Pattern Learning System for Real-Time Causal Event Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In intelligence, surveillance, and reconnaissance (ISR) activities, existing systems face challenges in processing high volumes of event messages from multiple sensors in real time to determine causal relationships, leading to difficulties in filtering out extraneous information and identifying useful causal events.
Innovation Solution
A pattern learning system that receives event messages sequentially, forms sub-sequences with trigger, consequence, and intermediary messages, generates graphs to represent these sequences, combines them, and determines causal sequences using the heaviest weighted directed path from trigger to consequence messages, effectively filtering out non-causal event messages.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If event messages from multiple sensors are processed in real time, then causal relationships can be determined timely, but the system cannot handle the high volume of event messages when event rates exceed processing capabilities
Solution Approach 1:
The patent segments the high-volume event message stream into smaller sub-sequences grouped by sensor identity. Each sub-sequence contains a manageable number of events from a single sensor, allowing the system to process multiple sensor streams in parallel without being overwhelmed by the total volume. This segmentation enables real-time processing by dividing the intractable whole into tractable parts.
Solution Approach 2:
The patent introduces a new dimension of organization by grouping events according to sensor identity rather than processing them as a single flat stream. This dimensional reorganization transforms the problem from processing N events sequentially to processing M sensor streams simultaneously, where each stream contains N/M events, effectively increasing processing throughput.
2Loss of information
If all event messages are processed to determine causal relationships, then complete information is analyzed, but extraneous non-causal information cannot be filtered out efficiently
Solution Approach 1:
The patent extracts and removes extraneous non-causal event messages from the event streams through automated filtering. By identifying and eliminating events that do not contribute to causal relationships, the system reduces the volume of data requiring detailed analysis while preserving all relevant causal information, thus simplifying processing without losing critical data.
Solution Approach 2:
The system employs feedback mechanisms where processed event data informs the filtering criteria for subsequent events. As the system learns from analyzed causal patterns, it refines its ability to distinguish causal from non-causal events, improving filtering efficiency over time while maintaining information completeness.
Data Source
AI summary
According to one embodiment, a pattern learning system includes a pattern learning tool that receives event messages in a sequential manner from multiple sensors and forms multiple sub-sequences that each includes a trigger event message, a consequence event message, and one or more intermediary event messages. The pattern learning tool then generates multiple graphs that each represents a sub-sequence of the plurality of event messages. The pattern learning tool then combines the graphs into a combined graph according to a type of each event message, and determines a causal sequence from the combined graph according to a heaviest weighted directed path from the trigger event message to the consequence event message.