Equalizer Training Sequence Shaping for Distorted Communication Channels
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing equalizer training technologies face challenges with severe distortion in channels, leading to low convergence speed, long link-setup times, and high transmission overhead due to intersymbol interference.
Innovation Solution
Implement spectral shaping and remapping at the transmitter to generate an intermediate sequence that counters channel response, reducing signal levels and transmitting this sequence to train the receiver's equalizer, or estimate and transmit a filter parameter from the receiver to the transmitter for spectral shaping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If traditional equalizer training is used in channels with severe distortion, then the equalizer can be trained, but the convergence speed becomes extremely low requiring many iterations
Solution Approach 1:
The patent applies preliminary action by pre-processing the training sequence at the transmitter before it enters the channel. The transmitter generates a first training sequence, applies spectral shaping to create a second sequence with modified frequency content, and remaps amplitude levels to create a third sequence. This pre-processing counteracts the expected channel distortion before the equalizer training begins, allowing the receiver's equalizer to converge much faster without requiring numerous iterations to compensate for severe distortion.
2Reliability
If more iterations are used to optimize convergence in severe distortion channels, then equalizer training can succeed, but the transmission overhead increases
Solution Approach 1:
By performing spectral shaping and remapping operations at the transmitter before transmission, the patent prepares the training sequence in advance to be optimally suited for the channel conditions. This preliminary processing ensures that the receiver can achieve reliable equalizer training with fewer iterations, thereby reducing the total transmission overhead while maintaining training success rate.
Solution Approach 2:
The patent changes parameters of the training sequence by applying spectral shaping that modifies the frequency spectrum and remapping that changes amplitude levels. These parameter transformations create a training sequence with characteristics that are more favorable for equalizer convergence in distorted channels, reducing the number of iterations needed and thus lowering transmission overhead.
3Speed
If spectral shaping is applied to counter channel response, then equalizer training convergence is accelerated, but additional processing complexity is introduced at the transmitter
Solution Approach 1:
The patent implements spectral shaping by modifying parameters of the training sequence, specifically the frequency spectrum and amplitude levels. This involves applying a spectral shaping filter and performing remapping of amplitude levels, which are standard signal processing operations that can be implemented with existing DSP resources at the transmitter, adding processing complexity only to the extent of these well-established techniques.
Data Source
AI summary
Examples of the present disclosure relate to a method, device and apparatus for communication, and a computer-readable medium. An example of the method includes: conducting, based on a channel response between a transmitter and a receiver, spectral shaping on a first sequence at the transmitter, so as to obtain an intermediate sequence, where the spectral shaping at least partially counters the channel response; remapping the intermediate sequence, so as to obtain a second sequence, where the second sequence has less signal levels than the intermediate sequence; and transmitting the second sequence to the receiver, so as to train an equalizer of the receiver. In this way, the method can accelerate training of the equalizer without sacrificing performance of the equalizer or introducing any additional hardware cost.