo code decoding, then error correction performance is improved, but computational complexity increases
Solution Approach 1:
The patent transforms the MAP algorithm from the linear domain to the logarithmic domain by applying logarithmic transformations to probability calculations. This parameter change converts multiplication operations into addition operations, significantly reducing computational complexity while preserving the optimal error correction performance of the MAP algorithm
Solution Approach 2:
The patent introduces simplified approximation algorithms (Max-Log-MAP and Log-MAP) that use pre-computed lookup tables instead of performing complex real-time calculations. These lookup tables store pre-calculated logarithmic values that can be quickly retrieved and combined through simple addition, replacing expensive computational operations with cheap table lookups
2Measurement precision
If the MAP algorithm is implemented with full computational precision, then decoding accuracy is improved, but implementation difficulty increases
Solution Approach 1:
The patent employs simplified logarithmic domain calculations with pre-computed lookup tables that provide sufficient precision for practical applications without requiring full computational precision. The lookup tables store quantized logarithmic values that balance accuracy requirements with implementation simplicity, making the algorithm feasible for real-world deployment
Solution Approach 2:
By working in the logarithmic domain rather than the linear probability domain, the patent transforms complex multiplicative probability calculations into simpler additive operations. This parameter transformation maintains decoding accuracy while dramatically simplifying the implementation requirements
3Productivity
If lookup table values are stored for all possible argument ranges, then computation speed is improved, but memory requirements increase
Solution Approach 1:
The patent stores lookup table values only for the most frequently occurring argument ranges in the logarithmic domain, rather than pre-computing and storing values for all possible input combinations. This partial action approach stores sufficient table entries to handle typical operating conditions, achieving fast computation for the majority of cases while keeping memory requirements manageable
Solution Approach 2:
By transforming to the logarithmic domain, the patent reduces the dynamic range of values that need to be stored in lookup tables. The logarithmic transformation compresses large probability values into smaller, more manageable numbers, reducing the memory required to store lookup table entries while maintaining computational accuracy
Data Source
AI summary
A method performs iterative decoding of information coded by an error correction code. The method includes: defining a transcendent first function representing a quantity to be evaluated for the decoding method; defining a quantized second function approximating the first function; computing first values of the second function obtained based on first arguments; the first values being not null and the first arguments being variable in a limited range having a maximum limit; computing second values of the second function obtained on the basis of second arguments, the second values being null; and generating a look-up table representing the first function and containing the first and second values associated to indexes correlated to said first arguments and to an expected maximum limit.