Personalized Fall Detection Algorithm via Segmented Data Transfer
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fall detection systems face challenges in reducing processing power and data transfer requirements, making them bulky and inefficient, especially when using machine-learning methods that require large datasets and significant processing capacity.
Innovation Solution
A method where a fall detector transmits partial sensor data and feedback information to an external device for updating a fall detection algorithm, allowing the external device to refine and train only specific coefficients, reducing the amount of data transferred and processing needed on the fall detector.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine-learning methods are used to improve fall detection accuracy, then detection accuracy is improved, but processing power requirements and data transfer requirements increase
Solution Approach 1:
The patent divides the machine learning model into two parts: a core model that runs locally on the fall detector device, and additional training data that is collected and processed separately. This segmentation allows the device to maintain high detection accuracy while reducing the processing power needed at the edge device, as the heavy training computations can be performed elsewhere using the collected sensor data segments.
Solution Approach 2:
The system performs preliminary data collection and preprocessing locally, storing sensor data segments and feedback information for later training. By preparing the data in advance and organizing it into manageable segments with corresponding feedback labels, the system reduces the computational burden during actual fall detection events, allowing accurate machine learning performance without requiring peak processing power at the moment of detection.
2Measurement precision
If machine-learning methods are used to improve fall detection accuracy, then detection accuracy is improved, but data transfer requirements increase
Solution Approach 1:
The patent segments the training data into manageable chunks of sensor data with corresponding feedback information. Instead of transferring entire datasets, the system divides training data into smaller segments that can be efficiently transferred between devices, reducing overall data transfer requirements while still providing sufficient data for accurate model training and updating.
Solution Approach 2:
The system extracts only the essential elements needed for training - specific sensor data segments and their corresponding feedback labels - and transfers only these extracted components. This extraction approach eliminates unnecessary data transfer by focusing only on the critical training elements, thereby reducing data transfer volume while maintaining training effectiveness.
3Measurement precision
If all sensor data is transmitted for training, then model accuracy is improved, but data transmission costs increase
Solution Approach 1:
The patent implements partial action by transmitting only selected segments of sensor data that are most relevant for training, rather than all available data. The system selectively identifies and transfers data segments that provide the most training value, achieving good model accuracy with reduced transmission energy by avoiding unnecessary data transfer.
Solution Approach 2:
By segmenting the sensor data into meaningful chunks with corresponding feedback information, the system enables efficient selective transmission. Only segments that contribute to improving model accuracy are transmitted, reducing overall transmission energy while maintaining training effectiveness through strategic selection of data segments.
Data Source
AI summary
A concept for personalizing a fall detection algorithm to a particular subject. Sensor data, responsive to a fall of a subject, is obtained at the fall detector, along with feedback information responsive to a confirmation of whether the subject has fallen and/or whether the subject had not fallen. Parts of the sensor data, and corresponding portions of the feedback information, are transmitted to an external device, which generate update information for the fall detection algorithm. The update information is then used by the fall detector to update, and thereby personalize, the fall detection algorithm.