Pattern-Dependent Adaptive Targets for Noise Predictive Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data processing circuits with noise predictive filtering face degraded performance due to the use of less-than-ideal targets in pattern-dependent filters, leading to suboptimal operation.
Innovation Solution
Implementing a data processing circuit with multiple pattern-dependent adaptive target circuits that calculate adaptive targets based on noise predictive outputs and a training sequence, allowing for independent adjustment of target values for each filter, thereby improving filtering accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a common target is used by all pattern dependent filters, then the device complexity is reduced, but the manufacturing precision and filtering accuracy are degraded
Solution Approach 1:
The patent divides the single common target into multiple pattern-specific adaptive targets (first adaptive target, second adaptive target, etc.), with each target dedicated to a specific pattern dependent filter. This segmentation allows each filter to use an optimized target value tailored to its specific pattern, thereby improving filtering accuracy while accepting increased device complexity through multiple adaptive target circuits.
Solution Approach 2:
The patent implements local quality by providing different adaptive target values to different pattern dependent filters based on their specific patterns. Each adaptive target circuit calculates and maintains target values optimized for its associated filter's specific pattern conditions, rather than using a uniform common target. This localized optimization improves overall system precision despite increased complexity.
2Ease of operation
If a single common target is used for all filters, then the ease of operation is improved, but the adaptability to different patterns is reduced
Solution Approach 1:
The patent transforms the static common target into dynamic pattern-specific adaptive targets that automatically adjust based on the detected pattern. Each adaptive target circuit dynamically calculates and updates its target value based on the specific pattern being processed, enabling the system to adapt to different patterns automatically without manual intervention, thus maintaining ease of operation while significantly improving adaptability.
Solution Approach 2:
The patent implements feedback mechanisms where each adaptive target circuit uses the output from its associated pattern dependent filter and the detected pattern information to continuously adjust and optimize its target value. This feedback loop enables automatic adaptation to different patterns while keeping the operation transparent and simple for the user.
Data Source
AI summary
Various embodiments of the present invention provide systems and methods for data processing. As an example, a data processing circuit is disclosed that includes a noise predictive filter circuit, a data detector circuit, and a first and a second pattern dependent adaptive target circuits. The noise predictive filter circuit includes at least a first pattern dependent filter circuit operable to perform noise predictive filtering on a data input for a first pattern using a first adaptive target to yield a first noise predictive output, and a second pattern dependent filter circuit operable to perform noise predictive filtering on the data input for a second pattern using a second adaptive target to yield a second noise predictive output. The data detector circuit is operable to apply a data detection algorithm to the first noise predictive output and the second noise predictive output to yield a detected output. The first pattern dependent adaptive target circuit is operable to adaptively calculate the first adaptive target based at least in part on the first noise predictive output and a training sequence. The second pattern dependent adaptive target circuit operable to adaptively calculate the second adaptive target based at least in part on the second noise predictive output and the training sequence.