Traveling Amount Estimation Using Sensor Clustering and Regression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating traveling distance using portable devices, such as smartphones, are not highly accurate as they rely on GPS or step counting, which can be unreliable due to variations in user stride and environmental factors.
Innovation Solution
A system comprising a model generation apparatus and a traveling amount estimation apparatus that extracts feature amounts from sensor data, clusters them, and uses regression functions to estimate traveling distance, improving accuracy by using supervised data and multiple regression analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If GPS function is used to determine position and traveling distance, then the user can check current position and traveling distance easily, but the measurement accuracy deteriorates due to environmental factors and signal reliability issues
Solution Approach 1:
The patent segments the traveling distance estimation problem into multiple feature dimensions (acceleration, angular velocity, barometric pressure, GPS coordinates) and processes each segment independently through clustering and regression analysis, then combines them for final estimation, thereby improving overall accuracy while maintaining ease of operation
Solution Approach 2:
The patent transforms the raw sensor data into clustered feature groups and applies different regression models for different clusters, dynamically changing the estimation parameters based on the characteristics of each data cluster to improve measurement precision
2Reliability
If step counting method is used to estimate traveling distance, then the system can work without GPS, but the measurement accuracy deteriorates due to variations in user stride
Solution Approach 1:
The patent applies local quality by creating different regression models for different clusters of walking patterns, where each cluster has its own optimized regression function tailored to specific walking characteristics, thereby improving accuracy for diverse user behaviors
Solution Approach 2:
The patent dynamically selects regression models based on real-time clustering of sensor data, allowing the system to adapt to changing walking conditions and patterns, making the estimation more reliable and accurate across different scenarios
3Measurement precision
If multiple sensor data and clustering analysis are used to improve estimation accuracy, then the measurement precision improves, but the device complexity increases
Solution Approach 1:
The patent performs preliminary clustering analysis on sensor data to categorize different walking patterns before applying regression analysis, preparing the data structure in advance to facilitate more efficient and accurate processing in subsequent steps
Solution Approach 2:
The patent creates simplified regression models that replicate the complex relationships between multiple sensor inputs and traveling distance, using clustered feature representations that capture essential patterns without requiring full complexity of the original multi-sensor system
Data Source
AI summary
One aspect of the present disclosure relates to a traveling amount estimation apparatus, comprising: at least one processor; and at least one memory that stores a program executed by the processor, wherein the processor is configured to: extract a feature amount from sensor data obtained from a traveling user and relating to traveling of the user; determine to which cluster the feature amount belongs; and estimate a traveling amount of the user from the feature amount in accordance with a regression function for the determined cluster.