Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02529545 2005-12-09
1
TITLE OF THE INVENTION
[0001] System and method for monitoring of activity and fall
FIELD OF THE INVENTION
[0002] The present invention relates to monitoring of a subject.
More specifically, the present invention is concerned with a system and a
method for monitoring of activity and fall of a subject.
SUMMARY OF THE INVENTION
[0003] More specifically, there is provided a system for monitoring
activity of at least one subject in an environment, comprising at least one
sensing assembly, detecting parameters in the environment; and a server
communicating with at least one of: i) the subject and ii) the at least one
sensing assembly; wherein the at least one sensing assembly comprises at
least a first sensor connected to the region of the back of the neck of the
subject, the first sensor unit comprising at least one accelerometer.
[0004] There is further provided a system for monitoring activity of a
subject in an environment, comprising at least one sensing assembly, detecting
parameters of the given environment; and a server communicating with at least
one of: i) the subject and ii) the at least one sensing assembly; wherein the
at
least one sensing assembly comprises at least a first sensor unit at the back
of
the neck of the subject.
[0005] There is further provided a method for monitoring activity of a
subject in an environment, comprising providing at least one sensing assembly
in the environment of the subject; providing a server communicating with at
CA 02529545 2005-12-09
2
least one of: i) the subject and ii) the at least one sensing assembly;
generating
property vectors from data collected by the at least one sensing assembly;
characterizing activity of the subject from the property vectors; and having a
result of said characterizing step accessible to the server.
[0006] Other objects, advantages and features of the present
invention will become more apparent upon reading of the following non-
restrictive description of embodiments thereof, given by way of example only
with reference to the accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
[0007] In the appended drawings:
[0008] Figure 1 is a diagram of a first embodiment of a system
according to the present invention;
[0009] Figure 2 is a diagram of a second embodiment of a system
according to the present invention;
[0010] Figure 3 is a diagram of a third embodiment of a system
according to the present invention;
[0011] Figure 4 is a diagram of a fourth embodiment of a system
according to the present invention;
[0012] Figure 5 is a diagram of a fifth embodiment of a system
according to the present invention;
CA 02529545 2010-07-27
3
[0013] Figure 6 a flowchart of an embodiment of a method according
to the present invention;
[0014] Figure 7 is a flowchart of another embodiment of a method
according to the present invention;
[0015] Figure 8 is a diagram of positioning obtained by a method
according to the present invention;
[0016] Figure 9 illustrates an embodiment of a neck assembly
according to the present invention;
[0017] Figure 10 illustrates an embodiment of a wrist assembly
according to the present invention;
[0018] Figure 11 is a flowchart of a further embodiment of a method
according to the present invention; and
[0019] Figure 12 is a diagram showing the results obtained by a
method according to the present invention.
DESCRIPTION OF EMBODIMENTS OF THE INVENTION
[0020] A system as illustrated in Figure 1 comprises a sensing
assembly 14 and a server 16, a subject to be monitored 12 being in its
environment 10.
[0021] The sensing assembly 14 illustrated in the embodiment of
CA 02529545 2010-07-27
4
Figure 1 comprises a first sensor unit 18, located in the region of the neck
of
the subject 12, for example integrated in a neck assembly worn by the subject
12, and a second sensor unit 20 located in the region of the diaphragm, or/and
the fist and/or of the leg of the subject 12, for example in a wrist band.
[0022] The sensor units include at least one 2- or 3- axes
accelerometers. They may further comprise a gyroscope. The respective
number, combination and location of the different sensors depend on target
monitoring data, as will be explained hereinbelow.
[0023] For example, the first sensor unit 18 may comprise a high G
accelerometer and a low G accelerometer, while the second unit sensor 20
comprises a low G accelerometer, both optionally further comprising a
gyroscope.
[0024] Alternatively, the sensor unit 18, located at the base of the
neck of the subject 12, integrated in a neck assembly that the subject 12
wears,
may comprise a three-axes high G sensor and a gyroscope. The sensor unit
20, worn as a bracelet, comprises a low G accelerometer and a gyroscope. The
sensing assembly 14 communicates with a base 22, located in the environment
10 of the subject 12. This base 22 is connected by a phone link 24 and by
Internet 26 to a server 16 for information exchange. Access to the remote
server 16 is controlled and allows target persons, such as a physician 28,
employees of a health center 30, members of the family 32 of a human subject
12 for example, as well as a call center 34 to monitor data and profiles
corresponding to the subject 12 from a distance. The remote server 16 is also
used as an interface for sending messages and instructions to the different
parts of the system.
[0025] The system automatically detects falls and critical activity
CA 02529545 2010-07-27
levels of the subject 12 and is able to emit a request for intervention or
alarm,
as will be further described hereinbelow.
[0026] The base 22 may support a remotely modifiable and
programmable reminder function useful for assisting subjects with a cognitive
5 deficiency, whereby remote-intervention functions are allowed. The base 22
may also comprise means for processing data and alarms received from the
sensing assembly 14, as well as means for bi-directional voice communication.
It may further support a mobile unit of the wireless type offering similar
features
as just described and optionally integrating GPS localization means allowing
monitoring the subject 12 outdoors for example.
[0027] In the embodiment illustrated in Figure 2, the sensor
assembly 14 consists of a sensor unit 18 comprising a 3-axes high G sensor
and a gyroscope, integrated in a neck assembly worm by the subject 12. The
neck assembly comprises RF communication means to the base unit 22, and a
device for asking help 36. The base 22 is a hands-free unit allowing wireless
communication, through a 2.4 GHz RF link of a ZigbeeTM network for example,
to the neck assembly and optional detectors 38. The base 22 includes a help
button and a reset button. The base 22 is linked to a remote server 16 by
standard telephone network 40.
[0028] In the embodiment illustrated in Figure 3, the base 22 is a
free-hands phone allowing wireless communication with the sensor unit 18 and
optional detectors 38 on the one hand, and to a remote server 16 via a
cellular
network 40 on the second hand.
[0029] As exemplified in Figure 1, optional detectors 38 may include
for example three-dimensional locaters and interphones (L), motion sensors
CA 02529545 2010-07-27
6
(M) and presence detectors (D), pillbox sensor (P), smoke detectors (F) and
household-appliance detectors (S).
[0030] In the embodiment illustrated in Figure 4, the subject 12
wears a first sensor unit 18 as a neck assembly and a second sensor unit 20 as
a wristband. Both sensor units are connected by a unidirectional low frequency
low power RF communication link. The sensor unit 20 is connected to a 900
MHz bidirectional receiver 42. The receiver 42, connected to a modem cable or
DSL 44, sends data in case of alarm to, or transfer data upon request of, a
remote server 16, using an external network accessed either by phone and/or
by the Internet.
[0031] In the embodiment illustrated Figure 5, a plurality of sensing
assemblies 14 are arranged as a 900 MHz network of the local network type for
example, for monitoring a plurality of subjects, for example in a shelter for
elderly in the case of human subjects, or a herd. The resulting network of
sensing assemblies 14 is linked to a central server 160 connected to other
servers 162 used for accessing information provided by the central server 160,
and connected to mobile units 165 by a wireless network.
[0032] These systems allow collecting data related to the dynamics
of the movements of least one subject to be monitored.
[0033] The sensor unit 18 in the upper part of the body of a person
being monitored, or in the front part of an animal being monitored for
example,
typically comprises a high G accelerometer to detect fall of the subject. It
may
further comprise a low G accelerometer to follow the position of movements of
the subject. The frequency, velocity and space orientation of movements of the
upper trunk of the subject is used to yield movement levels which can be
CA 02529545 2010-07-27
7
Graded from null to intense. These parameters may be processed using a
variety of tools such as fuzzy logic, threshold parameters or a weighting
mechanism for example.
[0034] For example, energy levels may be obtained as an average
of the sum of the absolute values of acceleration along the three axes of the
3
axis accelerometer for example, corrected by an off set characterizing the sum
of these accelerations at rest (since accelerometers measure not only the
energy involved during a movement of the person, but also the gravitational
force the person is submitted to), over a number of measurements per second.
This off set is to be taken into account, considering that acceleration
generated
by a moving person is within the range between about -2G and + 2G on a one
second cycle basis, while the gravitational force is generally about 9.8
m/s/s,
i.e. the signal corresponding to the gravitational force may be stronger than
those corresponding to the person's movements.
[0035] In a particular embodiment, the energy level (NE) are thus
obtained as follows: the OG (offset) value of each accelerometer is measured
acceleration along the three axes thereof, by placing each axis in perfect
alignment with the direction of the gravitational force, yielding the values
Xoffset,
Yoffset and Zoffset. In practice, the offset values may be set during the
fabrication
stage of the accelerometer, for example by making the OG value correspond to
an octet value of 128 (the octet 0 being related to -5G, and the octet 255
corresponding to +5G). The energy level NE is calculated as the average over
142 vectorials modules NE; during a period of one second, 142 being the
number of sample by second samples, wherein each vectorial module NE; is
the square root of the sum of the squared corrected values axes. A new NE
value is generated every second and stored. This indicator NE allows
quantifying the intensity of movements thus provides energy levels over
periods
CA 02529545 2010-07-27
8
of 1, 5, 30 minutes or more.
[0036] A second indicator NM, may be used to quantify the
movement levels, describing in particular movements of low amplitude. A
detector having a maximum and a minimum on each axis of the accelerometer
is used, on 1-second periods of time. By subtracting this minimum to that
maximum, the offset is obtained. The NM value is generated and stored
simultaneously with the NE.
[0037] A gain KM for the movement level NM and a gain Ke for the
energy level NE may be defined and used to generate an indicator of
movement level.
[0038] A third indicator, referred to as INC, may be used to identify a
fall event, as detected by an impact sensor, by comparison to an adjustable
threshold. The impact sensor measures a gradient and amplitude of shock
waves related to a fall, typically characterized by 10 waves over 0.25
seconds.
The fall indicator INC may be defined as the sum of absolute values of
amplitudes measured during an event. Typically, an INC of 25% corresponds to
low amplitude impacts, while severe falls are characterized by INC values of
100% and more.
[0039] Such data may further be used to determine sub-levels of
sleeping activity, including sleep phases and intensity, or levels of low-
intensity
activities such as rest or writing process.
[0040] The sensor assembly of the present invention allows
identifying critical levels of activity, as defined according to a target
population
of subjects to be monitored, such as persons suffering from functional
CA 02529545 2010-07-27
9
dependence for example. Critical levels may be set for a range of activities,
including a total lack thereof such as in case of death, breathing rhythms and
apnea, breathing in absence of minimal movement such as in case of coma or
faintness, and hyperactivity.
[0041] A preliminary classification of the persons to be monitored
according to their degree of physical sufficiency allows setting threshold and
control parameters adapted individually to each of these persons and to yield
data all the more representative of the state of each one of them.
[0042] Nycthemeral or circadian analysis may be used to obtain
activity patterns of a subject for time monitoring and identification of
abnormal
or undesirable variations in time of the subject.
[0043] Activity of the subject may be qualitatively assessed, be it
walking, feeding or sleeping for example, by analyzing the data collected by
the
sensing assembly of the system by processing based on neural networks in
combination with fuzzy logics or logic threshold values, depending on the
processing and memory capacity available.
[0044] As described hereinabove, the present system, using at least
one sensing assembly comprising at least one a high G accelerometer, may be
used for monitoring a fall of this subject and a low-G one for monitoring his-
her
activity level. Turning to Figures 6 to 10 of the appended drawings,
embodiments of a method according to the present invention method will now
be described.
[0045] In Figure 6, for example, the sensing assembly provides
analog signals emitted by at least a 3-axes accelerometer and two gyroscopes.
CA 02529545 2005-12-09
These signals are digitized by an analog-to-digital converter. The obtained
data
are placed in a memory stack storage used as a short-term memory, and which
size defines the term in seconds. Low-level algorithms are applied to the data
of the memory stack storage to extract data on the behavior and body posture
5 of the subject being monitored, by generating a property vector from these
data. Such property vector includes a number of parameters as follows (see
Figure 7):
- frequency of the body movement, defining the activity of the monitored
10 subject by the absence of activity, the rest, moderate and active awaken
states;
- trunk position variation: data collected by the accelerometer allow
determining the space orientation of the accelerometer, which is related
to the space orientation (x, y, z) of the monitored subject, through gravity
and knowledge of the location of the accelerometer in relation to the
body of the monitored subject (see Figure 8);
- height of the monitored subject along a z direction, , which may be used
to sort out actual fall events from false alarms by correlating this height
with the position of the body (see Figure 8);
- angular velocity of the trunk of the monitored subject, which may be
used to determine whether a trunk position variation results from a
controlled movement or from an accidental movement;
- shock wave amplitude, characterizing a fall event;
- number of shock waves, which may be used to sort out actual fall events
from false alarms.
[0046] As shown in Figure 8, the position of the monitored subject
includes his-her position along a vertical axis (noted z in the Figure), which
allows assessing the height of an event taking place in a (x,y) plane, i.e.
CA 02529545 2005-12-09
11
determining whether the monitored subject is lying on the floor, lying on top
of a
bed, seated, kneeling down, or standing. Measuring the position along the
vertical axis may be achieved by RFID, ultrasound or using a camera for
example.
[0047] The property vector is analyzed to determine whether the
monitored subject has fallen and to yield indications on the type of
activities the
subject is involved in. Fuzzy logics analysis may for example be used to yield
to
output information, relating to fall and activity respectively.
[0048] Using a sensing assembly comprising a sensor unit in the
upper part of the subject combining a high frequency-low accuracy (high G in
the range of 100G) accelerometer and a low frequency-high accuracy (low G in
the range between 2 and 5 G) accelerometer, allows detecting events such as
impact as well as body posture and fine movements.
[0049] Moreover, the locations of the sensor units of the sensing
assembly relative to the monitored subject's body may be selected to combine
a sensor unit at the wrist of the monitored subject with a sensor unit in the
region of the base of the neck for example. The combination of these sensor
units allows tracking the dynamics of the trunk of the monitored subject while
allowing discarding non-pertinent interferences due to non-significant
movements for example. Furthermore, this combination allows detecting a fall
event while reducing false-alarms generation, since, for example in the case
of
a single wrist sensor unit, even a knock of the hand wearing the wrist sensor
unit on a table for example would be detected as an impact.
[0050] Such a combination of the locations of the sensor units allows
sorting events, by allowing a validation between impacts or movements usually
CA 02529545 2010-07-27
12
of lower amplitude of the trunk of the monitored subject and impacts or
movements usually of higher amplitude of the arm of the monitored subject. It
further allows a qualitative assessment of events, by allowing for example to
identify a movement of the arm alone as a protection movement.
[0051] The sensor unit in the region of the trunk of the monitored
subject may be efficiently connected to the monitored subject without use of
straps, since a neck assembly for example may be used, as described earlier
hereinabove.
[0052] A gyroscope included in a sensor unit located at the base of
the neck or trunk of the monitored subject allows measuring angular
velocities,
i.e. velocity of lateral movements (left to right and right to left) and back
and
forth movements of the trunk of the monitored subject. Moreover, data from
such gyroscope are used to determine threshold values of angular velocities.
[0053] Figure 9 illustrates an embodiment of a neck assembly 120 of
a bolo tie type for locating a sensor unit. A clip 122, besides allowing
adjusting
a length of the assembly 120 around the neck of the person to be monitored,
acts as a balancing weight securing the assembly 120 in place as the
monitored person moves. The region of the assembly 120 placed in the back of
the neck comprises a flexible unit 124 housing sensors, including a 2-5G
accelerometer 128 and a 50G accelerometer 126, and optionally a gyroscope
134, a thermistor 130, and an impedance detector 132 for monitoring wearing
of the assembly. The assembly 120 comprises an RF antenna 138 and a link
136 between the flexible unit 124 and a pendant 100. The pendant 100 houses
a microphone 140, a 3D positioning unit 142, a help summon button 144 and a
battery 146.
CA 02529545 2010-07-27
13
[0054] Figure 10 illustrates an embodiment of a wrist assembly 150
of a wristwatch band type for locating a sensor unit. The bracelet part 154
integrates the sensors, including a 2-5 G accelerometer 156, a 50 G
accelerometer 158, a gyroscope 134, a thermistor 130 and an impedance
detector 164 for monitoring wearing of the assembly. The module 152
comprises a microphone 166, a 3D positioning unit 168, a help summon button
170, a battery 172 and a clock 174.
[0055] Data from a detector for monitoring wearing of the assembly
(132, 164 in Figures 9 and 10 for example) may be integrated with data from
the accelerometers and other sensors so as to further discriminate between
events, between a fall of the monitored person and the monitored person
merely dropping the neck assembly 120 or of the wrist assembly 150 on the
floor for example.
[0056] Such system and method allows identification of critical
activity levels, such as coma states, immobility over a period of time,
breathing
movements interruption, thereby allowing establishing a profile of daily
nycthemeral activities of the monitored subject for example. Such profile may
be used for detecting sudden variations, which may be significant of a decline
in the monitored subject's well being, and provide information concerning the
evolution of parameters of the profile of daily nycthemeral activities of the
monitored subject, weighted according to the initial functional independence
level of the monitored subject to permit assessment of functional independence
variations.
[0057] Acceleration, velocity and/or position signals sampled on at
least one sensing assembly comprising a sensor unit located on the trunk, and,
optionally, a sensor unit located on the wrist of the monitored subject, each
CA 02529545 2010-07-27
14
sensing assembly comprising an accelerometer and optionally a gyroscope
and/or a piezo-film, may be used to provide a representation for the behavior
of
the monitored subject through activity levels (see Figure 12).
[0058] The activity levels are characterized using indicators (such
as NE, NM, INC discussed hereinabove for example) on the body posture of
the monitored subject and of any change in her-his position in her-his
environment, of the velocity and quantity of movements of each part of her-his
body wearing a sensor unit, obtained from processing the acceleration,
velocity
and /or positional signals collected by the sensing assembly. As described
hereinabove, these indicators or property vectors are analyzed to yield the
state, phase of state and activities of the monitored subject, and the
evolution
thereof during a predetermined period of time (see Figure 11).
[0059] As illustrated in Figure 12, the state of the monitored subject
may be assessed between a critical state corresponding to a problem or an
activity level indicating a potentially deficient well being, and a normal
state. In
each case, an absence of movement as indicated by the absence of movement
detection, may be interpreted as a defective system or as the death of the
monitored subject, while slow heart beat, breathing movements and minimal
body movements may be interpreted as representative of a rest phase of the
monitored subject, and body or member movements, as characterized by their
frequency, velocity and orientation, may be evidence of an awareness phase of
the monitored subject. In each of these phases, different activity levels may
then be assessed, from null in the phase of absence of detection), to apnea
(heart beats are detected), coma (breathing is detected), and sleep
(movements are detected) in the rest phase, and to low, moderate and high
activity levels (movements are detected) in the awareness state.
CA 02529545 2005-12-09
[0060] The frequency analysis of signals collected at the base of the
neck of the monitored subject yields a quantitative assessment of the
movement of the monitored subject, including the variations of this quantity
of
movements during short periods of time.
5 [0061] The analysis of signals collected at the wrist of the monitored
subject may be combined to yield a qualitative assessment of this movement
by incorporating angular velocity measurements and positional measurement in
space (x, y, z). The combined analysis of signals emitted in the region of the
base of the neck and of signals emitted in the region of the wrist yield an
10 accurate activity profile and efficient positioning along the vertical axis
(z). This
in turn results in a possible identification of the very room in which an
event
occurs (bathroom, kitchen, bedroom etc...) by cross-correlation, and therefore
to an increasingly efficient monitoring system, since allowance levels may be
pre-determined individually for each room of the subject's habitat,
considering
15 for example that the subject detected lying on her-his bed in her-his-
bedroom
and sleeping is reasonably a normal event.
[0062] Therefore the present system and method allow monitoring a
subject very precisely in relation to her-his individual functional
independence
level as well as his-her environment.
[0063] Although embodiments were illustrated given hereinabove in
relation to a human, the present system and method may efficiently be applied
for monitoring a range of subjects, including for example farm animals,
domestic pets etc...
[0064] Although the present invention has been described
hereinabove by way of embodiments thereof, it may be modified, without
CA 02529545 2005-12-09
16
departing from the nature and teachings of the subject invention as defined in
the appended claims.