Layered LDPC Decoding With Rotating Metric Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The implementation of layered decoding algorithms for LDPC codes in hardware is complex due to the need to update channel cues with multiple metrics cues in each column, leading to increased routing constraints and potential deterioration in decoding performance.
Innovation Solution
A method and apparatus that simplify the hardware implementation by updating the channel cue associated with a particular column using only one selected metric cue at each iteration, with the row of the selected metric cue changing at each iteration, and using a simple indexing method such as k modulo Nm to determine the row index, where k is the iteration index and Nm is the number of metrics cues in the column.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple metrics cues are used to update channel cues in each column during layered decoding, then decoding performance is improved, but hardware complexity and routing constraints increase
Solution Approach 1:
The parity matrix H is divided into multiple layers, where each layer contains a subset of rows. During decoding, metrics cues are updated layer by layer rather than all at once. This segmentation allows the decoder to process multiple metrics cues systematically across different layers, improving decoding performance while managing hardware complexity through structured, incremental updates rather than simultaneous complex operations
Solution Approach 2:
The metrics matrix is pre-organized into layers with predetermined row groupings before decoding begins. This preliminary structuring enables the decoder to efficiently retrieve and update the appropriate metrics cues for each layer in sequence, reducing the real-time routing complexity and hardware requirements while maintaining the ability to process multiple metrics cues for accurate decoding
2Measurement precision
If multiple metrics cues are updated simultaneously in each column, then decoding accuracy is improved, but routing constraints and hardware implementation difficulty increase
Solution Approach 1:
The decoding process alternates between updating metrics cues for different layers in a periodic manner. Instead of updating all metrics cues simultaneously, the decoder cycles through layers sequentially, updating metrics cues for one layer at a time. This periodic action maintains decoding accuracy by ensuring all necessary metrics are updated, while significantly simplifying hardware implementation by reducing simultaneous routing requirements to manageable levels
3Reliability
If all metrics cues in a column are used for updating channel cues, then bit error rate performance is improved, but the complexity of managing metric updates increases
Solution Approach 1:
The set of metrics cues in each column is segmented by assigning them to different layers based on their row positions in the parity matrix. This segmentation allows the decoder to manage updates by layer rather than by column, reducing the complexity of tracking and updating multiple metrics cues while ensuring all are properly utilized for achieving low bit error rates
Solution Approach 2:
The assignment of metrics cues to layers is performed in advance based on the predetermined structure of the parity matrix. This preliminary organization creates a clear mapping between metrics cues and layers, simplifying the update management process during decoding by eliminating the need for complex real-time decisions about which metrics to update, while still utilizing all metrics for optimal bit error rate performance
Data Source
AI summary
The metrics matrix may include at least one particular layer including at least one particular column having several metrics cues, respectively, situated in different rows. For the particular layer, the updating of the channel cue is associated with the particular column involving at each iteration one updated metric cue selected from all the metrics cues of the particular column. The row of the selected metric cues may change at each iteration.