Portable Terminal Status Classification Using Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies fail to accurately determine the current status of users with portable terminals, such as whether they are moving or engaged in activities like watching videos or sleeping, leading to potential misunderstandings before communication.
Innovation Solution
A current status presentation system that collects and classifies data from GPS, acceleration, and operation management information using various algorithms to present accurate status images on portable terminals, enabling users to understand the status of acquaintances before initiating communication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor data is collected and classified to determine user status, then measurement precision of user status is improved, but device complexity increases
Solution Approach 1:
The system segments the user status determination into multiple independent classification algorithms (e.g., decision tree, neural network, support vector machine), each handling specific aspects of status classification. This modular approach improves measurement precision through ensemble methods while managing device complexity by organizing algorithms as separate, interchangeable components.
Solution Approach 2:
The system implements a universal classification framework that can handle multiple types of sensor data (acceleration, GPS, barometer) and classify various user statuses (walking, running, driving, sleeping) using the same algorithmic structure. This multi-functional approach improves measurement precision across diverse scenarios without proportionally increasing device complexity.
2Reliability
If multiple types of information are collected from sensors, then reliability of status classification is improved, but loss of energy increases
Solution Approach 1:
The system implements partial data collection by activating only the necessary subset of sensors based on current classification needs. For example, GPS is activated only when location-based status determination is required, and the barometer is used specifically for elevation-based activity detection. This approach maintains classification reliability while significantly reducing overall energy consumption compared to continuous full-sensor operation.
Solution Approach 2:
The system employs periodic sampling of sensor data rather than continuous collection, adjusting the sampling frequency based on the current classification confidence level. When classification is already high-confidence, sampling intervals are extended, reducing energy consumption. When uncertainty increases, sampling frequency increases to maintain reliability.
3Measurement precision
If classification algorithms are used to determine user status, then measurement precision is improved, but loss of time in processing increases
Solution Approach 1:
The system performs preliminary filtering and preprocessing of sensor data before applying complex classification algorithms. Simple features like acceleration magnitude and GPS speed are pre-calculated and used for quick initial status assessment. This preliminary action reduces the dimensionality and complexity of data requiring full algorithmic processing, thereby maintaining measurement precision while reducing overall processing time.
Solution Approach 2:
The system implements a cascaded classification approach where simple, fast algorithms are applied first to handle obvious cases (e.g., high acceleration clearly indicates exercise). Only when these quick assessments are inconclusive does the system proceed to more computationally intensive algorithms. This skipping approach maintains high measurement precision for clear cases while minimizing processing time by avoiding unnecessary complex computation.
Data Source
AI summary
A current status presentation system includes a collector that collects, for each user who possesses a portable terminal, a plurality of types of information including position information, an acceleration, and operation management information indicating an operation status of an application software, a classifier that classifies, for each user, whether a current status of a user is moving or playing a game, based on a combination of an amount of change in the position information and a magnitude of the acceleration, the operation management information, and any one of classification algorithms, and a presenter that, when the current status is being moving, presents a first status image representing a classified current status in association with a user image, and when the current status is being playing the game, presents a second status image representing the classified current status in association with the user image.