Wearable Exercise Tracking with Optical-Sensor Activity Recognition

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing wearable devices face challenges in accurately recognizing exercise activities due to reliance on sensor data that requires calibration and struggles with complex user movements and similar data patterns, lacking the fusion of rich mid-level features for improved classification.

Innovation Solution

A wearable device integrates sensor data with optical movement data from an image-based monitoring system to enhance exercise activity recognition by retraining the movement tracker model using confidence indicators from both sources, prioritizing the optical data for improved accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If a movement tracker model is executed on sensor data from the wearable device, then processing speed is improved and latency is reduced, but measurement precision deteriorates due to reliance on sensor data requiring calibration and producing similar data for different activities

Engineering Contradiction:
Improveprocessing speedVSAvoidactivity recognition accuracy
Core Design Contradiction:
SpeedVSMeasurement precision

Solution Approach 1:

The patent combines sensor data from the wearable device with optical movement data from an image-based monitoring system. The processing unit executes both a movement tracker model on sensor data and an optical tracker model on optical movement data, then merges these data sources to determine exercise activities. This merging allows the system to maintain fast local processing while improving measurement precision through the addition of optical data that provides better differentiation between similar activities.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent introduces an image-based monitoring system as an intermediary that captures optical movement data. This intermediary system processes visual information to generate optical movement representations, which then serve as additional input for the movement tracker model. The intermediary optical data acts as a mediator that enhances the accuracy of activity recognition without requiring changes to the core wearable device processing.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Device complexity

If sensor data is used for exercise tracking, then device complexity is minimized and ease of operation is improved, but reliability deteriorates due to calibration requirements and difficulty in recognizing complex user movements

Engineering Contradiction:
Improvedevice complexityVSAvoidactivity detection reliability
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent adds a new dimension of data collection by incorporating an image-based monitoring system that captures optical movement data. This dimensional addition provides visual information that complements the sensor data from the wearable device. The processing unit correlates sensor data with optical movement data, creating a multi-dimensional approach to activity recognition that improves reliability without significantly increasing device complexity, as the optical processing can occur on a connected device.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentEP4177853B1Devices and methods for configuring a wearable device for exercise tracking
Publication Date: 2026.04.29 SONY GROUP CORP
  • EP4177853B1 patent drawingFigure 1~2B
  • EP4177853B1 patent drawingFigure 3~4
  • EP4177853B1 patent drawingFigure 5

AI summary

An apparatus (20) for configuring a wearable device (10) for exercise tracking, the apparatus comprising: a data storage unit (213), for storing sensor data generated by a sensor arrangement in the wearable device based on sensed movement of the wearable device, and optical movement data comprising representations of an individual obtained from an image-based monitoring system (30); and a processing unit (211) configured to: execute a movement tracker model on the sensor data to determine a first exercise activity and an associated first confidence indicator; execute an optical tracker model on the optical movement data to determine a second exercise activity and an associated second confidence indicator; process the movement tracker model based on the first confidence indicator and the second confidence indicator.