Activity Tracking System Using Sensor Data Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing activity tracking devices fail to accurately sense and analyze a wide range of user activities, such as walking, running, cycling, and others, due to variations in sensor data and how the device is carried, leading to challenges in classification.
Innovation Solution
A system that includes movement sensors and a wireless interface, communicating with a mobile communication device and computing hardware to process sensor signals using specialized processors and classification algorithms like neural networks, decision forests, and support vector machines, generating a temporal log of activities and displaying them on a graphical user interface.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple sensor types and complex classification algorithms are used to identify diverse user activities, then activity recognition accuracy is improved, but device complexity and processing requirements increase
Solution Approach 1:
The system segments the activity recognition task by deploying multiple specialized processors, each dedicated to identifying specific activity types (walking, running, cycling, etc.). This division allows complex classification to be distributed across simpler, specialized components rather than requiring one monolithic complex processor.
Solution Approach 2:
The mobile communication device serves multiple functions: it acts as both the sensor platform and the processing unit. The device's existing sensors (accelerometer, GPS, etc.) are utilized for activity tracking, while its processor handles the classification algorithms, eliminating the need for separate dedicated hardware for each function.
2Speed
If parallel processing with specialized processors is used to identify multiple activity characteristics simultaneously, then processing speed and activity identification accuracy are improved, but computational resource requirements increase
Solution Approach 1:
The system dynamically adjusts processing based on detected activities. Not all processors run continuously at full capacity - they are activated based on the current activity context and sensor data patterns, allowing the system to optimize resource usage while maintaining fast response times when needed.
Solution Approach 2:
The classification algorithms are executed locally on the mobile device itself rather than requiring continuous cloud processing. The device's own computational resources handle the analysis, reducing ongoing energy consumption while maintaining processing speed through efficient local execution.
3Loss of information
If comprehensive sensor data collection from multiple sensors is performed to capture all user activities, then measurement completeness is improved, but data processing complexity and power consumption increase
Solution Approach 1:
The system performs preliminary filtering and preprocessing of sensor data before full classification. Basic patterns are detected and filtered early in the processing pipeline, allowing the system to discard irrelevant data before committing significant computational resources to detailed analysis, thus reducing overall power consumption while maintaining measurement completeness.
Solution Approach 2:
Different processing strategies are applied to different sensor data streams based on their characteristics and relevance to specific activities. Not all sensors are processed with the same level of intensity at all times - processing depth is adjusted locally for each sensor type and activity context, optimizing the balance between completeness and energy usage.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Accurately identifies and analyzes various user activities by processing sensor data from mobile communication devices, providing a timeline representation of activities, enhancing the accuracy and comprehensiveness of activity tracking.
Implementation Method 1
one or more sensors of the mobile communication device are implemented using at least one of: a gyroscopic angular sensor, an accelerometer
Implementation Method 2
The mobile communication device includes a Global Positioning System (GPS) receiver
Implementation Method 3
a gyroscopic angular sensor
Data Source
AI summary
In an embodiment, a mobile device determines sensor signals from one or more sensors of the mobile device, the sensor signals indicating motion that is associated with activities of a user of the mobile device. The mobile device determines activity types that correspond to at least a portion of the determined sensor signals. The mobile device determines activity periods of the user, each activity period corresponding to a period of time for one of the determined activity types. The mobile device then performs a heuristics-type analysis based on the length of time for each activity period and the corresponding activity types to determine whether two or more activity periods should be combined into a single activity period. The determined activity types are displayed on the mobile device.