Team Polar Decoder Architecture for High-Throughput FER Improvement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current polar decoders face challenges in achieving high-throughput Terahertz wireless communication requirements while maintaining energy and area efficiency, as existing Successive Cancellation (SC) decoders struggle to match the frame error rate (FER) performance of Successive Cancellation List (SCL) or CRC-aided SCL decoders due to increased complexity with list size.
Innovation Solution
A team of autonomous Randomized Successive Cancellation (RSC) decoders operates with derandomized decision-making, ensuring deterministic and efficient operation, achieving superior FER performance comparable to SC decoders in terms of complexity and throughput.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If Successive Cancellation List (SCL) decoder is used to improve FER performance, then FER performance is significantly improved, but device complexity increases rapidly with list size
Solution Approach 1:
The patent divides the decoding task into multiple independent Randomized Successive Cancellation (RSC) decoders operating in parallel, each handling a segment of the decoding process. This segmentation allows the system to achieve SCL-like FER performance through diversity combining while keeping individual decoder complexity low, similar to SC decoders.
Solution Approach 2:
The patent introduces a team decision maker (TDM) as an intermediary that collects decisions from multiple RSC decoders and combines them to produce the final decoded output. This TDM enables the system to leverage diverse decoding paths without requiring each individual decoder to maintain high complexity.
2Reliability
If list size is increased to achieve near state-of-the-art FER performance, then FER performance becomes competitive with other FEC schemes, but complexity increases rapidly rendering decoders impractical for Tb/s throughputs
Solution Approach 1:
The patent segments the decoding function into multiple parallel RSC decoders, each operating independently at SC decoder complexity levels. This allows the system to achieve high FER performance through ensemble diversity while maintaining the high throughput capability of simple SC decoders, avoiding the complexity bottleneck of traditional SCL approaches.
Solution Approach 2:
The patent changes the operational parameters of the decoding system by using multiple parallel decoders with randomized decision-making rather than a single complex SCL decoder. This parameter change enables the system to achieve competitive FER performance while maintaining Tb/s throughput capability through the simplicity and speed of individual RSC decoders.
3Reliability
If multiple SC decoders are used with different decoding paths to improve FER performance, then diversity of decisions is produced, but coordination overhead reduces throughput
Solution Approach 1:
Each RSC decoder in the patent operates autonomously and independently, making its own decoding decisions without requiring coordination or communication with other decoders during the decoding process. This self-service operation eliminates coordination overhead and preserves high throughput, while the team decision maker combines the independent decisions afterward to achieve improved FER performance.
Data Source
AI summary
A team polar decoder (TPD) includes polar decoders (PPDs) connected to a channel, and a team decision maker (TDM) connected to the PPDs and a destination. Component polar decoders (CPDs) decode a polar code in accordance with a polar code. Each CPD receives a noisy code block (NCB) from the channel, and decodes the NCB in consecutive steps to obtain a decoded transform input block (DTIB). Each CPD is generates, at an end of the decoding step, a candidate decoded data block from the DTIB by a data-demapping operation that is an inverse of a data-mapping operation applied at a polar encoder, then sends the CDDB to the TDM, which receives the CDDBs from the PPDs, generates a decoded data block (DDB), and sends the DDB to the destination.