Waveform Similarity Calculation via Axis Shifting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for extracting similar waveforms from historical data are labor-intensive and prone to misalignment errors, leading to undervaluation of similar waveforms due to absolute value misalignment and scaling differences.
Innovation Solution
A data similarity calculation method that involves converting time series data into two-dimensional coordinates, shifting and scaling the waveforms to maximize similarity by adjusting both the time and value axes, and extracting the optimal shift and scaling parameters to align and match waveforms accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If waveform similarity is calculated using absolute value alignment, then calculation speed is improved, but measurement precision deteriorates due to misalignment errors
Solution Approach 1:
The patent transforms the one-dimensional time series data into two-dimensional coordinate space, where the x-axis represents time and the y-axis represents the value axis. This dimensional transformation enables independent adjustment of time-axis shift and value-axis shift, allowing the system to simultaneously optimize both alignment accuracy and calculation efficiency by operating in a higher-dimensional parameter space.
Solution Approach 2:
The patent introduces two independent parameters: time-axis shift amount and value-axis shift amount. By changing these parameters systematically, the patent enables flexible adjustment of waveform alignment without requiring complex manual intervention. The dual-parameter approach allows the system to explore different alignment configurations efficiently while maintaining high measurement precision.
2Measurement precision
If manual searching of historical data is performed, then measurement precision is improved, but loss of time increases significantly
Solution Approach 1:
The patent enables the system to automatically perform waveform alignment and similarity calculation without requiring manual intervention. The computer automatically adjusts time-axis and value-axis shift parameters, calculates similarities, and identifies matching waveforms. This self-service capability eliminates the time-consuming manual search process while maintaining high accuracy through systematic parameter exploration.
3Measurement precision
If waveform alignment requires precise manual adjustment, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent separates the alignment process into two independent adjustment components: time-axis shift and value-axis shift. This segmentation allows each parameter to be adjusted independently through simple additive operations, avoiding the need for complex coupled adjustment mechanisms. The decomposition of the alignment problem into independent dimensional adjustments significantly reduces system complexity while maintaining high alignment precision.
Data Source
AI summary
There is provided a data similarity calculation method. The method includes: (a) acquiring a first waveform; (b) storing time series data; (c) converting the stored time series data into a waveform on two-dimensional coordinates, wherein the two-dimensional coordinates consists of a time axis and an a value axis representing values of the time series data, and the time axis is orthogonal to the value axis; (d) shifting the converted waveform in both directions of the time axis and the value axis so as to generate a second waveform; (e) calculating a similarity between the first waveform and the second waveform; and (f) extracting a shift amount in the direction of the time axis and a shift amount in the direction of the value axis when the similarity is the highest by repeatedly performing steps (d) and (e).