U.S. patent application number 13/427738 was filed with the patent office on 2013-04-11 for exercise-based entertainment and game controller to improve health and manage obesity.
This patent application is currently assigned to REGENTS OF THE UNIVERSITY OF CALIFORNIA. The applicant listed for this patent is Navid Amini, Majid Sarrafzadeh. Invention is credited to Navid Amini, Majid Sarrafzadeh.
Application Number | 20130090213 13/427738 |
Document ID | / |
Family ID | 48042437 |
Filed Date | 2013-04-11 |
United States Patent
Application |
20130090213 |
Kind Code |
A1 |
Amini; Navid ; et
al. |
April 11, 2013 |
Exercise-Based Entertainment And Game Controller To Improve Health
And Manage Obesity
Abstract
Monitoring and rewarding of physical activity are carried out
by: (1) receiving a measurement of a physical activity from a
sensor; (2) processing the measurement of the physical activity to
derive a valid extent of the physical activity; and (3) controlling
an entertainment device based on the valid extent of the physical
activity.
Inventors: |
Amini; Navid; (Los Angeles,
CA) ; Sarrafzadeh; Majid; (Anaheim Hills,
CA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Amini; Navid
Sarrafzadeh; Majid |
Los Angeles
Anaheim Hills |
CA
CA |
US
US |
|
|
Assignee: |
REGENTS OF THE UNIVERSITY OF
CALIFORNIA
Oakland
CA
|
Family ID: |
48042437 |
Appl. No.: |
13/427738 |
Filed: |
March 22, 2012 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
61467744 |
Mar 25, 2011 |
|
|
|
Current U.S.
Class: |
482/8 |
Current CPC
Class: |
G16H 20/30 20180101 |
Class at
Publication: |
482/8 |
International
Class: |
A63B 71/00 20060101
A63B071/00 |
Claims
1. A non-transitory computer-readable storage medium, comprising
executable instructions to: receive a first measurement of a
physical activity from a first sensor; process the first
measurement of the physical activity to derive a valid extent of
the physical activity; and control an entertainment device based on
the valid extent of the physical activity.
2. The non-transitory computer-readable storage medium of claim 1,
further comprising executable instructions to receive a second
measurement of the physical activity from a second sensor that is
different from the first sensor, and wherein the executable
instructions to process the first measurement of the physical
activity include executable instructions to derive the valid extent
of the physical activity based on the second measurement of the
physical activity.
3. The non-transitory computer-readable storage medium of claim 2,
wherein the first measurement is a measurement of acceleration, and
the second measurement is a measurement of pressure.
4. The non-transitory computer-readable storage medium of claim 3,
wherein the executable instructions to derive the valid extent of
the physical activity include executable instructions to correlate
the measurement of acceleration with sufficient pressure applied to
at least one area of a foot relative to a threshold pressure
value.
5. The non-transitory computer-readable storage medium of claim 1,
wherein the executable instructions to process the first
measurement of the physical activity include executable
instructions to: derive a physical activity template; and apply the
physical activity template to the first measurement to derive the
valid extent of the physical activity.
6. The non-transitory computer-readable storage medium of claim 5,
wherein the executable instructions to apply the physical activity
template include executable instructions to: derive a
cross-correlation between the physical activity template and the
first measurement; and detect a number of peaks in the
cross-correlation.
7. The non-transitory computer-readable storage medium of claim 1,
further comprising executable instructions to receive a second
measurement of the physical activity from the first sensor, and
wherein the executable instructions to process the first
measurement of the physical activity include executable
instructions to authenticate an identity of a performer of the
physical activity based on the second measurement of the physical
activity.
8. The non-transitory computer-readable storage medium of claim 7,
wherein the first measurement and the second measurement are
measurements of acceleration.
9. The non-transitory computer-readable storage medium of claim 7,
wherein the executable instructions to authenticate the identity of
the performer of the physical activity include executable
instructions to: derive a measured histogram and a template
histogram corresponding to the first measurement and the second
measurement, respectively; and derive a similarity score between
the measured histogram and the template histogram.
10. The non-transitory computer-readable storage medium of claim 1,
wherein the executable instructions to control the entertainment
device include executable instructions to activate the
entertainment device based on the valid extent of the physical
activity.
11. The non-transitory computer-readable storage medium of claim 1,
wherein the executable instructions to control the entertainment
device include executable instructions to allot a time budget for
the entertainment device based on the valid extent of the physical
activity.
12. A system for monitoring and rewarding physical activity,
comprising: a processing unit; and a memory connected to the
processing unit and including executable instructions to: receive
an identification of valid instances of a physical activity by a
user; and control access of the user to an entertainment device
based on the valid instances of the physical activity.
13. The system of claim 12, wherein the memory further includes
executable instructions to: receive a measurement from a sensor
that is applied to the user; and process the measurement to
identify the valid instances of the physical activity.
14. The system of claim 13, wherein the measurement is indicative
of multiple, candidate instances of the physical activity, and the
executable instructions to process the measurement include
executable instructions to identify a subset of the candidate
instances as the valid instances of the physical activity.
15. The system of claim 14, wherein the executable instructions to
process the measurement include executable instructions to compare
the measurement with a physical activity template to identify the
valid instances of the physical activity.
16. The system of claim 14, wherein the measurement is a first
measurement, the sensor is a first sensor, the memory further
includes executable instructions to receive a second measurement
from a second sensor that is applied to the user, and the
executable instructions to process the first measurement include
executable instructions to identify the subset of the candidate
instances as correlated with the second measurement.
17. The system of claim 13, wherein the executable instructions to
process the measurement include executable instructions to
authenticate an identity of the user to whom the sensor is
assigned.
18. The system of claim 17, wherein the executable instructions to
authenticate the identity of the user include executable
instructions to: derive a measured histogram corresponding to the
measurement; and derive a similarity score between the measured
histogram and a template histogram assigned to the user.
19. The system of claim 12, wherein the executable instructions to
control access to the entertainment device include executable
instructions to activate the entertainment device based on the
valid instances of the physical activity.
20. The system of claim 12, wherein the executable instructions to
control access to the entertainment device include executable
instructions to allot a time budget for the entertainment device
based on the valid instances of the physical activity.
Description
CROSS-REFERENCE TO RELATED APPLICATION
[0001] This application claims the benefit of U.S. Provisional
Application Ser. No. 61/467,744 filed on Mar. 25, 2011, the
disclosure of which is incorporated herein by reference in its
entirety.
FIELD OF THE INVENTION
[0002] The invention generally relates to the monitoring of
exercise or other forms of physical activity and, more
particularly, to the monitoring and rewarding of physical
activity.
BACKGROUND
[0003] The detrimental effects of childhood obesity on the health
and lifespan of an individual coupled with its far reaching grip on
today's youth have caused concern that approaches the level of a
nationwide pandemic. A sedentary lifestyle can be a significant
contributor to childhood obesity. On the other hand, exercise can
help children control their weight, and can help to reduce the risk
of illnesses such as high blood pressure, heart disease, and sleep
problems. However, many children fail to exercise because they
excessively spend time doing stationary activities such as playing
video games or watching television.
[0004] It is against this background that a need arose to develop
the apparatus, system, and method described herein.
SUMMARY
[0005] One aspect of the invention relates to a non-transitory
computer-readable storage medium. In one embodiment, the storage
medium includes executable instructions to: (1) receive a
measurement of a physical activity from a sensor; (2) process the
measurement of the physical activity to derive a valid extent of
the physical activity; and (3) control an entertainment device
based on the valid extent of the physical activity.
[0006] Another aspect of the invention relates to a system for
monitoring and rewarding physical activity. In one embodiment, the
system includes: (1) a processing unit; and (2) a memory connected
to the processing unit and including executable instructions to:
(a) receive an identification of valid instances of a physical
activity by a user; and (b) control access of the user to an
entertainment device based on the valid instances of the physical
activity.
[0007] Other aspects and embodiments of the invention are also
contemplated. The foregoing summary and the following detailed
description are not meant to restrict the invention to any
particular embodiment but are merely meant to describe some
embodiments of the invention.
BRIEF DESCRIPTION OF THE DRAWINGS
[0008] For a better understanding of the nature and objects of some
embodiments of the invention, reference should be made to the
following detailed description taken in conjunction with the
accompanying drawings.
[0009] FIG. 1: Block diagram of a system for monitoring and
rewarding children's physical activity, in accordance with one
embodiment of the invention.
[0010] FIG. 2: Block diagram of hardware and software components of
a physical activity monitor, in accordance with one embodiment of
the invention.
[0011] FIG. 3: Applying step-matching to an acceleration signal,
according to an embodiment of the invention.
[0012] FIG. 4: Histogram similarity approach, according to an
embodiment of the invention.
[0013] FIG. 5: A computer configured in accordance with one
embodiment of the invention.
[0014] FIG. 6: A gait cycle divided into four time intervals,
according to an embodiment of the invention.
DETAILED DESCRIPTION
1. General Description of System
[0015] FIG. 1 shows a block diagram of a system for monitoring and
rewarding children's physical activity, in accordance with one
embodiment of the invention. The system allows children to use a
set of entertainment appliances 1 through N, such as a television
set, a video game console, an audio equipment, or another
electronic entertainment device, depending on the amount of
exercise or other physical activity performed by the children. In
one embodiment, the system operates in a substantially automatic
manner. By providing positive reinforcement, the system can provide
the children with a sense of achievement, while addressing concern
by parents about the children's lack of physical activity.
[0016] As shown in FIG. 1, the system includes three main,
interconnected components: (1) a physical activity monitor (or
module) 1; (2) a main host 2 including a computer interface; and
(3) a power controller (or controller module) 3. The power
controller 3 is connected to power outlets 6, which supply power to
the entertainment appliances 1 through N, and the power controller
3 can activate and deactivate the entertainment appliances 1
through N by adjusting the amount of power supplied by the power
outlets 6. In some embodiments, and depending on the amount of
physical activity performed by the children, the power controller 3
can selectively activate (or selectively deactivate) a subset of
the entertainment appliances 1 through N, while a remaining subset
of the entertainment appliances 1 through N is deactivated (or
activated). As shown in FIG. 1, the system also includes a web
server 4, which provides functionalities for kids, parents, and
caregivers, such as sign-up, reporting, behavior monitoring,
diagnostic, and other functionalities. In some embodiments,
physical activity monitors 1 can be directly or indirectly
connected to entertainment appliances such as televisions. In such
embodiments, the power controller 3 and the main host 2 can be
incorporated in an entertainment appliance. Moreover, data uploads
and other communications with the web server 4 can be directly
performed by the entertainment appliance itself.
[0017] FIG. 2 shows a block diagram of hardware (a) and software
components (b) of the physical activity monitor 1, in accordance
with one embodiment of the invention. The physical activity monitor
1 can measure and process data associated with a physical activity,
such as acceleration data in one or more spatial dimensions as a
function of time.
[0018] As shown in FIG. 2, the physical activity monitor 1 includes
a processor 9 (e.g., a central processing unit (CPU)), which is
connected to a storage medium 10 and a set of sensors 7. The set of
sensors 7 can include a single sensor or a combination of sensors
of the same type or different types, which measure a physical
activity such as walking, running, load carrying, or jumping. The
set of sensors 7 can be included in the physical activity monitor
1, or can be separate from and connected to the physical activity
monitor 1. The physical activity monitor 1 can be physically
attached to, or carried by, a child (or another performer of a
physical activity), and can be connected to the main host 2 through
a wireless or a wired connection of a communication unit 8. The
physical activity monitor 1 can include a pedometer, for example.
Alternatively or in addition, the physical activity monitor 1 can
include other types of sensors, such as a heart-rate monitor or
pressure sensors. The physical activity monitor 1 also can be in
the form of a software application on a smart phone that includes
an accelerometer. For example, the application can incorporate
computer code to detect physical activity based on data measured by
an accelerometer built in or connected to the smart phone. The
application can be developed for a particular operating system of
the smart phone, such as Android.
[0019] In one embodiment, the physical activity monitor 1 can
identify or recognize the type of exercise being carried out by a
child. For example, the physical activity monitor 1 can be
configured to distinguish between different activities such as
walking uphill or downhill, walking on a level surface, running,
heavy load carrying, and so forth. The physical activity monitor 1
also can be configured to distinguish between different types of
environments in which a physical activity is performed, such as the
type surface including walking on grass, uneven ground, gravel,
sand, carpet, and so forth. The recognition of the type of activity
and the type of environment can be carried out in accordance with
supervised techniques or unsupervised techniques. For example,
certain aspects of an unsupervised technique for exercise
recognition is set forth in U.S. Provisional Application Ser. No.
61/448,602 filed on Mar. 2, 2011, the disclosure of which is
incorporated herein by reference in its entirety. Such function can
be carried out by a module 13 stored or residing in the storage
medium 10. Alternatively or in addition, either of, or both, the
main host 2 and the web server 4 can perform such function by
processing data measured by the physical activity monitor 1.
[0020] In one embodiment, the physical activity monitor 1 can
calculate or derive parameters indicative of an extent of a
physical activity, such as distance traveled, duration of exercise,
intensity of exercise (e.g., pace of a walk or run), and calories
or energy burned as a result of exercise. The calculation of such
parameters can be based on the type of exercise that is performed,
the type of environment in which the exercise is performed, or
both. For example, the physical activity monitor 1 can use the
notion of MET (Metabolic Equivalent of Task) to derive the number
of calories burned by a child. The calculation of exercise
parameters can be carried out by a module 14 stored or residing in
the storage medium 10. Alternatively or in addition, either of, or
both, the main host 2 and the web server 4 can perform such
function by processing data measured by the physical activity
monitor 1.
[0021] In one embodiment, the physical activity monitor 1 can
perform processing of measurements of physical activity by the set
of sensors 7 to reduce or minimize the vulnerability of the system
to false positives and cheating. The processing of measurements to
mitigate against false positives and cheating can be carried out by
modules 11 and 12 stored or residing in the storage medium 10.
Alternatively or in addition, either of, or both, the main host 2
and the web server 4 can perform such function by processing data
measured and provided by the physical activity monitor 1.
[0022] In one embodiment, the physical activity monitor 1 can
validate, or determine a valid extent of, measurements of a
physical activity made by the set of sensors 7. A measurement of a
physical activity can be invalid (or have a valid extent of zero).
For example, if some form of cheating has occurred, such as when a
first child attaches the physical activity monitor 1 to a second
child who performs an exercise, a measurement of the exercise can
be deemed invalid. In another example, a measurement of a physical
activity, such as a step count, can be deemed invalid because other
types of physical activities, such as shaking of a pedometer, can
be mistakenly interpreted by the pedometer as steps. Also, a
measurement of a physical activity can be substantially valid (or
have a positive valid extent). For example, the physical activity
monitor 1, through processing of pressure data or acceleration
data, can validate or determine a valid (accurate) extent of a step
count associated with the pressure data or acceleration data
resulting from a walking or running activity.
[0023] Referring back to FIG. 1, the main host 2 can reside in, or
correspond to, a parent's computer. When a child activates the
physical activity monitor 1, a software component of the main host
2 can retrieve data from the physical activity monitor 1 and store
that data in the parent's computer. The data from the physical
activity monitor 1 can be transmitted to either of, or both, the
main host 2 (through a home network via Wi-Fi or LAN, or using
protocols such as Bluetooth, Zigbee, or other similar protocols)
and the web server 4 through the Internet (such as via cellular
communications). In this manner, a database of the main host 2 and
a database of the web server 4 can be updated with the latest
information regarding the child's daily activity. In one
embodiment, these databases can be synchronized even if, for
example, cellular coverage is not available or the child is not in
the vicinity of the main host 2.
[0024] Based on a valid extent of a physical activity, the main
host 2 can issue a command to the power controller 3, which
activates (or deactivates) one or more of the entertainment
appliances 1 through N in accordance with the command. In this
manner, the main host 2, in combination with the power controller
3, can control operation of the entertainment appliances 1 through
N as a reward for physical activity, while mitigating against false
positives and cheating. In one embodiment, the power controller 3
can be integrated into a home automation system that controls the
power outlets 6 in a wireless fashion or via underlying power
lines. Based on a physical activity level of a child, the main host
2 can allot a time budget for one or more of the power outlets 6
corresponding to specific appliances. In this embodiment, the power
controller 3 can deactivate a corresponding appliance when the time
budget expires, so that the child is forced to leave the appliance.
In this manner, the main host 2, in combination with the power
controller 3, can control a child's access to the functionality of
the corresponding appliance. To do so, the main host 2 can issue a
command to the power controller 3, and the power controller can
transmit a radiofrequency (RF) signal to activate (or deactivate)
one or more of the power outlets 6.
[0025] Alternatively or in addition, a controller module can be
included as a software application that interacts with
entertainment applications residing in a child's computer 5, such
as video games. If a physical activity by the child has a
sufficient valid extent, the child can be rewarded with a stronger
avatar for local or web-based games on the child's computer 5, or
can be rewarded, based on the child's exercise records, with other
types of visual feedback incentives through interaction with the
child's computer 5. These additional types of incentives, in
addition to control of access to the entertainment appliances 1
through N, can further persuade the child to engage in healthy
physical activity.
[0026] In one embodiment, a mapping between a valid extent of a
physical activity by a child and an allotted time budget (or amount
of another type of incentive described above) can be adjusted or
otherwise configured. For example, the time budget can linearly
increase as a function of the valid extent of the physical activity
by the child. Alternatively or in addition, the time budget can
increase as a step function. For example, the child can receive no
time budget until the valid extent of the physical activity by the
child exceeds or reaches some minimum value. Allocation of the time
budget can be up to a maximum value that a child can receive per
period of time, such as per day. In one embodiment, the mapping
between the valid extent of the physical activity by the child and
the time budget (amount of another type of incentive described
above) can vary across different types of appliances and across
different applications, such as video games. In one embodiment, the
main host 2 has a recommended mode that calculates a time budget
based on characteristics such as age, sex, height, and weight. It
is contemplated that adults also can benefit from the system in
addition to children.
2. Determining Valid Extent of Physical Activity with a Second Type
of Sensor
[0027] This section and the following sections describe the
processing of measurements of a physical activity to mitigate
against false positives and cheating. As previously described, in
one embodiment, a measurement of a physical activity, such as a
step count for walking or running, can be deemed invalid because
other types of physical activities, such as shaking of a pedometer,
can be mistakenly interpreted by the pedometer as steps.
[0028] In one embodiment, a first sensor can be a pedometer
attached to a person's clothing, placed in the person's pocket, or
embedded in the person's shoe. The first sensor can provide a first
measurement of a physical activity such as walking or running. In
some instances, a significant fraction of detected steps can be a
result of false positives stemming from intentional or
unintentional movements along a vertical axis. To combat this
issue, a second sensor of a different type from the first sensor
can provide a second measurement of the physical activity. The
first sensor and the second sensor can be included in the physical
activity monitor 1, or can be separate sensors that communicate
with the physical activity monitor 1.
[0029] In one embodiment, the second sensor can include a first
pressure sensor located in a first area of a shoe insole
corresponding to a heel area of a foot, and a second pressure
sensor located in a second area of the shoe insole corresponding to
a ball area of the foot. As indicated by FIG. 6, in each gait cycle
for walking, there is a time interval when a considerable amount of
pressure is applied to the heel area, namely a heel strike. Also,
there is a time interval when a considerable amount of pressure is
applied to the ball of the foot, namely right before a toe off.
Therefore, checking an output of the pressure sensors at the right
time (or checking a pattern or time variation of the output versus
an expected gait cycle) can be used to verify that a recent
detected step is indeed an actual step, rather than some other type
of physical activity. It is worth noting that similar strategies
can be applied for running (with a potentially different pattern)
to decrease the number of false positives.
[0030] Alternatively or in addition, the second sensor can include
additional pressure sensors. For example, the second sensor can
include three, four, or more pressure sensors located in the heel
area, the ball area, and other areas of the foot.
[0031] For the example set forth in FIG. 6, a pressure threshold
for the first pressure sensor located in the heel area of the shoe
insole can be configured as a value T1=(weight of person (kg)g
(N/Kg))/(50 (cm.sup.2)) N/cm.sup.2. If a pressure detected by the
first pressure sensor exceeds or reaches T1, then a heel strike
(corresponding to a "1" in the leftmost bit of the two-bit pairs
shown in FIG. 6) is detected. Otherwise a "0" is detected by the
first pressure sensor. Similarly, a pressure threshold for the
second pressure sensor located in the ball area of the shoe insole
can be configured as a value T2=(weight of person (kg).times.g
(N/Kg))/(50 (cm.sup.2)) N/cm.sup.2. If the pressure detected by the
second pressure sensor exceeds or reaches T2, then a ball strike
(corresponding to a "1" in the rightmost bit of the two-bit pairs
shown in FIG. 6) is detected. Otherwise a "0" is detected by the
second pressure sensor. The value of 50 cm.sup.2 for T1 and T2
represents a typical value of the heel area or the ball area, and
can be adjusted or otherwise configured for a particular
person.
[0032] The data measured by the pressure sensors in the shoe insole
can be valuable when the physical activity monitor 1 uses a typical
pedometer to measure a physical activity. In this case, to prevent
false positives (e.g., as a result of the pedometer being
intentionally or unintentionally shaken), recent pressure data is
checked to see if a considerable amount of pressure has been
applied to at least one of the heel area and the ball area. In one
embodiment, the physical activity monitor 1 validates, or
determines a valid extent of, data measured by the pedometer for a
physical activity based on data measured by the pressure sensors
for the same physical activity. Alternatively or in addition,
either of, or both, the main host 2 and the web server 4 can
perform such function by processing data measured by the physical
activity monitor 1.
[0033] To determine a valid extent of a first measurement by the
first sensor (such as a pedometer), data from the first pressure
sensor and the second pressure sensor are checked for at least a
subset of possible (or candidate) steps detected by the pedometer.
For example, if output of the pedometer is active (e.g., a possible
step is detected by the pedometer), then recent pressure data
measured by the first pressure sensor (heel area) and the second
pressure sensor (ball area) are checked for a "01" pattern, a "11"
pattern, or a "10" pattern (as shown in FIG. 6). In one embodiment,
the recent pressure data is pressure data measured in a time period
of one second or less prior to the detection of the possible step
by the pedometer. Note that a time interval between a heel strike
and a ball strike is typically less than 0.5 seconds, even for a
slow walk. The detection of one of these patterns (or a sequence of
these patterns) validates the detected step by the pedometer, as
the step detected by the pedometer correlates to a recent
application of sufficient pressure to at least one of the pressure
sensors. In one embodiment, the step detected by the pedometer can
be deemed valid based on the detection of any of the "01", "11", or
"10" patterns in the recent time period.
[0034] In one embodiment, each possible step detected by the first
sensor (such as a pedometer) can be checked against data from the
second sensor (such as first and second pressure sensors).
Alternatively, a subset of possible steps detected by the first
sensor can be checked against data from the second sensor. The
number of detected steps can be corrected or otherwise adjusted
based on a percentage of steps that were detected correctly by the
first sensor. For example, if 90% of the detected steps by the
first sensor are validated based on the second sensor, then a valid
extent of the step data can be 90% of the step data measured by the
first sensor. An entertainment appliance, such as an electronic
entertainment device, can then be controlled based on this valid
extent of the step data.
3. Determining Valid Extent of Physical Activity with a Physical
Activity Template
[0035] As previously described, in one embodiment, processing by
the physical activity monitor 1 can determine a valid (accurate)
extent of a physical activity, such as walking or running. For
example, the physical activity monitor 1 can determine a valid step
count based on acceleration data. This processing can be
facilitated by determining a physical activity template, and by
applying the physical activity template to measured acceleration
data to count a number of instances of the activity that have
occurred, such as a number of steps for a walking or running
activity.
[0036] In one embodiment, an accelerometer can be included in a
sensor that measures a physical activity such as walking or
running. The accelerometer can be included in the physical activity
monitor 1. A template matching approach can reliably measure a
number of steps taken by a person. The template matching approach
is based on a template, which represents a typical step cycle. More
generally, a template can identify or represent an aspect of a
physical activity of a particular type, such as a walking step, a
running step, and so forth.
[0037] In one embodiment, an acceleration data signal can be
divided into multiple data blocks having a particular duration in
time, such as data blocks of about 10 seconds. FIG. 3(a) depicts a
10-second data block measured from a left foot along the x-axis,
containing six step cycles. A low-pass filter with a cutoff
frequency of about 20 Hz is applied to the signal to facilitate
further processing, as shown in FIG. 3(b).
[0038] The template matching approach can then examine whether any
physical activity template already exists, such as residing in the
storage medium 10. The physical activity template can be derived
from measurements of the same type of physical activity during a
training period, which can be a time duration of at least about 10
seconds, such as a time duration of about 1 minute. Alternatively
or in addition, a first step cycle (e.g., a time duration at the
beginning of the measured data signal, such as an initial 10 second
period of the measured data signal) can be extracted as a temporary
template. Another portion of the measured data signal also can be
extracted as the temporary template. In one embodiment, the
template can be derived by the physical activity monitor 1.
Alternatively or in addition, the template can be derived by either
of, or both, the main host 2 and the web server 4.
[0039] An instance of a physical activity (or another event) can be
detected in a measured data signal when there is a sufficient
degree of similarity between the measured data signal and the
template. In one embodiment, the template is slid across the entire
or a portion of the data signal, and a normalized cross-correlation
is calculated between the template and the measured data signal
(see FIG. 3(c)). The normalized cross-correlation indicates the
similarity between the template and the measured data signal, as
set forth in
R N [ k ] = X , Y X Y = R XY ( k ) R XX ( 0 ) R YY ( 0 ) , ( 1 )
##EQU00001##
[0040] In equation (1), X represents the template, Y represents the
measured data signal, k is an index representing a time lag, <X,
Y> is the inner product of X and Y, .parallel.X.parallel. is the
norm of X, .parallel.Y.parallel. is the norm of Y, R.sub.XY(k) is
the cross-correlation of X and Y for arbitrary k, R.sub.XX(0) is
the auto-correlation of X at zero lag, and R.sub.YY(0) is the
auto-correlation of Y at zero lag. In one embodiment, the
cross-correlation can be derived by the physical activity monitor
1. Alternatively or in addition, the cross-correlation can be
derived by either of, or both, the main host 2 and the web server
4.
[0041] In one embodiment, a maximum value for the normalized
cross-correlation is 1 for absolute identity, which allows a
uniform threshold to be set for all data despite varying
amplitudes. Peaks in the normalized cross-correlation in FIG. 3(c)
can indicate significant similarity between the template and the
data signal segment, and thus the occurrence of an event (e.g., a
step). An interval during which the cross-correlation exceeds or
reaches a threshold T (e.g., 0.4) can be defined as a peak
searching interval. Such intervals are marked with a solid line in
FIG. 3(c) and with dashed lines in FIG. 3(d). Local maxima falling
within the peak searching intervals in the filtered signal are
marked as fiducial points of steps, and denoted by asterisks in
FIG. 3(d). Using the template matching approach, initial positive
peaks of steps, which occur when feet lift off the ground, are
detected, and the number of steps is correspondingly counted.
[0042] In one embodiment, a physical activity template can be
derived based on an average of multiple step cycles, which can be a
more representative template than a temporary template. For
example, in FIG. 3, six step cycles can be detected based on
locating the peaks in the data block. These six step cycles can be
aligned based on the location of their peaks and averaged together
to derive a new template, which can be applied for further
processing of subsequent data blocks. Alternatively or in addition,
multiple step cycles in a longer time duration, such as about 1
minute, can be detected based on locating peaks in an initial data
block of a shorter time duration. These multiple step cycles can be
aligned based on the location of their peaks and averaged together
to derive a new template, which can be applied for further
processing.
[0043] Step cycles can be detected based on various techniques,
such as (a) finding zeros of a signal; (b) computing the signal's
energy; or (c) using the concept of salience used in speech
processing. In one embodiment, given that a signal from a sensor,
such as an accelerometer, is mixed with noise, the third technique
can yield a higher accuracy. The salience of a given data sample
can be defined as the length of the longest interval over which the
sample is a maximum. The term salience vector denotes a signal
containing the salience of each sample in an original, input
signal. As a result of feet striking the ground while walking, a
start of each walking cycle typically has a large salience.
Therefore, cycles can be detected by locating such distinct
points.
[0044] In one embodiment, the number of steps in a data block, such
as a block of acceleration data for walking, can be derived as
follows. First, a salience of each sample of the accelerometer data
in an input signal r can be found, and a corresponding salience
vector, s, can be created. Then, the vector u=(rs)/max(s) is
computed, where "" represents an element-wise multiplication. This
transformation makes the peaks of r more pronounced and attenuates
other elements of r. Then, elements in u beyond or reaching a
certain threshold are extracted as potential cycle indices. Then, a
difference d between these results is computed. Any results
differing by a single sample can be discarded. Based on an original
signal assumption, an average (or mean) of the difference d can
correspond to be an average (or mean) of a cycle length. Then, the
difference d is normalized around its mean, and the indices which
fall within the threshold are extracted. These are the cycle
starting and ending points. The number of steps in the data block
is then derived as: (number of extracted indices -1).
[0045] If a template is already present, peaks can be detected
using the techniques stated above. Before a next data block is
processed using the same template, a determination can be made as
to whether the template will be updated. A step signal may change
dynamically with time; accordingly, the template may not accurately
represent a current step signal. In one embodiment, if major peaks
in a normalized cross-correlation are lower than 0.55 (or another
threshold), a new template can be derived using step cycles in a
current data block. Otherwise, peak detection is carried out in the
current data block using the existing template.
4. Determining Valid Extent of Physical Activity Through User
Authentication
[0046] As previously described, in one embodiment, a measurement of
a physical activity by a first sensor, such as a step count for
walking or running, can be deemed invalid because a detected
identity of a performer of the physical activity (as measured by
the first sensor) does not correspond to an identity associated
with, or assigned to, the first sensor. In one embodiment, an
unique identifier can be assigned to a given physical activity
monitor 1 (which can include the first sensor) to associate the
physical activity monitor 1 with a given person. If exercise
performed by a different person is measured by the physical
activity monitor 1, then the system can detect the exercise as
cheating.
[0047] To detect this type of cheating, the system can determine
whether the physical activity monitor 1 is being carried by the
person to whom the physical activity monitor 1 is assigned, or by a
different person. In one embodiment, such determination is carried
out through a histogram comparison to derive a similarity
score.
[0048] A first histogram, namely a template histogram, can be
derived that identifies or represents characteristics of a first
instance of a physical activity of a particular type, such as
walking or running, where it is known that the first instance of
the physical activity is performed by the person to whom the
physical activity monitor 1 is assigned. The template histogram can
be derived during a training period. In one embodiment, the
template histogram can be derived by the physical activity monitor
1. Alternatively or in addition, the template histogram can be
determined by either of, or both, the main host 2 and the web
server 4.
[0049] A second histogram, namely a measured histogram to be
validated, can be derived that identifies or represents
characteristics of a second instance of the same type of physical
activity, where it is desired to determine whether the second
instance of the physical activity is performed by the person to
whom the physical activity monitor 1 is assigned, or by a different
person. In one embodiment, the measured histogram can be derived by
the physical activity monitor 1. Alternatively or in addition, the
measured histogram can be determined by either of, or both, the
main host 2 and the web server 4.
[0050] Based on a comparison of the template histogram to the
measured histogram, a valid extent of measurements of the second
instance of the physical activity can be derived. In particular,
the system can determine whether the second instance of the
physical activity is performed by the person to whom the physical
activity monitor 1 is assigned (if the template histogram is
sufficiently similar to the measured histogram), or by a different
person. The comparison of the template histogram to the measured
histogram can occur during a verification period separate from the
training period.
[0051] Certain factors can affect the accuracy of a histogram
similarity determination for user authentication. First, each
histogram can include a number of bins (or resolution) that can
correspond to the number of different recognizable outputs that a
sensor (such as an accelerometer) can provide. Second, each
histogram can include a number of observations (data points) that
can correspond to a sampling rate of the sensor times a duration of
the physical activity, such as each step.
[0052] In one embodiment, the template histogram is derived by
measuring a first instance of a physical activity of a known type
(such as walking or running) and performed by a known person to
whom the physical activity monitor 1 is assigned. A duration of the
first instance of the physical activity can be long enough to
include at least several cycles of the physical activity, such as
multiple steps or strides. For example, the duration of the first
instance of the physical activity can be at least about ten
seconds, and can be about one minute or more, such as up to about
five minutes, about ten minutes, or more.
[0053] In one embodiment, the measured histogram is determined by
measuring a second instance of the same type of physical activity.
A duration of the second instance of the physical activity can be
long enough to include at least several cycles of the physical
activity, such as multiple steps or strides. For example, the
duration of the second instance of the physical activity can be at
least about ten seconds, and can be about one minute or more, such
as up to about five minutes, about ten minutes, or more. The
duration of the second instance of the physical activity can be the
same as, or different from, the duration of the first instance of
the physical activity.
[0054] In one embodiment, an output range of a sensor, such as an
accelerometer, is divided into n intervals (typically 100 bins for
sampling rates no more than about 100 Hz). Each interval can
correspond to a bin of at least one of the template histogram and
the measured histogram. With regard to the template histogram, each
data point from the first instance of the physical activity can be
included in the template histogram. For a sampling rate of about
100 Hz, all data points can be taken into consideration, namely the
number of observations is equal to the number of data points.
[0055] Various metrics can be used to determine a similarity
between the template histogram and the measured histogram. In one
embodiment, an absolute distance metric can be used to derive a
similarity score between these histograms. The template histogram
and the measured histogram can be normalized prior to their
comparison, and an absolute distance can be derived as set forth in
equation (2):
dist ( x , y ) = i = 1 n x i - y i ( 2 ) ##EQU00002##
[0056] Here, x.sub.i is the probability of a data point residing in
bin i of the normalized template histogram, and y.sub.i is the
probability of a data point residing in bin i of the normalized
measured histogram. In one embodiment, this distance value
represents a similarity score between two acceleration signals.
This metric is both computationally streamlined and effective at
measuring similarity between histograms for authenticating the
identity of a person performing various types of physical
activities.
[0057] If a physical activity is performed by the same person to
whom the physical activity monitor 1 is assigned, the distance
value is typically smaller than a resulting distance value when the
physical activity is performed by an impostor. A combined
acceleration signal, namely an acceleration signal that combines
accelerations along multiple axes (e.g., all three axes), can yield
further improvements in authenticating the identity of a performer
of the physical activity. In one embodiment, operations involved in
comparing two gait samples using the histogram similarity approach
are visualized in FIG. 4.
[0058] FIG. 5 illustrates a computer 800 configured in accordance
with one embodiment of the invention. The computer 800 includes a
CPU 802 connected to a bus 806. Input/output (I/O) devices 804 are
also connected to the bus 806, and can include a keyboard, mouse,
display, and the like. A computer program for determining a valid
extent of measurements as described above is stored in a memory
808, which is also connected to the bus 806. Computer programs
providing functionality corresponding to at least one of the
physical activity controller 1, the main host 2, the power
controller 3, and the web server 4 can also be stored in the memory
808.
[0059] An embodiment of the invention relates to a non-transitory
computer-readable storage medium having computer code thereon for
performing various computer-implemented operations. The term
"computer-readable storage medium" is used herein to include any
medium that is capable of storing or encoding a sequence of
instructions or computer codes for performing the operations
described herein. The media and computer code may be those
specially designed and constructed for the purposes of the
invention, or they may be of the kind well known and available to
those having skill in the computer software arts. Examples of
computer-readable storage media include, but are not limited to:
magnetic media such as hard disks, floppy disks, and magnetic tape;
optical media such as CD-ROMs and holographic devices;
magneto-optical media such as floptical disks; and hardware devices
that are specially configured to store and execute program code,
such as application-specific integrated circuits ("ASICs"),
programmable logic devices ("PLDs"), and ROM and RAM devices.
Examples of computer code include machine code, such as produced by
a compiler, and files containing higher-level code that are
executed by a computer using an interpreter or a compiler. For
example, an embodiment of the invention may be implemented using
Java, C++, or other object-oriented programming language and
development tools. Additional examples of computer code include
encrypted code and compressed code. Moreover, an embodiment of the
invention may be downloaded as a computer program product, which
may be transferred from a remote computer (e.g., a server computer)
to a requesting computer (e.g., a client computer or a different
server computer) via a transmission channel. Another embodiment of
the invention may be implemented in hardwired circuitry in place
of, or in combination with, machine-executable software
instructions.
[0060] While the invention has been described with reference to the
specific embodiments thereof, it should be understood by those
skilled in the art that various changes may be made and equivalents
may be substituted without departing from the true spirit and scope
of the invention as defined by the appended claims. In addition,
many modifications may be made to adapt a particular situation,
material, composition of matter, method, operation or operations,
to the objective, spirit and scope of the invention. All such
modifications are intended to be within the scope of the claims
appended hereto. In particular, while certain methods may have been
described with reference to particular operations performed in a
particular order, it will be understood that these operations may
be combined, sub-divided, or re-ordered to form an equivalent
method without departing from the teachings of the invention.
Accordingly, unless specifically indicated herein, the order and
grouping of the operations is not a limitation of the
invention.
* * * * *