Multivariate Correlated Data Compression via Common Information Vectors

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current telecommunication and information systems face challenges in efficiently encoding and transmitting correlated data vectors over multiple channels, particularly in noisy environments, where existing methods fail to optimize resource usage and fidelity effectively.

Innovation Solution

A method for compressively encoding two correlated data vectors using a pre-encoder unit that transforms the data into canonical form, followed by an encoder unit that computes common and private information vectors, and compresses them using lattice coding and joint entropy techniques, ensuring efficient transmission over multiple channels.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional source encoding is used to compress correlated data vectors, then compression efficiency is improved, but transmission fidelity deteriorates in noisy multi-channel environments

Engineering Contradiction:
Improvecompression efficiencyVSAvoidtransmission fidelity
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent segments the encoding process into two distinct parts: source encoding that exploits correlation between data vectors to achieve compression, and channel encoding that adds redundancy to protect against noise. This segmentation allows each encoding type to optimize for its specific function without compromising the other.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements nested encoding where channel encoding is applied within the framework of source encoding. The source encoder first compresses the correlated data vectors, then the channel encoder adds protective redundancy to the compressed output. This nested structure enables simultaneous achievement of compression efficiency and transmission reliability.

Inventive Principle:
Principle #7Nested doll (Nesting)

2Reliability

If channel encoding is applied to protect against noise, then transmission reliability is improved, but resource efficiency deteriorates due to added redundancy

Engineering Contradiction:
Improvetransmission reliabilityVSAvoidresource efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent applies preliminary source encoding to compress the data vectors before channel encoding is applied. By performing the compression action first, the subsequent channel encoding operates on a smaller, more efficient data structure, thereby reducing the overall resource consumption while maintaining the protective redundancy needed for reliable transmission.

Inventive Principle:
Principle #10Preliminary action

3Productivity

If lossy coding is used to increase channel capacity utilization, then resource efficiency is improved, but information loss increases

Engineering Contradiction:
Improvechannel capacity utilizationVSAvoiddata fidelity
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The patent incorporates feedback mechanisms where the encoder and decoder use knowledge of the correlation structure and noise characteristics to optimize the lossy compression parameters. This feedback allows the system to achieve high channel capacity utilization while maintaining data fidelity within acceptable distortion bounds by continuously adapting the compression level based on channel conditions.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3562043B1Methods for compression of multivariate correlated data for multi-channel communication
Publication Date: 2023.06.07 VAN SCHUPPEN PATENTS BV
  • EP3562043B1 patent drawingFigure 1
  • EP3562043B1 patent drawingFigure 2
  • EP3562043B1 patent drawing

AI summary

Methods are provided for efficiently encoding and decoding multivariate correlated data sequences for transmission over multiple channels of a network. The methods include transforming data vectors from correlated sources into vectors that comprise substantially independent and correlated components, and generating a common information vector based on the correlated components, and two private information vectors. The methods also include computing the amount of information, such as Wyner's lossy common information, in the common information vector, computing rates that lie on the Gray-Wyner rate region, and choosing compression rates based on the amount of common information. The methods may be applicable, in general, to a wide range of communications and/or storage systems and, particularly, to sensor networks and multi-user virtual environments for gaming and other applications.