Living-Body Data Interpolation Using Time and Position Models

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing data processing methods for living body data interpolation, such as PTL 1, face precision issues due to lack of consideration for time and place conditions when interpolating missing data, leading to decreased accuracy.

Innovation Solution

A data processing device and method that classifies living-body information data based on user attributes, including sensor values, measurement time, and position, and generates models using machine learning to estimate interpolation data, accounting for correlations among these factors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If interpolation is performed using accumulated past data sets focusing on similarity, then data deficit can be filled, but precision of interpolated data decreases when time and place conditions differ

Engineering Contradiction:

Improvedata deficitVSAvoidinterpolation precision

Core Design Contradiction:

Loss of informationVSMeasurement precision

Solution Approach 1:

The patent applies local quality by creating separate interpolation models for different time periods (work days vs. weekend days) and different measurement positions (wrist, ankle, etc.). This ensures that interpolation is performed using data from locally similar conditions rather than globally similar data, thereby maintaining precision while filling data deficits.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent segments the past data sets into multiple groups based on time attributes (work days, weekend days, holidays) and measurement position attributes. By dividing the data into these segments, the system can select the most appropriate segment for interpolation, improving precision while still filling data gaps effectively.

Inventive Principle:

Principle #1Segmentation

2Duration of action of moving object

If measurement frequency is reduced due to battery capacity or body motion noise, then device usage is extended, but necessary living-body information data may be lost

Engineering Contradiction:

Improvedevice usage durationVSAvoidliving-body information data

Core Design Contradiction:

Duration of action of moving objectVSLoss of information

Solution Approach 1:

The patent performs preliminary action by proactively interpolating missing living-body information data using pre-built models that account for time and position attributes. Instead of waiting for data loss to occur, the system prepares interpolation models in advance and automatically fills gaps, ensuring continuous data availability while extending device usage duration.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system uses feedback by continuously learning from accumulated measurement data to improve its interpolation models. The models are refined based on actual measurement patterns, time attributes, and position attributes, enabling more accurate data reconstruction and reducing information loss over time.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20240127964A1Data processing device, system, data processing method, and recording medium

Publication Date: 2024.04.18 NEC CORP

AI summary

A data processing device that includes a classification unit that classifies at least one piece of living-body information data into at least one group based on an attribute of at least one user, the at least one piece of living-body information data including a sensor value relating to a living body of the user, and a measurement time and a measurement position of the sensor value, and a learning unit that generates, for each of the group, a model for estimating interpolation data for interpolating a deficit of the living-body information data using a correlation among the sensor value included in the living-body information data, the measurement time, and the measurement position.