Polar SCL Decoder Special-Node Processing for Lower Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional successive cancellation (SC) and simplified successive cancellation list (SSCL) decoding methods for polar codes face challenges in reducing decoding latency while maintaining performance, particularly due to inadequate handling of special nodes and lack of well-established logic for path splitting in SSCL methods.
Innovation Solution
An apparatus and method for constituent code processing in polar successive cancellation list (SCL) decoding that determines an activation value and number of candidate paths based on node reliability, information nodes, and leaf nodes, selecting the most probable paths by calculating path metrics and flipping combinations of least reliable bits to optimize latency and hardware complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If conventional SC or SCL decoding methods are used, then decoding can be performed with standard processing, but long latency limits practical applications
Solution Approach 1:
The patent segments the decoding process by identifying and specially processing only certain intermediate nodes (special nodes) that correspond to specific constituent polar codes (Rate-0, Rate-1, Rate-R, Rate-S codes), while other nodes are processed using conventional methods. This selective segmentation enables latency reduction without requiring complete redesign of the entire decoding architecture.
Solution Approach 2:
The patent applies preliminary action by pre-determining which intermediate nodes are special nodes based on their constituent code types, and preparing specific processing rules for these nodes before the actual decoding process. This allows the decoder to quickly identify and apply optimized processing paths for special nodes during decoding, reducing overall latency.
2Productivity
If simplified SSCL decoding is applied to reduce latency, then processing speed improves, but there is no well-established logic for determining path splitting amount and differentiation within special node types
Solution Approach 1:
The patent applies local quality by differentiating processing methods based on the specific type of special node encountered. Each constituent code type (Rate-0, Rate-1, Rate-R, Rate-S) has its own characteristic processing rules for path splitting and metric updating. This localized differentiation captures the unique polarization features of each node type while maintaining overall system efficiency.
Solution Approach 2:
The patent changes parameters dynamically by determining the activation value and number of candidate paths as a function of node reliability, number of information nodes, and number of leaf nodes. This parameter adaptation allows the decoder to optimize its behavior for each specific node type and reliability condition, achieving better performance without excessive complexity.
3Loss of time
If aggressive path splitting and metric updating are applied in special nodes, then latency reduction increases, but hardware complexity and potential performance degradation occur
Solution Approach 1:
The patent applies partial action by selectively applying simplified SSCL decoding only to special nodes rather than all nodes in the decoding tree. This partial application achieves latency reduction while avoiding the hardware complexity that would result from applying the same aggressive optimization uniformly across the entire decoder. The activation value I controls whether simplified processing is applied (I=1) or conventional processing is used (I=0).
Data Source
AI summary
An apparatus for constituent code processing in polar successive cancellation list (SCL) decoding and a method thereof. The apparatus includes a processor configured to determine an activation value I and a number r of the candidate paths, where I is a binary value and r is an integer, (I, r)=ƒ(R, k, m), ƒ is a function, R is a number indicating node reliability, k is an integer indicating a number of information nodes, and m is an integer indicating a number of leaf nodes; determine min1, min2, . . . , minq, wherein q is a number of least reliable bits; determine r candidate paths; determine path metrics PMt<sub2>j </sub2>of a codeword j for each candidate path t; and select r most probable paths based on PMt<sub2>j</sub2>.