Procedure and system for detecting a person's fall

• US 20060279426A1
• Filed: 06/05/2006
• Published: 12/14/2006
• Est. Priority Date: 06/07/2005
• Status: Active Grant

• Current Assignee: Commissariat a L'Energie Atomique (AES Corporation)
• Original Assignee: Commissariat a L'Energie Atomique (AES Corporation)
• Inventors: Bonnet, Stephane, Guillemaud, Regis
• Granted Patent: US 7,423,537 B2
• Time in Patent Office: Days
• US Class Current: 340/573.100
• CPC Class Codes: A61B 2562/0219 Inertial sensors, e.g. accelerometers, gyroscopes, tilt switches
• CPC Class Codes: A61B 5/1117 Fall detection
• CPC Class Codes: G06F 2218/00 Aspects of pattern recognition specially adapted for signal processing
• CPC Class Codes: G06V 40/23 Recognition of whole body movements, e.g. for sport training

Abstract

A person under supervision wears a sensor consisting of at least one accelerometer (31) and a magnetometer (41), oriented in his vertical direction. A fall event is picked up when a significant and rapid oscillation of the acceleration signal coincides with a shift in the ambient magnetic field between two levels (between t=4000 and t=5000). Additional criteria that may also make use of the magnetometer enable the diagnosis to be made, and it is easier and safer to establish this than with accelerometers alone.

Claims

1. ) Procedure for detecting the fall of a person, consisting of the following stages:

place on person (1) a sensor (2) comprising at least one accelerometer and at least one magnetometer,

continuously monitor the acceleration signal and the ambient magnetic-field signal provided by the sensor, and analyse their changes,

indicate a fall when a number of criteria have been met, such criteria including;

a) a period of instability in the acceleration signal, followed by stability in the acceleration signal;

b) a gradual change in the ambient magnetic-field signal during the instability period of the acceleration signal, between two different levels of the ambient magnetic-field signal measured before and after the instability period of the acceleration signal.

2. ) Procedure for detecting the fall of a person according to claim 1, characterized by the acceleration signal consisting of at least one vertical acceleration component along a vertical direction (VT) associated with the person.

3. ) Procedure for detecting the fall of a person according to claim 1, characterized by the ambient magnetic-field signal including a vertical component of the field along a vertical direction (VT) associated with the person.

4. ) Procedure for detecting the fall of a person according to claim 1, characterized by the fact that the criteria also include:

c) detection of the recumbent position in the stability of the acceleration signal by an effectively zero measurement along the vertical axis (VT) associated with the person.

5. ) Procedure for detecting the fall of a person according to claim 1, characterized by the fact that the criteria also include:

d) recognition of normal events likely to be confused with a fall.

6. ) Procedure for detecting the fall of a person according to claim 5, characterized by the fact that such recognition includes a measurement of the person'"'"'s azimuth.

7. ) Procedure for detecting the fall of a person according to claim 5, characterized by the fact that such recognition includes a comparison between at least one of the signals with an equivalent signal recorded previously and corresponding to one of the normal events, for the instability period of the acceleration signal.

8. ) Procedure for detecting the fall of a person according to claim 4., characterized by the fact normal events include going to bed.

9. ) Procedure for detecting a fall according to claim 1, characterized by the fact that the fall is indicated after a threshold (S) is crossed by a weighted numerical combination of criteria.

10. ) Procedure for detecting a fall according to claim 1, characterized by the fact that readings are converted according to the orientation of sensor (2) to express readings in a frame of axes associated with the person and consisting of a vertical axis, an antero-posterior axis and a medio-lateral axis, and including calibration of the sensor to determine the afore-mentioned orientation.