Dynamic Time Warping for Multisensor Trace Data Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current machine-learning technologies face challenges in training self-learning applications using trace data from sensors due to difficulties in aligning time series with different sampling rates, start times, and segmentation methods, limiting the incorporation of diverse data streams into corpora for methods like SVM or decision-tree procedures.
Innovation Solution
A dynamic time-warping procedure is employed to align and segment time series from multiple sensors, selecting a standard series for each attribute and aligning non-standard series to it, allowing for the creation of a corpus suitable for machine-learning modules.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a sliding-window method is used to divide time series into segments, then the segmentation process is simple, but the method is only useful when the two time series are very similar and cannot handle different sampling rates or variable lengths
Solution Approach 1:
The patent applies dynamic time warping to create a dynamic segmentation approach that adapts to the specific characteristics of each time series pair. Instead of using fixed sliding windows, the method dynamically identifies corresponding segments based on temporal alignment, allowing the segmentation to flexibly accommodate different sampling rates, variable lengths, and timing offsets while maintaining compatibility with machine learning corpus requirements.
2Adaptability or versatility
If time series with different sampling rates and variable lengths are incorporated into a machine learning corpus, then the diversity and applicability of the corpus increases, but the difficulty of aligning and processing these time series increases
Solution Approach 1:
The patent introduces dynamic time warping as an intermediary process that mediates between diverse time series with different characteristics and the requirements of machine learning corpus. This intermediary alignment mechanism transforms incompatible time series into a standardized format suitable for corpus incorporation, handling different sampling rates, variable lengths, and timing offsets without requiring complex custom processing for each pair.
3Adaptability or versatility
If dynamic time-warping procedure is used to align time series, then the ability to handle diverse sampling rates and time offsets is improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent applies segmentation by dividing the dynamic time warping process into identifying corresponding segments between time series. This segmentation approach breaks down the complex alignment task into manageable segments that can be processed more efficiently, reducing the overall computational burden and processing time while maintaining the ability to handle diverse sampling rates and time offsets.
Data Source
AI summary
A machine-learning system receives from multiple sensors a set of trace-data time series. Each time series contains a chronological sequence of sensor measurements of one attribute of one instance of a manufacturing product or process. The system partitions each series into a set of contiguous segments and selects one received series to be a standard series for each attribute. The starting and ending measurements of each non-standard time series are then time-aligned to the starting and ending points of the non-standard series' corresponding standard series, using a dynamic time-warping procedure. One or more segments of each aligned non-standard series are then aligned to each segment of the corresponding standard series. The resulting time-aligned, segmented time series are then incorporated into a corpus that is used by a machine-learning module to train a self-learning application.