Compression Context Learning via Secondary Link Rule Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data compression and decompression techniques for LPWAN networks require an initialization phase that reduces available throughput on communication links, limiting adaptability and impacting real-time data exchanges due to the need for predefined compression/decompression rules shared between devices.
Innovation Solution
A method that learns and updates compression/decompression rules using a secondary communication link, allowing selective modification of basic rules based on current data stream values, thereby minimizing throughput degradation on the primary communication link and enabling more frequent rule updates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If compression/decompression rules are predefined and shared between devices, then compression efficiency is improved, but adaptability to dynamic data flow changes deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-defining a set of basic compression/decompression rules that cover common data patterns. These rules are prepared in advance and stored in both compressor and decompressor devices, enabling immediate compression efficiency without requiring real-time negotiation or learning during data transmission.
Solution Approach 2:
The patent implements dynamics by allowing the compression context to be dynamically updated during operation. The compressor and decompressor can learn from actual data traffic patterns and adaptively modify which basic rules to apply or create new rules, enabling the system to evolve its compression strategy based on observed data flows while maintaining the foundation of pre-defined rules.
2Reliability
If an initialization phase is implemented to share rules between devices, then compression/decompression functionality is improved, but available throughput on the communication link deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-configuring basic compression rules in both devices before actual data transmission begins. This initialization happens once during device setup or connection establishment, rather than requiring continuous rule exchange during data transmission, thereby preserving throughput during the main communication phase.
Solution Approach 2:
The patent implements self-service by enabling the compressor and decompressor to autonomously learn and synchronize compression rules during normal operation without requiring a dedicated initialization phase. The devices independently analyze data patterns and update their rule sets, eliminating the need for separate rule-exchange communication that would consume throughput.
3Adaptability or versatility
If compression rules are updated frequently to adapt to data changes, then adaptability is improved, but the duration of learning and rule updates deteriorates
Solution Approach 1:
The patent applies partial action by updating only the necessary portions of the compression context rather than performing complete rule re-learning. When data patterns change, the system selectively modifies specific basic rules or adds new rules for observed patterns, rather than re-establishing the entire compression framework, thereby reducing learning time while maintaining adaptability.
Solution Approach 2:
The patent implements discarding and recovering by allowing the system to discard outdated or ineffective compression rules and recover by learning new rules from observed data patterns. This selective replacement approach enables frequent adaptation without the overhead of complete rule set re-learning, as the system maintains a library of basic rules and only updates those relevant to current data characteristics.
Data Source
Figure 1~3
AI summary
The invention concerns a method for learning from a compression/decompression context comprising at least one compression/decompression rule intended to be used by a compression device and a decompression device, to respectively compress and decompress a first flow of data transmitted via a first communication link. According to such a process, at least one device, among the compression device, the decompression device and a third device, carries out the following steps: • analysis (E200) of at least one second flow of non-compressed data, transmitted between the compression device and the decompression device via at least one second communication link; and • learning (E210) of the at least one compression/decompression rule, depending on the analysis.