Randomizing Crosstalk Probing Signals for DSL Vectoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for mitigating crosstalk in Multiple Input Multiple Output (MIMO) wired communication systems, such as DSL, face challenges in accurately estimating crosstalk channels due to non-linear effects like demapping errors and signal clipping, leading to biased precoder and postcoder coefficients and performance degradation over time.
Innovation Solution
A vectoring controller that iterates through successive crosstalk acquisition cycles, randomizing crosstalk probing symbols by rotating or scaling them, and iteratively configures the vectoring processor to mitigate these biases, ensuring optimal precoder or postcoder coefficients are achieved.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deterministic pilot sequences are used for crosstalk estimation, then the estimation process is simple and reproducible, but non-linear effects cause biased coefficients and performance degradation over time
Solution Approach 1:
The patent applies dynamics by making the pilot sequences dynamic rather than static. Specifically, it introduces time-varying randomization where the pilot sequence changes over time according to a random process. This dynamic approach prevents the systematic bias that occurs with deterministic sequences under non-linear effects, while maintaining estimation accuracy through proper statistical properties of the randomized sequences.
Solution Approach 2:
The patent changes the parameter of the pilot sequence from deterministic to randomized. By introducing randomness in the pilot sequence values (while maintaining orthogonality and other critical properties), the system transforms the estimation process to be robust against non-linear distortions. The randomization parameter is controlled to ensure sequences remain usable for correlation-based estimation while avoiding bias accumulation.
2Measurement precision
If pilot sequences are randomized to reduce non-linear effects, then unbiased coefficients are achieved, but sequence design becomes more complex
Solution Approach 1:
The patent employs periodic action by using pseudo-random sequences with periodic structures that maintain orthogonality. The randomized pilot sequences are designed to repeat certain statistical properties periodically, ensuring that correlation-based estimation remains effective. This periodic structure simplifies the randomization process compared to completely aperiodic random sequences.
Solution Approach 2:
The patent uses copying by generating randomized pilot sequences that replicate the essential properties of deterministic orthogonal sequences (such as orthogonality and correlation properties) while introducing randomization. Instead of designing entirely new sequences, the system copies the structure of known good sequences and applies randomization transformations that preserve critical estimation properties while eliminating bias.
Data Source
AI summary
In accordance with an embodiment, the vectoring controller is configured to iterate through successive crosstalk acquisition cycles and, within respective ones of the crosstalk acquisition cycles, to configure sequences of crosstalk probing symbols for transmission over the respective communication lines, to receive sequences of error samples as successively measured by respective receivers coupled to the respective communication lines while the sequences of crosstalk probing symbols are being transmitted, and to determine crosstalk estimates between the respective communication lines based on the sequences of error samples. The vectoring controller is further configured to randomize the successive sequences of crosstalk probing symbols used during the successive crosstalk acquisition cycles, and to iteratively configure the vectoring processor based on the successive crosstalk estimates.