Asynchronous Intercorrelated Time Series Datasets Alignment Method

Simple SummaryContent extracted from patent full text and abstract with AI.

This invention provides a computer-implemented method for accurately aligning two asynchronous, intercorrelated time series datasets, such as signals from different sensors, even when the signals have significant noise or are morphologically distinct. The method divides the datasets into smaller segments and uses two neural networks—each trained to transform and filter one dataset—to map the signals into a correlated space. Calculated time shifts between paired segments are then refined using a robust multi-model fitting algorithm, enabling precise synchronization of the datasets.

Use CasesContent extracted from patent full text and abstract with AI.

• Synchronizing physiological signals from different biomedical sensors (e.g., ECG and PPG or BCG) for advanced medical diagnostics and monitoring.
• Aligning heterogeneous sensor data in industrial applications, such as predictive maintenance in Industry 4.0 systems.
• Sensor calibration and benchmarking by comparing sensor outputs against a gold standard reference over time.
• Synchronizing multi-sensor data streams for research in experimental physics, astronomy, or oceanography.
• Enhancing autonomous vehicle sensor fusion by aligning independent sensor streams with disparate internal clocks.
• Signal alignment in telecommunications, radar, or aerospace signal intelligence systems for improved data analysis.

BenefitsContent extracted from patent full text and abstract with AI.

• Enables automatic, high-precision alignment of asynchronous, noisy, and morphologically different time series datasets from various sources.
• Reduces or eliminates the need for manual data preprocessing, masking, and filtering thanks to adaptive neural networks.
• Easily adaptable to new sensor types and datasets via transfer learning, providing significant flexibility and scalability.
• Improves the robustness and accuracy of downstream data analysis and sensor fusion tasks.
• Parallelizable processing for efficient alignment of large datasets, saving computational resources and time.
• Outperforms traditional alignment methods (like Dynamic Time Warping or Cross-Correlation Lag Analysis) especially for high-frequency, noisy, or partially correlated signals.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

The present invention concerns a computer-implemented method for aligning intercorrelated timeseries comprising the steps of:(a) retrieving a first time series dataset (x) and a second time series dataset (y), the first and second time series dataset being intercorrelated,(b) segmenting each of the first and second time series dataset (x, y) into a plurality of consecutive smaller segments (xi, yi), all segments of the first and second time series dataset (x, y) having the same length,(c) determining pairs of corresponding segments by associating successive segments of the first time series dataset (x) with corresponding segments of the second time series dataset (y),(d) optimizing, for each pair of corresponding segments, a correlation function to obtain an approximation of a first times series transformation function (f1) and of a second time series transformation function (f2), the first time series transformation function (f1) being parametrized by weights (w1) of a first neural network (N1) and the second time series transformation function (f2) being parametrized by weights (w2) of a second neural network (N2),(e) using the first and second time series transformation functions (f1,f2) to determine a vector of segment shifts s whose components contain approximations of the shifts between the first and the second segment in a pair of corresponding segments (xi, yi),(f) applying a multi-model fitting algorithm to the segment shift vector (s), said multi-model algorithm outputting a shift function (fopt) for aligning segments of each pair of corresponding segments (xi, yi), and(g) aligning the first time series dataset (x) with the second series dataset by applying said shift function (fopt) to all pairs of corresponding segments (xi, yi).