Hierarchical Tree Model for Smartphone Movement Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for distinguishing user movement behavior on smartphones, such as those using GPS or microphones, are inefficient and consume excessive battery power, and existing data processing techniques lack accuracy in differentiating movement behaviors.
Innovation Solution
A method and mobile device that utilize a hierarchical tree model constructed from acceleration sensor data, dividing data into first and second frame groups with overlapping segments to extract characteristic factors for accurate movement behavior classification, eliminating the need for additional devices like GPS or microphones.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GPS or microphone devices are used to distinguish user movement behavior, then measurement precision is improved, but use of energy deteriorates
Solution Approach 1:
The acceleration sensor performs multiple functions: it not only detects movement behavior but also provides data for both first and second frame groups, eliminating the need for separate GPS or microphone devices. This multi-functional use reduces device count and energy consumption while maintaining measurement precision through sophisticated data processing of the acceleration data.
Solution Approach 2:
The patent changes the temporal parameters of data processing by creating two different frame groups with different time units (first frame group with one time unit, second frame group with another time unit). This parameter transformation allows the single acceleration sensor to provide sufficient information for accurate movement behavior distinction without requiring additional energy-intensive sensing devices.
2Measurement precision
If conventional data processing techniques are used, then device complexity is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent segments the acceleration data into two distinct frame groups: a first frame group divided based on a first time unit and a second frame group divided based on a second time unit. This segmentation allows the system to extract different characteristic factors from each group, improving measurement precision by analyzing data at multiple temporal resolutions without requiring complex hardware.
Solution Approach 2:
The patent introduces a temporal dimension by creating frame groups with different time units. This dimensional transformation of the data processing approach allows extraction of characteristic factors that are not visible in single-time-unit analysis, thereby improving movement behavior distinction accuracy while maintaining relatively simple device architecture.
3Measurement precision
If acceleration data is collected and processed using hierarchical tree model, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The patent performs preliminary action by pre-dividing acceleration data into first and second frame groups with different time units before actual movement behavior analysis. This pre-processing organization allows the hierarchical tree model to efficiently extract characteristic factors from structured data, improving measurement precision while reducing the time required during actual movement classification through the pre-organized data structure.
Data Source
AI summary
A mobile device for distinguishing a user's movement, a method therefor, and a method for generating a hierarchical tree model therefor. Particularly, the mobile device includes: an acceleration sensor; a buffer for collecting the acceleration data outputted from the acceleration sensor according to a user's specific movement; an extraction unit for extracting the characteristic elements of the user's specific movement based on the acceleration data collected by the buffer; and a movement judgment unit for determining to which class the user's specific movement belongs by inputting the characteristic elements extracted by the extraction unit into a pre-structured hierarchical tree model, wherein the hierarchical tree model is pre-structured based on the characteristic elements extracted for each movement.