Personalized Fall Detection Using Subject-Specific Movement Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automatic fall detectors struggle with low accuracy in distinguishing between falls and everyday movements, leading to high false alarm rates and the need for a cancel function that users may find difficult to use, especially when unconscious or panicked.
Innovation Solution
A computer-based method trains a classifier using subject-specific movement data to generate personalized fall and non-fall event probability distributions, reducing false alarms by personalizing the detection to individual user behaviors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fall detectors are configured to have a sufficiently high false alarm rate to minimize missed falls, then fall detection sensitivity is improved, but false alarm rate increases
Solution Approach 1:
The system performs preliminary learning during a data collection period before actual fall detection begins. It collects movement data during everyday activities to establish a subject-specific baseline of normal behavior patterns, then uses this baseline to configure the classifier thresholds that will be used during subsequent fall detection operations.
Solution Approach 2:
The system dynamically adjusts the false alarm rate threshold based on the subject's individual movement characteristics. Instead of using a fixed threshold for all users, the classifier is trained on each subject's specific movement patterns during the learning period, allowing the detection criteria to adapt to individual variations in movement behavior.
2Object-generated harmful factors
If a cancel function is provided to handle false alarms, then false alarm management is enabled, but system complexity and user burden increase
Solution Approach 1:
The system performs self-calibration during the data collection period by automatically analyzing the subject's movement patterns and establishing personalized baseline characteristics. This self-learning process eliminates the need for manual configuration or user intervention to optimize detection parameters, allowing the system to adapt to each user automatically.
3Measurement precision
If subject-specific probability distributions are used to improve detection accuracy, then fall detection precision is improved, but data collection and processing requirements increase
Solution Approach 1:
The system collects movement data for a predetermined period that is sufficient to establish reliable subject-specific patterns, rather than requiring exhaustive long-term data collection. This partial action approach gathers only the necessary amount of data needed to create an accurate baseline model for that specific user.
Data Source
AI summary
A method and system for training a fall detection classifier using subject-specific movement data. Movement data responsive to a subject's movements during everyday activities are obtained over a predetermined data collection period. For each detected event in the movement data, values for one or more parameters that may (together or individually) indicate a fall are obtained. The obtained values are used to generate a subject-specific probability distribution for non-fall events. A fall event probability distribution is obtained using the non-fall event probability distribution. This fall event probability distribution can then be subsequently processed, with reference to a threshold value, to distinguish fall events from non-fall events.