Dynamic Threshold Activity Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fitness tracking systems face challenges in accurately counting steps and repetitions during various activities due to false positives from non-specific sensor data, leading to inaccurate tracking and user experience issues.
Innovation Solution
A system that dynamically determines a threshold value for counting steps and repetitions based on predicted activities using machine learning models, incorporating data from accelerometers and other sensors, and adjusts these values based on previous activity patterns to differentiate between specific activities like walking, running, cycling, and others.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a fixed threshold value is used for counting steps and repetitions, then the device complexity is reduced, but the measurement precision deteriorates due to false positives from non-specific sensor data
Solution Approach 1:
The patent applies dynamics by transitioning from a fixed threshold value to a dynamic threshold determination system. The system now adjusts threshold values based on predicted activities using machine learning models, allowing the threshold to adapt to different exercise types and intensities. This resolves the contradiction by improving measurement precision through activity-specific thresholds while managing complexity through automated ML-based determination.
Solution Approach 2:
The patent changes the parameter of threshold values from static to dynamic based on activity predictions. Machine learning models analyze sensor data to predict activities, and these predictions drive changes in threshold parameters. This allows the system to improve counting accuracy by using appropriate thresholds for each activity type without requiring manual configuration, thereby resolving the precision-complexity contradiction.
2Measurement precision
If machine learning models are used to dynamically determine threshold values, then the measurement precision improves, but the use of energy increases due to computational processing
Solution Approach 1:
The patent applies preliminary action by pre-training machine learning models offline to recognize activity patterns. During actual exercise tracking, the pre-trained models efficiently process sensor data with minimal computational overhead. This approach improves measurement precision through accurate activity recognition while reducing energy consumption during the actual measurement phase, as the heavy computational work was performed in advance.
3Measurement precision
If activity-specific threshold values are used, then the measurement precision improves, but the device complexity increases due to multiple threshold management
Solution Approach 1:
The patent applies self-service by enabling the system to automatically determine appropriate threshold values based on machine learning activity predictions. The system autonomously selects and adjusts thresholds without requiring manual configuration or user input. This resolves the contradiction by improving measurement precision through activity-specific thresholds while managing complexity through automated self-adjustment mechanisms.
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
The system significantly improves accuracy in counting steps and repetitions during intended activities while minimizing false positives, providing a better user experience by accurately tracking fitness sessions and workouts.
Implementation Method 1
The user activity component 120 may include a filter component 135 configured to determine a magnitude value of the motion sensor data
Data Source
AI summary
Techniques for detecting user activity and repetitions (rep) counting are described. A system may dynamically determine a threshold value for counting reps. The threshold value may be based on an activity predicted by the system using sensor data. The threshold value may be based on a rep rate. Changes in the threshold value may enable more accurate rep counting when a user is performing an activity that involves taking reps, such as, walking, running, rowing bicycling, etc., and may further cease counting of reps when a user is performing an activity that does not involve taking repetitions but may result in sensor data that triggers rep counting.