Network Encoding Grouping to Reduce RLNC Data Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In random linear network coding (RLNC), linearly correlated encoded data can affect receiving reliability, necessitating a reduction in linear correlation to improve data reception.
Innovation Solution
The method involves dividing data units into groups for separate encoding, using encoding matrices or kernels, and employing oblique diagonals in encoding matrices to reduce correlation, thereby enhancing the data's resistance to channel fading and changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If random linear network coding is used to encode data units, then encoding efficiency is improved, but linear correlation of encoded data increases, worsening receiving reliability
Solution Approach 1:
The patent divides T to-be-encoded data units into L groups, where each group contains at most M data units. Each group is separately encoded to produce encoded data units. This segmentation reduces linear correlation between encoded data units while maintaining encoding efficiency, as the grouping structure limits the propagation of correlation across all data units.
Solution Approach 2:
The patent applies different encoding treatments to different groups of data units. Specifically, encoded data units corresponding to the first group are treated differently from those corresponding to the second group, with different types of redundancy information added. This local differentiation reduces overall linear correlation while preserving the benefits of network coding.
2Ease of operation
If all T data units are encoded together as a single group, then encoding simplicity is maintained, but correlation of encoded data increases, reducing resistance to channel fading
Solution Approach 1:
Instead of encoding all T data units together, the patent segments them into L groups with at most M units each. This segmentation maintains relative encoding simplicity while significantly reducing linear correlation and improving resistance to channel fading, as failures in one group do not propagate to all encoded data units.
Solution Approach 2:
The patent introduces parameters T (total data units), L (number of groups), and M (max units per group) to control the encoding process. By adjusting these parameters, the system optimizes the balance between encoding simplicity and channel fading resistance, allowing flexible adaptation to different channel conditions.
3Speed
If encoded data units are generated with high correlation, then encoding speed is maintained, but spectral efficiency decreases due to reduced receiving reliability
Solution Approach 1:
The patent segments T data units into L groups processed in parallel, maintaining encoding speed while reducing correlation. Each group is encoded independently or with limited interaction, preserving fast encoding performance while improving spectral efficiency through reduced linear correlation of the output encoded data units.
Solution Approach 2:
The patent applies partial grouping rather than complete segmentation, with each group containing at most M units out of T total units. This partial action approach maintains much of the encoding speed advantage while achieving sufficient correlation reduction to improve spectral efficiency, avoiding the overhead of complete individual encoding of each data unit.
Data Source
AI summary
Embodiments of this application provide an encoding method and an apparatus for network encoding. A plurality of to-be-encoded data units are divided into a plurality of groups for separate encoding, and the quantity of to-be-encoded data units included in each group is less than the quantity of all to-be-encoded data units. According to the method, correlation of the encoded data is reduced, and a capability of the encoded data against channel fading and channel change is improved, to increase the receiving reliability of the encoded data.