Convolutional Decoder Traceback Using Alternating State Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current convolutional decoders in radio communication systems, such as those used in LTE systems, face challenges in achieving high reliability and efficient operation speed for decoding convolutional codes, particularly in correcting errors in data transmission over Physical Downlink Control Channel (PDCCH) using Tail Biting Convolutional Code (TBCC).
Innovation Solution
A convolutional decoder is designed with a branch metric processor, an Add-Compare-Select (ACS) processor, and trace back logic, utilizing first and second storage units to determine and update maximum likelihood path metrics and selectively trace back survival paths, enhancing the reliability and speed of decoding convolutional codes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a conventional Viterbi decoding algorithm is used with multiple training operations, then decoding reliability can be improved, but operation speed deteriorates due to the complex multi-step process
Solution Approach 1:
The patent pre-calculates and stores branch metrics for all possible state transitions in a branch metric table before the actual decoding process. This preliminary action eliminates the need for repeated metric calculations during multiple training operations, thereby maintaining decoding reliability while significantly improving operation speed.
Solution Approach 2:
The patent performs a single training operation followed by maximum likelihood path finding, rather than multiple training operations. By performing only the necessary partial action (one training operation) and relying on pre-computed branch metrics, the system achieves comparable reliability with reduced operational complexity and improved speed.
2Measurement precision
If multiple training operations are performed in the Viterbi decoding algorithm, then decoding accuracy is improved, but the complexity of the decoding process increases
Solution Approach 1:
The branch metric table is pre-computed and stored in advance, containing all necessary metric values for state transitions. This preliminary action eliminates the need for repeated metric calculations during multiple training operations, thereby maintaining decoding accuracy while significantly reducing the complexity of the decoding process.
Solution Approach 2:
The patent uses a pre-computed branch metric table that copies and stores all possible branch metric values in advance. During decoding, the system simply retrieves these pre-stored values rather than recalculating them, thereby maintaining accuracy while reducing computational complexity.
3Reliability
If branch metrics are recalculated during each training operation, then decoding reliability is maintained, but operation speed decreases due to repeated calculations
Solution Approach 1:
The patent pre-calculates all branch metrics and stores them in a branch metric table before the decoding process begins. This preliminary action eliminates the need for repeated calculations during each training operation, thereby maintaining decoding reliability while significantly reducing the time lost to redundant computations.
Solution Approach 2:
The branch metric table serves as a self-service resource that provides pre-computed metrics during the decoding process. Instead of requiring the system to recalculate metrics during each operation, the pre-computed table automatically provides the necessary values, thereby maintaining reliability while eliminating time-consuming recalculations.
Data Source
AI summary
A convolutional decoder includes a first storage, a second storage, a branch metric processor to determine branch metrics for transitions of states from a start step to a last step according to input bit streams, an ACS processor to select maximum likelihood path metrics to determine a survival path according to the branch metrics and to update states of the start step to the first storage and the second storage alternately based on the selection of the maximum likelihood path metrics, and a trace back logic to selectively trace back the survival path based on the states of the start step stored in a selected storage among the first storage and the second storage.