RF Fall Classification Using Signal-Quality Activity Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current fall detection systems face challenges in accurately classifying fall events and types, particularly in distinguishing between different types of falls and determining medical urgency, due to limitations in signal quality and sensitivity, especially when subjects move or fall in areas with interference.
Innovation Solution
A fall-classification device that receives wireless radiofrequency communication signals, determines signal-quality values, and uses subject-activity metrics to detect falls and classify types by analyzing signal variations before and after the event, with adjustable transmission power and sensitivity settings to improve accuracy and reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If signal-quality determination is enhanced to improve fall classification accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The wireless communication signals serve dual purposes: both for data transmission and for fall detection. The receiver unit processes communication signals to extract subject-activity metrics, eliminating the need for separate dedicated sensors and reducing overall device complexity while maintaining high measurement precision
Solution Approach 2:
Signal-quality values act as an intermediary parameter that bridges wireless communication and fall detection. By determining signal-quality values from received wireless signals and using them to derive subject-activity metrics, the system achieves accurate fall classification without requiring direct complex sensing mechanisms
2Measurement precision
If sensitivity is increased to detect subtle fall movements, then measurement precision is improved, but reliability deteriorates due to false alarms from interference
Solution Approach 1:
The fall detection process is segmented into multiple analysis stages: signal-quality determination, subject-activity metric extraction, and fall classification. By dividing the detection process and analyzing different aspects separately, the system can maintain high sensitivity for detecting subtle movements while filtering out interference through multiple verification steps
Solution Approach 2:
The system dynamically adjusts detection parameters based on signal-quality values. By adapting the analysis to current signal conditions and using real-time signal-quality assessment, the system maintains optimal sensitivity while reducing false alarms caused by varying interference levels
3Measurement precision
If multiple subject-activity metrics are analyzed to improve fall type classification, then measurement precision is improved, but loss of information increases due to data processing requirements
Solution Approach 1:
The system extracts only the essential subject-activity metrics from wireless communication signals that are directly relevant to fall detection. By selectively extracting specific metrics rather than processing all available data, the system achieves accurate fall type classification while minimizing data processing overhead and information loss
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention is directed to a fall-classification device (100) with a receiver unit (102) configured to receive wireless communication signals (W) from wireless transmitters (101, 103) and a subject-activity data determination unit (106) configured to determine and provide, using signal-quality values determined by a quality signal determination unit, respective subject-activity data indicative of at least two subject-activity metrics of a subject (S) within a sensing volume (V1). A fall-event detection unit (108) is configured to determine, based on a time variation of the subject-activity data of at least one subject-activity metric, whether a fall-event has occurred. A fall-event type classification unit (110) is configured, using a predetermined algorithm and subject-activity data of at least two subject-activity metrics determined during a first time span before and a second time span after the determination of the fall-event, to determine a fall type, from a predetermined list of fall types, with increased accuracy.