Task Estimation Model Using Wearable Sensor Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating task states from wearable device sensor data struggle with accurately distinguishing between tasks with different contents, especially when movement variations are similar, leading to difficulties in preparing teaching data and achieving high estimation accuracy.
Innovation Solution
An information processing apparatus that acquires measurement information from wearable devices, extracts feature vectors, clusters them using various division methods, calculates detection ratios, and generates a task estimation model through machine learning to determine whether a task is being performed, thereby improving estimation accuracy and reducing the effort required for preparing teaching data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If simple waveform examination methods are used to estimate task states, then the ease of operation is improved, but the measurement precision deteriorates because higher-order states with movement variations cannot be accurately distinguished
Solution Approach 1:
The patent segments the continuous measurement information into discrete operation sections by detecting predetermined operations as delimiters. This segmentation transforms the continuous sensor data into discrete segments that can be individually analyzed, enabling accurate identification of task states within each segment while maintaining ease of operation through automated processing.
Solution Approach 2:
The patent extracts feature vectors from the measurement information within operation sections, isolating the essential characteristics needed for task state estimation. This extraction process removes irrelevant data and focuses on key features that distinguish different task states, improving measurement precision without complicating the overall system.
2Measurement precision
If multiple division methods are used to cluster feature vectors, then the measurement precision of task state estimation is improved, but the device complexity increases due to multiple processing steps
Solution Approach 1:
The patent employs multiple division methods dynamically to cluster feature vectors, where each division method processes the data through different clustering approaches. This dynamic application of multiple methods allows the system to capture various patterns in the data, improving task state estimation accuracy while the automated selection and application of these methods prevents excessive manual complexity.
Solution Approach 2:
The patent creates multiple copies of the feature vectors and processes them through different division methods simultaneously. This copying approach allows parallel processing of the same data through multiple clustering perspectives, improving measurement precision by comparing results from different methods without requiring sequential processing that would increase device complexity.
3Measurement precision
If clustering analysis is performed on operation sections to identify element operations, then the measurement precision is improved, but the loss of time increases due to additional processing steps
Solution Approach 1:
The patent performs preliminary clustering analysis on operation sections to pre-identify element operations before final task state estimation. This preliminary action groups similar operations together in advance, creating a structured foundation that speeds up the subsequent task state determination process, thereby reducing overall processing time while maintaining high measurement precision.
Solution Approach 2:
The patent applies partial clustering analysis only to operation sections that contain predetermined operations, rather than processing all measurement information uniformly. This selective approach focuses computational resources on relevant segments, improving operation detection accuracy where needed while minimizing time loss by skipping unnecessary processing in non-operational segments.
Data Source
AI summary
An information processing apparatus includes a processor that acquires measurement information measured by a device put on a task performer; detects a predetermined operation in the measurement information; extracts feature vectors from an operation section in which the predetermined operation has been detected; divides the feature vectors into clusters by division methods; obtains ratios at which the feature vectors are classified into the respective clusters in a predetermined time; performs learning by using the ratios as an input and using, as an output label, information indicating whether the task performer performs a predetermined task in the predetermined time; generates a task estimation model by determining weighting for the clusters based on a result of the learning; and estimates, by using the task estimation model, whether the task performer is in a process of performing the predetermined task.