Adaptive Nonlinear Equalization Using Temporary Decision Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical disk reproduction systems face challenges in adaptively reducing both linear and nonlinear distortion as recording density increases, particularly due to the need for training signals to adjust neural network coefficients, which limits their ability to track waveform variations.
Innovation Solution
An equalizer system comprising a linear equalizer unit, a temporary decision unit, and a nonlinear equalizer unit that derives coefficients using temporary decisions as teacher signals, allowing for adaptive nonlinear distortion reduction without training signals, and includes a delay mechanism to align signal timings and converge coefficients.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a training signal is recorded on an optical disk to update neural network coefficients, then the nonlinear distortion reduction performance is improved, but the use efficiency of the optical disk is reduced
Solution Approach 1:
The system uses the reproduction signal itself (after temporary decision) as the teacher signal for coefficient updates, eliminating the need for separate training signals. The equalizer adapts autonomously using the available signal, making the system self-sufficient and improving disk usage efficiency.
Solution Approach 2:
The reproduction signal serves dual purposes: it is both the signal to be equalized and the teacher signal for coefficient updates. This multi-functionality eliminates the need for dedicated training signals, improving disk utilization while maintaining equalization performance.
2Device complexity
If the neural network coefficients are fixed after learning, then the system complexity is reduced, but the ability to track variation in reproduction waveform characteristics is reduced
Solution Approach 1:
The coefficient update mechanism transitions from static (fixed after learning) to dynamic (continuous adaptation). The system continuously updates coefficients using the reproduction signal as teacher signal, enabling real-time tracking of waveform variations while maintaining manageable complexity through automated adaptation.
Solution Approach 2:
The system implements feedback by using the temporary decision result of the reproduction signal as the teacher signal for coefficient updates. This closed-loop mechanism enables continuous adaptation to changing waveform characteristics, improving tracking ability without requiring complex external control systems.
3Measurement precision
If a training signal is used to initialize coefficients, then the initial equalization performance is improved, but the system cannot adapt to power variations in recorders
Solution Approach 1:
The coefficient update process continues throughout operation using the reproduction signal as teacher signal, rather than being limited to an initial training phase. This continuous adaptation ensures the system remains effective despite power variations in recorders, maintaining equalization performance across different operating conditions.
Data Source
AI summary
A linear equalizer unit sequentially subjects a signal to be processed to linear equalization. A temporary decision unit sequentially subjects a signal subjected to linear equalization by the linear equalizer unit to temporary decision. A nonlinear equalizer unit derives a plurality of coefficients using a signal subjected to temporary decision as a teacher signal and sequentially subject a signal subjected to linear equalization by the linear equalizer unit to nonlinear equalization based on the plurality of coefficients.