`(12) Patent Application Publication (10) Pub. No.: US 2010/0217533 A1
`
`
` Nadkarni et al. (43) Pub. Date: Aug. 26, 2010
`
`US 20100217533A1
`
`(54)
`
`IDENTIFYING A TYPE OF MOTION OF AN
`OBJECT
`
`(75)
`
`Inventors:
`
`Vij ay Nadkarni, San Jose, CA
`(US); Jeetendra Jangle, Fremont,
`CA (US); John Bentley, Santa
`Clara, CA (US); Umang Salgia,
`Nigadi (IN)
`
`Correspondence Address;
`Law Office of Brian Short
`PO. Box 641867
`
`San Jose, CA 95164-1867 (US)
`
`(73) Assignee:
`
`LABURNUM NETWORKS,
`INC., San Jose, CA (US)
`
`(21) APPL N05
`
`12/560,069
`
`(22)
`
`Filed:
`
`Sep. 15, 2009
`
`Related US. Application Data
`(60) Provisional application No. 61/208,344, filed on Feb.
`23, 2009.
`
`_
`_
`_
`_
`Publlcatlon Class1ficat10n
`
`(51)
`
`Int. Cl.
`(2006.01)
`G01P 15/00
`(2006.01 )
`G06F 19/00
`(2006.01)
`G06F 1 7/18
`(52) US. Cl. ............................ 702/19; 702/141; 702/179
`(57)
`ABSTRACT
`
`A method of identifying a type of motion of an animate or
`inanimate object is disclosed. The method includes generat-
`mg an acceleration Signature based on the sensed acceleratlon
`of the object. The acceleration signature is matched with at
`least one of a plurality of stored acceleration signatures,
`wherein each stored acceleration signatures corresponds with
`type of motion. The type of motion of the object is identified
`based on the statistical matching or exact matching of the
`acceleration signature.
`
`Generating an acceleration signature based on the sensed acceleration of the object
`
`61
`
`60
`
`Matching the acceleration signature with at least one of a plurality of stored
`acceleration signatures, wherein each stored acceleration signatures corresponds with
`type of motion
`
`Identifying the type of motion of the object based on the statistical matching or exact
`matching of the acceleration signature
`
`
`
`63
`
`1
`
`APPLE 1008
`
`APPLE 1008
`
`1
`
`
`
`Patent Application Publication
`
`Aug. 26, 2010 Sheet 1 0f 9
`
`US 2010/0217533 A1
`
`FiGURE ”i
`
`2
`
`
`
`Patent Application Publication
`
`Aug. 26, 2010 Sheet 2 0f 9
`
`US 2010/0217533 A1
`
`Slow Falling and lying down summersault
`
`
`
`
`
`
`m.r‘" Acceleration1’
`
`350
`
`HGUREZA
`
`
`Slippipggand. falling.99,,,b%9¥s9n la beggcy surface (air mattress).
`
`Acceleration
`
`
`
`
`
`
`
`X Axis250 350
`
`HGUREZB
`
`3
`
`
`
`Patent Application Publication
`
`Aug. 26, 2010 Sheet 3 0f 9
`
`US 2010/0217533 A1
`
`
`539911ffi9¢,With,k¥1§¢§fl§¥§d A ..
`
`AcceIeration
`
`
`
`FIGURE 2C
`
`4
`
`
`
`Patent Application Publication
`
`Aug. 26, 2010 Sheet 4 0f 9
`
`US 2010/0217533 A1
`
`
`
`312
`
`Ace.
`
`314
`
`Ace,
`316
`
`
`
`
`
`Q
`
`z mAID
`320
`
`Compare
`
`330
`
`Library
`
`34“
`
`Audio.
`360
`
`370
`
`Controller
`
`i
`2
`
`Communication
`
`
`
`:
`1
`
` i'
`
`5 WI
`
`'—,>
`
`FIGURE 3
`
`5
`
`
`
`Patent Application Publication
`
`Aug. 26, 2010 Sheet 5 0f 9
`
`US 2010/0217533 A1
`
`Monitoring an activity of a person the motion detection device is attached
`ilQ
`
`Performing instantaneous computations over raw signals to compute atomic motions
`along with gravity vector and tilt vector
`5.2.9
`
`430
`
`Applying series of digital filters to remove noise in the atomic motions data
`
`\
`
`WM
`
`Mertorming state analysis on series of atomic data samples“
`m
`NRN M
`
`Perioidic State Analysis
`44_5
`
`Transcient State Analysis
`@
`
`W A
`
`Formation of macro motion signatures
`5.6.9.
`
`learning system providing the right model for the user from a set of model
`m
`
`4_8_Q
`
`Pre—building a motion database of motion libraries
`
`FIGURE 4
`
`6
`
`
`
`Patent Application Publication
`
`Aug. 26, 2010 Sheet 6 0f 9
`
`US 2010/0217533 A1
`
`
`
`Another large
`acceleration
`
`
`
`
`
`
` Deviation in Motion
`
`Monitor
`Pattern Detected,
`
`recorded and
`Activity
`5.19
`
`reported
`
`
`m
`
`
`Fall Report
`
`
`Large acceleration
`Acknowledged
`
`
`5%
`threshold exceeded,
`
`
`
`record audio
`
` Q9
`
`
`
`Monitor
`
`
`
`Fall
`Prob. Fail
`
`
`magnitude detected
`Repo rted
`fl
`5.2.5
`
`
`52,5:
`
`
`Short Period of
`
`
`Inactivity
`
`
`iii
`
` Send Alert, recording
`
`and analysis sent
`
`
`fl
`
`Normal Movement
`Prob. Fall
`
`
`Detected, Audio Rec.
`Detected
`
`Stopped
`.5.£LQ
`
` 54—5
`
`Fall
`
`
`Detected
`
`
`5.6.5.
`Period of Inactivity
`
`fl
`
`
`
`
`
`
`
`Analysis of motion,
`positions indicates
`normal activity
`@
`
`
`Analysis
`§§§
`
`
`Analysis of motion,
`positions indicates a
`
`fall occurred
`
`
`5%
`
`FIGURE 5
`
`7
`
`
`
`Patent Application Publication
`
`Aug. 26, 2010 Sheet 7 0f 9
`
`US 2010/0217533 A1
`
`Generating an acceleration signature based on the sensed acceleration of the object
`
`610
`
`60
`
`Matching the acceleration signature with at least one of a plurality of stored
`acceleration signatures, wherein each stored acceleration signatures corresponds with
`type of motion
`
`identifying the type of motion of the object based on the statistical matching or exact
`matching of the acceleration signature
`
`
`
`30
`
`FIGURE 6
`
`8
`
`
`
`Patent Application Publication
`
`Aug. 26, 2010 Sheet 8 0f 9
`
`US 2010/0217533 A1
`
`
`
`The motion detection device determining what network connections are available to
`the motion detection device
`
`10
`
`
`
`
`
`
`The motion detection device distributing at least some of the acceleration signature
`matching processing if processing capability is available to the motion detection device
`
`though available network connections
`m
`
`
`
`
`
`FIGURE 7
`
`9
`
`
`
`Patent Application Publication
`
`Aug. 26, 2010 Sheet 9 of 9
`
`US 2010/0217533 A1
`
` Processor
`
`\
`
`Noflllilwe’fiuork
`
`Available
`
`810
`
`Blue Tooth
`
`Processor
`
`8312
`
`Cellular
`820
`
`\‘
`A”
`
`350
`
`
`
`Processor
`
`
`
`Home
`Base Station
`840
`
`FIGURE 8
`
`10
`
`10
`
`
`
`US 2010/0217533 A1
`
`Aug. 26, 2010
`
`IDENTIFYING A TYPE OF MOTION OF AN
`OBJECT
`
`RELATED APPLICATIONS
`
`[0001] This patent application claims priority to US. pro-
`visional patent application Ser. No. 61/208,344 filed on Feb.
`23, 2009 which is incorporated by reference.
`
`FIELD OF THE DESCRIBED EMBODIMENTS
`
`[0002] The described embodiments relate generally to
`motion detecting. More particularly, the described embodi-
`ments relate to a method and apparatus for identifying a type
`of motion of an animate or inanimate object.
`
`BACKGROUND
`
`[0003] There is an increasing need for remote monitoring
`of individuals, animals and inanimate objects in their daily or
`natural habitats. Many seniors live independently and need to
`have their safety and wellness trackcd. A large percentage of
`society is fitness conscious, and desire to have, for example,
`workouts and exercise regimen assessed. Public safety offic-
`ers, such as police and firemen, encounter hazardous situa-
`tions on a frequent basis, and need their movements, activities
`and location to be mapped out precisely.
`[0004] The value in such knowledge is enormous. Physi-
`cians, for example, like to know their patients sleeping pat-
`terns so they can treat sleep disorders. A senior living inde-
`pendently wants peace of mind that if he has a fall it will be
`detected automatically and help summoned immediately. A
`fitness enthusiast wants to track her daily workout routine,
`capturing the various types of exercises, intensity, duration
`and caloric burn. A caregiver wants to know that her father is
`living an active, healthy lifestyle and taking his daily walks.
`The police would like to know instantly when someone has
`been involved in a car collision, and whether the victims are
`moving or not.
`[0005] Existing products for the detection of animate and
`inanimate motions are simplistic in nature, and incapable of
`interpreting anything more than simple atomic movements,
`such as jolts, changes in orientation and the like. It is not
`possible to draw reliable conclusions about human behavior
`from these simplistic assessments.
`[0006]
`It is desirable to have an apparatus and method that
`can accurately monitor motion of either animate of inanimate
`objects.
`
`SUMMARY
`
`[0007] An embodiment includes a method of identifying a
`type ofmotion of an animate or inanimate object. The method
`includes generating an acceleration signature based on the
`sensed acceleration of the object. The acceleration signature
`is matched with at least one of a plurality of stored accelera-
`tion signatures, wherein each stored acceleration signatures
`corresponds with type of motion. The type of motion of the
`object is identified based on the statistical matching or exact
`matching of the acceleration signature.
`[0008] Another embodiment includes a method of identi-
`fying a type of motion of a person. The method includes
`generating an acceleration signature based on the sensed
`acceleration of an object attached to the person. The accel-
`eration signature is matched with at least one of a plurality of
`stored acceleration signatures, wherein each stored accelera-
`tion signatures corresponds with type of motion. The type of
`
`motion of the object is identified based on the statistical
`matching or exact matching of the acceleration signature.
`[0009] Other aspects and advantages of the described
`embodiments will become apparent from the following
`detailed description, taken in conjunction with the accompa-
`nying drawings, illustrating by way of example the principles
`of the described embodiments.
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`FIG. 1 shows examples of different types ofmotions
`[0010]
`of a human being that an object attached to the human being
`can be used to detect or sense.
`
`FIGS. 2A, 2B, 2C shows examples of time-lines of
`[0011]
`several different acceleration curves (signatures), wherein
`each signature is associated with a different type of sensed or
`detected motion.
`
`FIG. 3 is an example ofa block diagram ofa motion
`[0012]
`detection device.
`
`FIG. 4 is a flowchart that includes the steps of an
`[0013]
`example of a method for detecting various motions of daily
`living activities and emergency situations, such as, a fall.
`[0014]
`FIG. 5 is a flowchart that includes the steps of a
`method for detection of a fall.
`
`FIG. 6 is a flow chart that includes the steps of one
`[0015]
`example of a method of identifying a type of motion of an
`animate or inanimate object.
`[0016]
`FIG. 7 is a flow chart that includes steps of one
`example of a method of a motion detection device checking
`network availability for improvements in speed and/or pro-
`cessing power of acceleration signature matching.
`[0017]
`FIG. 8 shows a motion detection device that can be
`connected to one of multiple networks.
`
`DETAILED DESCRIPTION
`
`[0018] The monitoring of human activities generally falls
`into three categories: safety, daily lifestyle, and fitness. By
`carefully interpreting human movements it is possible to draw
`accurate and reasonably complete inferences about the state
`of well being of individuals. A high degree of sophistication
`is required in these interpretations. Simplistic assessments of
`human activity lead to inaccurate determinations, and ulti-
`mately are of questionable value. By contrast, a comprehen-
`sive assessment leads to an accurate interpretation and can
`prove to be indispensable in tracking the well being and safety
`of the individual.
`
`To draw accurate inferences about the behavior of
`[0019]
`humans, it turns out that the atomic movements become sim-
`ply alphabets that include elemental motions. Furthermore,
`specific sequences of elemental motions become the vocabu-
`lary that comprises human behavior. As an example, take the
`case of a person who leaves the home and drives to the
`shopping center. In such a scenario, the behavioral pattern of
`the person is walking to the door or the house, opening and
`closing the door, walking further to the car, settling down in
`the car, starting the engine, accelerating the car, going
`through a series of stops, starts and turns, parking the car,
`getting out and closing the car door, and finally walking to the
`shopping center. This sequence of human behavior is com-
`prised of individual motions such as standing, walking, sit-
`ting, accelerating (in the car), decelerating, and turning left or
`right. Each individual motion, for example walking, is com-
`prised of multiple atomic movements such as acceleration in
`an upward direction, acceleration in a downward direction, a
`
`11
`
`
`
`US 2010/0217533 A1
`
`Aug. 26, 2010
`
`modest forward acceleration with each step, a modest decel-
`eration with each step, and so on.
`[0020] With written prose, letters by themselves convey
`almo st no meaning at all. Words taken independently convey
`individual meaning, but do not provide the context to com-
`prehend the situation. It takes a complete sentence to obtain
`that context. Along the same line of reasoning, it requires a
`comprehension of a complete sequence of movements to be
`able to interpret human behavior.
`[0021] Although there is an undeniable use for products
`that are able to detect complex human movements accurately,
`the key to the success of such technologies lies in whether
`users adopt them or not. The technology needs to capture a
`wide range of human activities. The range of movements
`should ideally extend to all types of daily living activities that
`a human being expects to encounterisleeping, standing,
`walking, running, aerobics, fitness workouts, climbing stairs,
`vehicular movements, falling, jumping and colliding, to name
`some of the more common ones.
`
`It is important to detect human activities with a great
`[0022]
`deal of precision. In particular, activities that relate to safety,
`fitness, vehicular movements, and day to day lifestyle pat-
`terns such as walking, sleeping, climbing stairs, are important
`to identify precisely. For example, it is not enough to know
`that a person is walking. One needs to know the pace and
`duration of the walk, and additional knowledge of gait,
`unsteadiness, limping, cadence and the like are important.
`[0023]
`It is critical that false positives as well as false nega-
`tives be eliminated. This is especially important for cases of
`safety, such as falls, collisions, and the like. Human beings
`come in all typesishort, tall, skinny, obese, male, female,
`athletic, couch potato, people walking with stick/rolator,
`people with disabilities, old and young. The product needs to
`be able to adapt to their individuality and lifestyle.
`[0024] The embodiments described provide identification
`of types of motion of an animate or inanimate object. Motion
`is identified by generating acceleration signatures based on
`the sensed acceleration of the object. The acceleration signa-
`tures are compared with a library of motion signature, allow-
`ing the motion of the object
`to be identified. Further,
`sequences of the motions can be determined, allowing iden-
`tification of activities of, for example, a person the object is
`attached to.
`
`Just as the handwritten signatures of a given human
`[0025]
`being are substantively similar from one signature instance to
`the next, yet have minor deviations with each new instance, so
`too will the motion signatures of a given human be substan-
`tively similar from one motion instance to the next, yet have
`minor deviations.
`
`[0026] Algorithms used for pattern recognition (signature
`matching) should have the sophistication to accurately handle
`a wide range of motions. Such algorithms should have the
`ability to recognize the identical characteristics ofa particular
`motion by a given human being, yet allow for minor varia-
`tions arising from human randomness. Additionally,
`the
`devices used to monitor peoples’ movement need to be min-
`iature and easy to wear. These two objectives are fundamen-
`tally opposed. However, the described embodiments provide
`a single cohesive system that is both sophisticated enough to
`detect a wide range of motions.
`[0027]
`FIG. 1 shows examples of different types ofmotions
`of a human being that an object attached to the human being
`can be used to detect or sense. The human motions can
`
`include, for example, standing, sleeping, walking, and run-
`
`ning. A first motion 110 can include walking. A second
`motion 120 can include falling. A third motion 130 can
`include running. Each of the motions generates a unique
`motion signature. As will be described, the signatures can be
`universal to, for example, many individuals. Additionally, the
`signatures can have additional characteristics that are unique
`to an individual.
`
`FIGS. 2A, 2B, 2C shows examples ofdifferent types
`[0028]
`of acceleration and orientation signatures for various sample
`motions by human beings. It should be noted that these sig-
`natures are expected to have certain components that are
`common from one human being to the next, but also have
`certain components that vary from one human to the next.
`[0029] The signatures of FIGS. 2A, 2B, 2C are depicted in
`only one orientation. That is, three accelerometers can be
`used to generate acceleration signatures in the X, Y and Z
`(three) orientations. The signatures of FIGS. 2A, 2B, 2C only
`show the signature of one of the three orientations. It is to be
`understood that matching can use the other orientations as
`well.
`
`FIG. 2A shows an example of an acceleration sig-
`[0030]
`nature of a person doing a slow fall and lying down summer-
`sault. FIG. 2B shows an example of an acceleration signature
`of a person slipping and falling back on a bouncy surface (for
`example, an air mattress). FIG. 2C shows an acceleration
`signature of a person fall on their face with their knees flexed.
`By matching an acceleration signature that has been gener-
`ated by sensing the motion of a person with one of many
`stored signatures, the motion ofthe person can be determined.
`[0031]
`FIG. 3 is an example ofa block diagram ofa motion
`detection device. The motion detection device can be attached
`
`to an object, and therefore, detect motion ofthe object that can
`be identified. Based on the identified motion, estimates ofthe
`behavior and conditions of the object can be determined.
`[0032] The motion detection device includes sensors (such
`as, accelerometers) that detect motion of the object. One
`embodiment ofthe sensors includes accelerometers 312, 314,
`316 that can sense, for example, acceleration of the object in
`X, Y and Z directional orientations. It is to be understood that
`other types of motion detection sensors can alternatively be
`used.
`
`[0033] An analog to digital converter (ADC) digitizes ana-
`log accelerometer signals. The digitized signals are received
`by compare processing circuitry 330 that compares the digi-
`tized accelerometer signals with signatures that have been
`stored within a library of signatures 340. Each signature
`corresponds with a type of motion. Therefore, when a match
`between the digitized accelerometer signals and a signature
`stored in the library 340, the type of motion experienced by
`the motion detection device can determined.
`
`[0034] An embodiment includes filtering the accelerometer
`signals before attempting to match the signatures. Addition-
`ally, the matching process can be made simpler by reducing
`the possible signature matches.
`[0035] An embodiment
`includes identifying a previous
`human activity context. That is, for example, by knowing that
`the previous human activity was walking, certain signatures
`can intelligently be eliminated from the possible matches of
`the present activity that occurs subsequent to the previous
`human activity (walking).
`[0036] An embodiment includes additionally reducing the
`number of possible signature matches by performing a time-
`domain analysis on the accelerometer signal. The time-do-
`main analysis can be used to identify a transient or steady-
`
`12
`
`12
`
`
`
`US 2010/0217533 A1
`
`Aug. 26, 2010
`
`is, for
`state signature of the accelerometer signal. That
`example, a walk may have a prominent steady-state signature,
`whereas a fall may have a prominent transient signature.
`Identification of the transient or steady-state signature of the
`accelerometer signal can further reduce or eliminate the num-
`ber of possible signature matches, and therefore, make the
`task ofmatching the accelerometer signature with a signature
`within the library of signature simpler, and easier to accom-
`plish. More specifically, the required signal processing is
`simpler, easier, and requires less computing power.
`[0037] Upon detection of certain types of motion, an audio
`device 360 and/or a global positioning system (GPS) 370 can
`engaged to provide additional information that can be used to
`determine the situation of, for example, a human being the
`motion detection device is attached to.
`
`[0038] The condition, or information relating to the motion
`detection device can be communicated through a wired or
`wireless connection. A receiver of the information can pro-
`cess it, and make a determination regarding the status of the
`human being the motion detection device is attached to.
`[0039]
`FIG. 4 is a flowchart that includes the steps of an
`example of a method for detecting various motions of daily
`living activities and emergency situations, such as, a fall. A
`first step 410 includes monitoring an activity of a person the
`motion detection device is attached. Raw signal data is col-
`lected from, for example, an accelerometer sensor. A second
`step 420 includes performing instantaneous computations
`over raw signals to compute atomic motions along with grav-
`ity vector and tilt vector. A step third step 430 includes apply-
`ing series of digital filters to remove noise in the atomic
`motions data. A fourth step 440 includes performing state
`analysis on series of atomic data samples and forming con—
`text. Depending on the state analysis, the series of atomic data
`is passed through either a step 445 periodic or steady state
`data analysis or a step 450 transient state data analysis.A sixth
`step 460 includes formation of macro motion signatures. The
`macro motion signatures are built from an output of state
`analysis vectors using known wavelet transformation tech-
`niques (for example, a Haar Transform). The transform per-
`forms pattern matching on current motion pattern with exist-
`ing motion pattern library using, for example, DWT (Discreet
`Wavelet Transform) techniques. Complex motion wavelets
`are later matched using statistical pattern matching tech-
`niques, such as, HHMM (Hidden Heuristic Markov Model).
`The statistical pattern matching includes detecting and clas-
`sifying events of interest. The events of interest are built by
`observing various motions and orientation states data of an
`animate or inanimate object. This data is used to train the
`statistical model which performs the motion/activity detec-
`tion. Each activity will have its own model trained based on
`the observed data. A seventh step 470 includes a learning
`system providing the right model for the user from a set of
`model. It also aids in building newer (personal) patterns
`which are not in the library for the person who is wearing the
`motion detection device. An eighth step 480 includes pre-
`building a motion database of motion libraries against which
`motion signatures are compared. The database adds new
`motion/states signature dynamically as they are identified.
`[0040]
`FIG. 5 is a flowchart that includes the steps of an
`example of a method for detecting a fall. A first step 510
`includes monitoring an activity of, for example, a person the
`motion detection device is attached to. A step 515 includes
`recording and reporting in deviations in normal motion pat-
`terns of the person. A step 520 includes detecting the accel-
`
`eration magnitude deviation exceeding a threshold. The
`acceleration magnitude deviation exceeding the threshold
`can be sensed as a probable fall, and audio recording is
`initiated. Upon detection ofthis condition, sound recording of
`the person the motion detection device is connected to can be
`activated. The activation of sound can provide additional
`information that can be useful in assessing the situation ofthe
`person. A step 530 includes monitoring the person after the
`probable fan. A step 525 includes detection of another accel-
`eration having magnitude lesser than the threshold, and c011-
`tinuing monitoring of audio. A step 535 includes detecting a
`short period of inactivity. A step 540 includes monitoring the
`person after determining a fall probably occurred. A step 545
`includes subsequently detecting normal types of motion and
`turning offthe audio because the person seems to be perform-
`ing normal activity. A step 550 includes monitoring a period
`of inactivity. A step 555 includes additional analysis of
`detected information and signals. A step 560 includes further
`analysis including motion data, orientation detection all indi-
`cating the person is functioning normally. A step 560 includes
`determining that a fall has occurred based on the analysis of
`the motion data, and analysis of a concluded end position and
`orientation of the person. The sound recording can be de-
`activated. A step 565 includes concluding that a fall has
`occurred. A step 570 includes sending an alert and reporting
`sound recordings. A step 575 includes the fall having been
`reported. A step 580 includes an acknowledgement ofthe fall.
`[0041]
`FIG. 6 is a flow chart that includes the steps of one
`example of a method of identifying a type of motion of an
`animate or inanimate object. A first step 610 includes gener-
`ating an acceleration signature (for example, a tri -axial) based
`on the sensed acceleration of the object. A second step 620
`includes matching the acceleration signature with at least one
`of a plurality of stored acceleration signatures, wherein each
`stored acceleration signatures corresponds with type of
`motion. A third step 630 includes identifying the type of
`motion of the object based on the statistical (pattern) match-
`ing or exact matching ofthe acceleration signature. As will be
`described, the acceleration signal can be created using a
`wavelet transformation.
`
`For embodiments, the type of motion includes at
`[0042]
`least one of atomic motion, elemental motion and macro-
`motion.
`
`[0043] Though embodiments of generating matching
`acceleration signatures are described, it is to be understood
`that additional or alternate embodiments can include gener-
`ating and matching of orientation and/or audio signatures.
`Correspondingly, the first step 610 can include generating an
`acceleration signature, (and/or) orientation and audio signa-
`ture based on the sensed acceleration, orientation ofthe object
`and audio generated by the object, for example, a thud of a
`fall, or a cry for help.
`[0044] Atomic motion includes but is not limited to a sharp
`jolt, a gentle acceleration, complete stillness, a light accelera-
`tion that becomes stronger, a strong acceleration that fades, a
`sinusoidal or quasi-sinusoidal acceleration pattern, vehicular
`acceleration, vehicular deceleration, vehicular left and right
`turns, and more.
`[0045] Elemental motion includes but is not limited to
`motion patterns for walking, running, fitness motions (e.g.
`elliptical machine exercises, rowing, stair climbing, aerobics,
`skipping rope, bicycling .
`.
`. ), vehicular traversal, sleeping,
`sitting, crawling, turning over in bed, getting out of bed,
`getting up from chair, and more.
`
`13
`
`13
`
`
`
`US 2010/0217533 A1
`
`Aug. 26, 2010
`
`[0046] Macro-motion includes but is not limited to going
`for a walk in the park, leaving home and driving to the shop-
`ping center, getting out of bed and visiting the bathroom,
`performing household chores, playing a game of tennis, and
`more.
`
`[0047] Each of the plurality of stored acceleration signa-
`tures corresponds with a particular type ofmotion. By match-
`ing the detected acceleration signature of the object with at
`least one of a plurality of stored acceleration signatures, an
`estimate or educated guess can be made about the detected
`acceleration signature.
`[0048] An embodiment includes a common library and a
`specific library, and matching the acceleration signature
`includes matching the acceleration signature with stored
`acceleration signatures of the common library, and then
`matching the acceleration signature with stored acceleration
`signatures of the specific library. For a particular embodi-
`ment, the general library includes universal acceleration sig-
`natures, and the specific library includes personal accelera-
`tion signatures. That is, for example, the stored acceleration
`signatures of the common library are useable for matching
`acceleration signatures of motions of multiple humans, and
`the stored acceleration signatures of the specific library are
`useable for matching acceleration signatures of motions of a
`particular human. Additionally, each library can be further
`categorized to reduce the number of possible matches. For
`example, at an initialization, a user may enter physical char-
`acteristics of the user, such as, age, sex, and/or physical char-
`acteristics (such as, the user has a limp). Thereby, the possible
`signature matches within the general library can be reduced.
`The signature entries within the specific library can be
`leamed (built) over time as the human wearing the motion
`detection device goes through normal activities ofthe specific
`human. The specific library can be added to, and improved
`over time.
`
`[0049] An embodiment includes filtering the acceleration
`signals. Additional embodiment include reducing the number
`of stored acceleration signature matches by identifying a
`previous activity of the object, and performing a time domain
`analysis on the filtered acceleration signal to identify tran-
`sient signatures or steady-state signatures of the filtered
`acceleration signal. That is, by identifying a previous activity
`(for example, a human walking of sleeping) the possible
`number ofpresent activities can be reduced, and therefore, the
`number of possible stored acceleration signature matches
`reduced. Additionally, the transient and/or steady-state sig-
`natures can be used to reduce the number of possible stored
`acceleration signature matches, which can improve the pro-
`cessing speed.
`includes activating audio
`[0050] Another embodiment
`sensing of the object if matches are made with at least por-
`tions of particular
`stored acceleration signatures. For
`example, if the acceleration signature exceeds a threshold
`value, then audio sensing of the object is activated. This is
`useful because the audio information can provide additional
`clues as to what, for example, the condition of a person. That
`is, a fall may be detected, and audio information can be used
`to confirm that a fall has in fact occurred.
`
`tion to determine, for example, if the user is in pain, uncon-
`scious or in a dangerous situation (for example, in a shower or
`in a fire).
`[0052] An embodiment includes the object being associ-
`ated a person, and the stored acceleration signatures corre-
`sponding with different types of motion related to the person.
`A particular embodiment includes identifying an activity of
`the person based on a sequence of identified motions of the
`person. The activity of the person can include, for example,
`falling (the most important in some applications), walking,
`running, driving and more. Furthermore, the activities can be
`classified as daily living activities such as walking, running,
`sitting, sleeping, driving, climbing stairs, and more, or spo-
`radic activities, such as falling, having a car collision, having
`a seizure and so on.
`
`[0053] An embodiment includes transmitting information
`related to the identified type of motion if matches are made
`with particular stored acceleration signatures. The informa-
`tion related to the identified type ofmotion can include at least
`one of motions associated with a person the object is associ-
`ated with. The motions can include, for example, a heartbeat
`of the person, muscular spasms, facial twitches, involuntary
`reflex movements which can be sensed by, for example, an
`accelerometer. Additionally, the information related to the
`identified type of motion can include at least one of location
`of the object, audio sensed by the object, temperature of the
`object.
`[0054] Another embodiment includes storing at least one of
`the plurality of stored acceleration signatures during an ini-
`tialization cycle. The initializing cycle can be influenced
`based on what the object is attached to. That is, initializing the
`stored acceleration signatures (motion patterns) can be based
`on what the object is attached to, which can both reduce the
`number of signature required to be store within, for example,
`the general library, and therefore, reduce the number of pos-
`sible matches and reduce the processing required to identify a
`match. Alternatively or additionally, initializing the stored
`acceleration signatures can be based on who the object is
`attached to, which can influence the specific library. The
`initialization can be used to determine motions unique, for
`example, to an individual. For example, a unique motion can
`be identified for a person who walks with a limp, and the
`device can be initialized with motion patterns of the person
`walking with a limp.
`[0055] An embodiment includes initiating a low-power
`sleep mode of the object if sensed acceleration is below a
`threshold for a predetermined amount of time. That is, if, for
`example, a person is sensed