Logical Symbol Indexing for Multi-Sensor Pattern Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for searching logical patterns in multi-sensor time-series data from industrial environments face challenges such as variable event durations, misalignment of patterns across sensors, and the need for manual annotation, leading to inefficiencies and inaccuracies in detecting events of interest.
Innovation Solution
A system and method that preprocesses time-series data into logical symbols, clusters, and indexes them using a TRIE-based structure, allowing for efficient retrieval of relevant patterns through a search query, ranking, and retrieval of the most relevant patterns based on relevance, addressing the challenges of variable durations and sensor misalignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional near-exact waveform matching techniques are used, then the search process is simple, but many instances of events are missed due to variable durations and sensor misalignment
Solution Approach 1:
The patent segments the time-series data into discrete logical symbols representing different sensor states and transitions. This segmentation allows the system to handle variable duration events by matching sequences of logical symbols rather than requiring exact waveform alignment, thereby improving detection accuracy while maintaining search efficiency through structured pattern matching.
Solution Approach 2:
The patent transforms continuous sensor data into discrete logical symbols by changing the parameter representation from continuous values to categorical states. This parameter transformation enables the system to accommodate variations in event duration and sensor timing by matching logical patterns rather than exact temporal waveforms, resolving the contradiction between accuracy and efficiency.
2Measurement precision
If manual annotation of events is performed to capture domain knowledge, then the detection accuracy improves, but the effort and time required increases significantly
Solution Approach 1:
The patent enables the system to automatically learn and represent domain knowledge through the logical pattern matching framework without requiring manual annotation of each event. The system self-services by transforming sensor data into logical symbols that inherently capture event patterns, eliminating the need for time-consuming manual annotation while maintaining detection accuracy.
Solution Approach 2:
The patent creates a simplified copy of the physical sensor data in the form of logical symbols that preserve the essential event patterns. This symbolic representation copies the meaningful structure of events without requiring manual interpretation, allowing automated detection while reducing the time investment needed compared to direct manual annotation of raw sensor data.
3Ease of manufacture
If hard-coded rules are used to detect events, then the implementation is straightforward, but the rules become ineffective and difficult to implement due to variable event characteristics
Solution Approach 1:
The patent introduces dynamics by allowing the pattern matching system to adapt to variable event characteristics through flexible logical symbol sequences. Rather than relying on fixed hard-coded thresholds and time windows, the system dynamically matches sequences of logical symbols that can accommodate variations in event duration, intensity, and timing across different sensors, thereby improving detection reliability while maintaining implementation simplicity.
Solution Approach 2:
The patent changes the parameters from fixed hard-coded values to flexible logical symbol sequences. This transformation allows the system to represent variable event characteristics using sequences of symbolic states rather than rigid numerical thresholds, improving detection reliability while keeping the implementation approachable through structured pattern definitions.
4Adaptability or versatility
If machine learning techniques are used for event detection, then the system can handle variable patterns, but the computational intensity and complexity increase
Solution Approach 1:
The patent uses lightweight logical symbols as disposable representations of sensor states that can be quickly generated and discarded during pattern matching. These simple symbolic tokens provide the necessary adaptability to handle variable event patterns without requiring complex machine learning models, thereby reducing system complexity while maintaining pattern matching flexibility.
Solution Approach 2:
The patent substitutes complex machine learning mechanisms with a simpler logical pattern matching system. By replacing heavy computational models with structured symbolic sequence matching, the system achieves adaptability to variable patterns while significantly reducing computational intensity and system complexity.
5Productivity
If fixed-length time-series windows are used for pattern detection, then the processing is simplified, but events of variable length cannot be detected accurately
Solution Approach 1:
The patent introduces dynamics by allowing the logical symbol sequence length to vary according to the actual event duration. Instead of forcing all events into fixed-length windows, the system dynamically determines the sequence length based on the number of logical symbols that constitute the event pattern, enabling accurate detection of variable-length events while maintaining processing efficiency through structured symbolic representation.
Solution Approach 2:
The patent changes the window length parameter from fixed to variable by using sequences of logical symbols whose length naturally adapts to the event being detected. This parameter transformation allows the system to process events of any duration using appropriately sized symbolic sequences, achieving both processing efficiency and detection accuracy.
Data Source
AI summary
Methods and systems for searching logical patterns in voluminous multi sensor data from the industrial internet is provided. The method retrieves instances of patterns in time-series data where patterns are specified logically, using a sequence of symbols. The logical symbols used are a subset of the qualitative abstractions specifically, the concepts of steady, increasing, decreasing. Patterns can include symbol-sequences for multiple sensors, approximate duration as well as slope values for each symbol. To facilitate efficient querying, each sensor time-series is pre-processed into a sequence of logical symbols. Each position in the resulting compressed sequence is registered across a TRIE-based index structure corresponding to the multiple logical patterns it may belong to. Logical multi-sensor patterns are efficiently retrieved and ranked using such a structure. This method of indexing and searching provides an efficient mechanism for exploratory analysis of voluminous multi-sensor data.