Iterative Decoder Early Termination for Lower Power Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Iterative decoders in communication systems, such as LDPC and turbo codes, consume excessive power due to the fixed number of iterations, even though fewer iterations are often sufficient for decoding, leading to inefficient energy usage.
Innovation Solution
An apparatus and method that includes a memory device to store bit number differences between iterations and an iteration termination device to compare these differences to thresholds, incrementing a confidence value and terminating iterations when the value exceeds a confidence threshold, thereby reducing power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a fixed maximum number of iterations is used in iterative decoders, then decoding performance is guaranteed, but power consumption increases unnecessarily
Solution Approach 1:
The patent applies dynamics by transitioning from a fixed iteration count to a dynamic, adaptive iteration termination mechanism. The decoder now adjusts the number of iterations based on real-time confidence metrics and bit difference calculations, allowing the system to terminate early when convergence is detected while maintaining performance guarantees when needed.
Solution Approach 2:
The patent implements feedback through confidence value tracking and bit difference comparison between iterations. The system continuously monitors decoding progress, compares current iteration results with previous iterations, and uses this feedback to dynamically determine when to terminate, thereby optimizing power consumption while maintaining reliability.
2Use of energy by moving object
If the number of iterations is reduced to save power, then power consumption decreases, but decoding reliability may be compromised
Solution Approach 1:
The system uses feedback mechanisms to monitor decoding confidence and bit differences between iterations. By continuously assessing whether the decoded bits are stabilizing (low bit difference) and whether confidence thresholds are met, the system can safely terminate early without compromising reliability, or continue iterating when convergence hasn't been achieved.
Solution Approach 2:
The patent makes the iteration process dynamic by adjusting the termination point based on actual decoding progress rather than using a fixed count. This allows the system to adapt the number of iterations to the specific characteristics of each received signal, reducing iterations when possible while maintaining performance when necessary.
3Use of energy by moving object
If adaptive iteration termination is implemented, then power consumption is reduced, but device complexity increases
Solution Approach 1:
The patent segments the iteration termination decision into distinct functional components: confidence value calculation, bit difference computation, threshold comparison, and termination control. This modular segmentation allows the complex adaptive termination logic to be implemented as separate, manageable modules rather than a monolithic complex system.
Data Source
AI summary
There are provided a method, an apparatus and a computer program product for reducing power consumption in an iterative decoder. The apparatus includes a memory device and an iteration termination device. The memory device is for storing a bit number difference indicating a number of bits that are different between a decoded codeword for a current iteration and a decoded codeword for a previous iteration, for each iteration of the iterative decoder prior to a maximum number of iterations. The iteration termination device is for comparing the bit number difference to a pre-specified bit number difference threshold value, incrementing a confidence value when the bit number difference exceeds the pre-specified bit number difference threshold value, and terminating further iterations of the iterative decoder when the confidence value exceeds a pre-specified confidence threshold value.