LDPC Decoding with Energy-Sum Feedback for Overflow Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In low density parity check (LDPC) decoding, high bit error rates occur due to severely damaged received messages, excessive iteration times, or inadequate bit width, leading to decoding failures and overflow issues.
Innovation Solution
A decoding method that monitors energy summations of belief values across iterations, stopping the decoding process early if the energy summation suddenly drops or exceeds thresholds, ensuring the decoding result is generated based on previous iterations to prevent bit errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If iteration decoding is executed with limited iteration times and bit width, then decoding speed is improved, but bit error rate increases due to overflow and inadequate precision
Solution Approach 1:
The patent calculates energy summation of belief values before completing all iterations to predict potential overflow in advance. By computing the energy summation metric during the decoding process, the system can proactively detect when belief values are approaching overflow conditions and adjust iteration termination accordingly, preventing bit errors before they occur while maintaining efficient decoding speed
Solution Approach 2:
The patent implements a feedback mechanism where the energy summation of belief values is continuously monitored during iteration decoding. This feedback information is used to dynamically control the decoding process - when the energy summation indicates approaching overflow thresholds, the system adjusts iteration termination to prevent overflow, thereby maintaining low bit error rates while avoiding unnecessary iterations that would waste time
2Reliability
If bit width is increased to prevent overflow, then reliability is improved, but device complexity and resource consumption increase
Solution Approach 1:
The patent changes the monitoring parameter from individual belief value magnitudes to the energy summation of all belief values. This parameter transformation allows the system to detect overflow conditions using a compact summation metric rather than tracking each belief value's bit width individually, thereby maintaining overflow prevention capability without requiring increased bit width in the underlying data structures
Solution Approach 2:
The patent introduces energy summation as an intermediary metric that mediates between the belief values and overflow detection. Instead of directly monitoring whether individual belief values exceed representable ranges (which would require increased bit width), the system uses the energy summation as an intermediate indicator to predict potential overflow conditions, allowing overflow prevention with standard bit width allocations
Data Source
AI summary
A decoding method for low density parity check (LDPC) and an electric device using the decoding method are provided. The decoding method includes: receiving a message and executing an iteration decoding to the message; obtaining first belief values of the message in an (i−1)th iteration; obtaining a first energy summation of the first belief values; obtaining second belief values of the message in an ith iteration; obtaining a second energy summation of the second belief values; determining whether the second energy summation is smaller than the first energy summation and whether a difference between the first energy summation and the second energy summation is larger than a first threshold; if yes, generating a decoding result according to the first belief values. Therefore, a bit error rate of the decoding is decreased.