Real-Time User Action Recognition With Two-Level Screening

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing fitness motion monitoring devices require users to pre-identify their actions, leading to poor user experience and non-real-time feedback, as they only provide information after the exercise is completed.

Innovation Solution

A method and system for identifying user actions in real-time by comparing user action data with candidate reference action data using a two-level screening process involving difference degree-based and probability-based operations, allowing immediate feedback without prior action type input.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If users pre-identify their action types before exercising, then the motion monitoring device can select appropriate reference action data for accurate monitoring, but the user experience deteriorates due to the additional input requirement

Engineering Contradiction:
Improveaction monitoring accuracyVSAvoiduser operation convenience
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The system automatically identifies action types by analyzing motion data from multiple sensors without requiring user input. The device performs self-service by autonomously determining the user's current action type and selecting appropriate reference data, eliminating the need for manual action type selection while maintaining monitoring accuracy

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The manual mechanical process of users selecting action types is replaced with an automated system that uses sensor data processing and pattern recognition algorithms to identify actions, substituting human operation with an automated detection mechanism

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Device complexity

If the motion monitoring device provides feedback only after exercise completion, then the system complexity is reduced, but the user experience deteriorates due to lack of real-time feedback

Engineering Contradiction:
Improvesystem complexityVSAvoidfeedback delay time
Core Design Contradiction:
Device complexityVSLoss of time

Solution Approach 1:

The system performs preliminary action by continuously analyzing motion data during exercise and identifying action types in real-time, preparing and providing feedback information as the exercise progresses rather than waiting until completion, thus reducing feedback delay while maintaining manageable system complexity

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system maintains continuous useful action by continuously processing sensor data and providing ongoing action identification and feedback throughout the exercise session, ensuring uninterrupted real-time monitoring and information delivery to the user

Inventive Principle:
Principle #20Continuity of useful action

3Measurement precision

If the system uses a two-level screening operation with difference degree and probability calculations, then the action identification accuracy is improved, but the computational complexity increases

Engineering Contradiction:
Improveaction identification accuracyVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The computational process is segmented into two distinct levels: first-level screening using difference degree calculation to quickly eliminate unlikely action types, and second-level screening using probability calculations to precisely identify the current action. This segmentation reduces overall computational complexity by dividing the complex task into manageable stages

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs partial action by first conducting a rough screening with difference degree calculation to filter out obviously incorrect action types, then applying the more computationally intensive probability calculations only to the remaining candidate actions, thus reducing total computational burden while maintaining high accuracy

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP4167129B1Method and system for recognizing user actions
Publication Date: 2025.07.09 SHENZHEN SHOKZ CO LTD
  • EP4167129B1 patent drawingFigure 1
  • EP4167129B1 patent drawingFigure 2
  • EP4167129B1 patent drawingFigure 3

AI summary

The embodiment of the present disclosure provides a method and a system for identifying a user action. The method and system may obtain user action data collected from a plurality of measurement positions on a user, the user action data corresponding to an unknown user action, identify that the user action includes a target action when obtaining the user action data based on at least one set of target reference action data, the at least one set of target reference action data corresponding to the target action, and send information related to the target action to the user.