a2) United States Patent
`US 7,463,997 B2
`(10) Patent No.:
`Pasoliniet al.
`(45) Date of Patent:
`Dec. 9, 2008
`
`
`US007463997B2
`
`(54) PEDOMETER DEVICE AND STEP
`DETECTION METHOD USING AN
`ALGORITHM FOR SELF-ADAPTIVE
`COMPUTATION OF ACCELERATION
`THRESHOLDS
`.
`.
`.
`..
`,
`Inventors: Fabio Pasolini, S. Martino Siccomario
`(IT); Ivo Binda, Voghera(IT)
`
`75
`(75)
`
`(73) Assignee: STMicroelectronics S.r.]., Agrate
`Brianza (IT)
`
`(*) Notice:
`
`Subject to any disclaimer, the term ofthis
`patent is extended or adjusted under 35
`U.S.C. 154(b) by 3 days.
`
`(21) Appl. No.: 11/537,933
`(22)
`Filed:
`Oct. 2, 2006
`oo.
`,
`Prior Publication Data
`US 2007/0143068 Al
`Jun. 21, 2007
`
`(65)
`
`Foreign Application Priority Data
`(EP)
`eee eeeeeeeseeseeseeeeeeeeee 05425683
`
`(30)
`Oct. 3, 2005
`(51)
`Int.Cl
`nt.
`.
`(2006.01)
`GOIC 22/00
`
`702/160
`(52) US. Ch.
`.....-.porereteerenessrenrecencces
`(58) Field of Classification Search ................. 702/141,
`702/150-154, 158, 160; 600/595; 73/490,
`—
`;
`73/510
`See applicationfile for complete search history.
`References Cited
`U.S. PATENT DOCUMENTS
`
`(56)
`
`4/2000 Gaudet et al... 702/160
`6,052,654 A
`10/2000 Richardsonet al.
`......... 600/300
`6,135,951 A
`6,826,477 B2* 11/2004 Ladetto etal... TOL/217
`6,898,550 Bl
`5/2005 Blackadar et al.
`........... 702/182
`2006/0020177 A1*
`1/2006 Seoet al.
`seeessseeeeeeeee 600/300
`2007/0073514 Al*
`3/2007 Nogimori et al.
`........... 702/160
`2007/0143069 AL*
`6/2007 Pasolini et al. «2... 702/160
`2007/0198187 AL*
`8/2007 Pasolini et al.
`......0....6 701/220
`
`FOREIGN PATENT DOCUMENTS
`
`GB
`
`2 359 890
`
`9/2001
`
`* cited by examiner
`Primary Examiner—Michael P Nghiem
`(74) Attorney, Agent, or Firm—Lisa K.Jorgenson; Dennis M.
`de Guzman; Seed IP Law Group PLLC
`
`(57)
`
`ABSTRACT
`
`In a pedometer device for detecting and counting steps of a
`user on foot, an accelerometer sensor detects a vertical accel-
`:
`:
`:
`:
`eration generated during the step. A processing unit, con-
`nected to the accelerometer sensor, processes an acceleration
`signal relating to the acceleration in orderto detect the occur-
`rence of a step, and in particular compares the acceleration
`signal with a first reference threshold. The processing unit
`automatically adapts thefirst reference threshold as a func-
`tion of the acceleration signal. In particular, the processing
`unit modifies the first reference threshold as a function of an
`envelope of the amplitude of the acceleration signal.
`
`5,583,776 A
`
`12/1996 Levietal. we. 364/450
`
`30 Claims, 5 Drawing Sheets
`
`yo!
`
`PROCESSING UNIT
`
`SETTING
`
`FIRST COMPARATOR
`
`DISTANCE-CALCULATION
`
`
`
`
`
`oor]
`
`
`THRESHOLD-ADAPTATION
`SECOND COMPARATOR
`
`
`
`ENVELOPE CALCULATION MEAN-VALUE CALCULATION
`
`[
`
`_LENGTH-ESTIMATION
`
`AXIS-DETERMINATION
`
`HTCv. Uniloc Luxembourg
`
`Page | of 12
`
`HTC Ex. 1005
`
`HTC v. Uniloc Luxembourg
`
`Page 1 of 12
`
`HTC Ex. 1005
`
`

`

`U.S. Patent
`
`Dec. 9, 2008
`
`Sheet 1 of 5
`
`US 7,463,997 B2
`
`i!
`
`PROCESSING UNIT
`
`SETTING
`
`FIRST COMPARATOR
`
`DISTANCE-CALCULATION
`
`THRESHOLD-ADAPTATION
`
`SECOND COMPARATOR
`
`
`
`ENVELOPE CALCULATION
`
`MEAN-VALUE CALCULATION
`
`LENGTH—-ESTIMATION
`
`AXIS—DETERMINATION
`
`DISPLAY
`
`INTERFACE
`
`FIG. 1
`
`HTCv. Uniloc Luxembourg
`
`Page 2 of 12
`
`HTC Ex. 1005
`
`HTC v. Uniloc Luxembourg
`
`Page 2 of 12
`
`HTC Ex. 1005
`
`

`

`U.S. Patent
`
`Dec. 9, 2008
`
`Sheet 2 of 5
`
`US 7,463,997 B2
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`HTCv. Uniloc Luxembourg
`
`Page 3 of 12
`
`HTC Ex. 1005
`
`HTC v. Uniloc Luxembourg
`
`Page 3 of 12
`
`HTC Ex. 1005
`
`

`

`U.S. Patent
`
`Dec. 9, 2008
`
`Sheet 3 of 5
`
`US 7,463,997 B2
`
`
`
`
`
`
`
`DETECTION
`CalAcc > St
`
`
`
`
`
`
`
`
`
`NEGATIVE PHASE BEEN
`DETECTED?
`
`STEP
`INCREMENT
`
`STEP LENGTH
`
`INCREMENT OF CALORIE
`SPEED COMPUTATION
`
`HTCv. Uniloc Luxembourg
`
`Page 4 of 12
`
`HTC Ex. 1005
`
`HTC v. Uniloc Luxembourg
`
`Page 4 of 12
`
`HTC Ex. 1005
`
`

`

`U.S. Patent
`
`Dec. 9, 2008
`
`Sheet 4 of 5
`
`US 7,463,997 B2
`
`DETERMINATION OF ACCELERATION DATUM
`CalAcc AND THRESHOLD ADAPTATION
`
`
`11.18
`’
`
`ACQUISITION OF
`ACCELERATION SAMPLEAcc
`
`30
`
`
`ELIMINATION OF D.C. COMPONENT |
`34
`AND DETERMINATION OF CalAcc
`
`34
`
`32
`
`
`Env* = o,"Env*
`(a, < 1)
`
`he
`
`Path
`;
`
`35
`
`NO<<YES
`
`[eres
`
`Env = o,"Env-
`(a< 1)
`
`37
`
`St = B* Env?
`(<1).
`38
`
`
`
`S°= B* Env
`(B < 1)
`
`
`
`
`
`HTCv. Uniloc Luxembourg
`
`Page 5 of 12
`
`HTC Ex. 1005
`
`HTC v. Uniloc Luxembourg
`
`Page 5 of 12
`
`HTC Ex. 1005
`
`

`

`U.S. Patent
`
`Dec. 9, 2008
`
`Sheet 5 of 5
`
`US 7,463,997 B2
`
`CalAcc
`
`EET
`MAHA EL
`EH WA 5
`| NGnyAMAalu —s
`oohATTATTACTMTTINIRTUa
`s, TRATES '
`
`i ri Hy i Wall , HT fe : Vy eB }
`AA TAT TUTE8
`
`PARAMETER
`INITIALIZATION
`
`DETERMINATION OF ACCELERATION DATUM
`CalAcc AND THRESHOLD ADAPTATION
`
`POSITIVE PHASE DETECTION
`CalAcc > S*
`
`
`HAS A
`POSITIVE PHASE BEEN
`DETECTED?
`
`STEP INCREMENT
`
`STEP LENGTH ADAPTATION
`
`DISTANCE INCREMENT
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`INCREMENT OF
`
`SPEED COMPUTATION 8]
`
`CALOR
`
`HTCv. Uniloc Luxembourg
`
`Page 6 of 12
`
`HTC Ex. 1005
`
`HTC v. Uniloc Luxembourg
`
`Page 6 of 12
`
`HTC Ex. 1005
`
`

`

`US 7,463,997 B2
`
`1
`PEDOMETER DEVICE AND STEP
`DETECTION METHOD USING AN
`ALGORITHM FOR SELF-ADAPTIVE
`COMPUTATION OF ACCELERATION
`THRESHOLDS
`
`BACKGROUND OF THE INVENTION
`
`1. Field of the Invention
`
`2
`counted as steps; on the other hand, if the threshold is too
`high, some steps may notbe detected.
`
`BRIEF SUMMARY OF THE INVENTION
`
`invention provides a
`One embodiment of the present
`pedometer device and a method for detecting and counting
`steps which will enable the aforesaid disadvantages and prob-
`lems to be overcome.
`
`BRIEF DESCRIPTION OF THE SEVERAL
`VIEWS OF THE DRAWINGS
`
`20
`
`40
`
`45
`
`60
`
`Thepresent invention relates to a pedometer device and to
`One embodimentofthe invention is a pedometer device for
`a step detection method using an algorithm for self-adaptive
`detecting and counting the steps ofa user. The device includes
`computation of acceleration thresholds.
`an accelerometer sensor configured to detect an acceleration
`2. Description of the Related Art
`generated during a step; and a processing unit connected to
`the accelerometer sensor, and configured to process an accel-
`Step-counting devices (referred to in general as pedom-
`eration signal relating to the acceleration to detect the occur-
`eters) are known, which, being carried by a user, enable
`rence of a step. The processing unit includesafirst compara-
`measurementofthe numberof steps made, and calculation of
`tor configured to compare the acceleration signal with a first
`the distancetraveled, as well as supplying of additionalinfor-
`reference threshold, and a threshold-adaptation circuit con-
`mation, such as, for example, the average speed, or the con-
`figured to modify thefirst reference threshold as a function of
`sumption of calories.
`the acceleration signal.
`Pedometers are advantageously used in inertial navigation
`systems (the so-called dead-reckoning systems) applied to
`One embodiment of the invention is a step detection
`human beings. Such systemstrace the movements ofa user,
`method for detecting steps in the gait of a user. The method
`by identifying and measuring his/her displacements starting
`includes producing an acceleration signal relating to an accel-
`from a knownstarting point, without resorting to the use of a
`eration generated during a step; and processing the accelera-
`Global Positioning System (GPS), or by acting as aid to a
`tion signal to detect the occurrenceofthe step. The processing
`
`GPS. In said systems, a compass supplies the information step includes comparing the acceleration signal withafirst
`linked to the direction of displacement, and the pedometer
`reference threshold, and modifyingthe first reference thresh-
`supplies the information linked to the amount of said dis-
`old as a function of the acceleration signal.
`placement. Pedometersare also used in a wide range of appli-
`cations in the clinical sector (for example, in rehabilitation),
`and in general in the field of fitness (for example, as instru-
`ments for monitoring a physical activity).
`In particular, pedometers are known that use integrated
`accelerometers of a MEMS (micro-electromechanical sys-
`tem) type for step detection. In particular, such pedometers
`have particularly compact dimensions, and can be advanta-
`geously integrated within portable devices, such as mobile
`phones, Mp3 readers, camcorders, etc.
`The aforesaid pedometers implement a step detection
`method based upon the analysis of the pattern of a vertical
`acceleration, which is generated during the various phases of
`the step by the contact of the foot to the ground, and which is
`detected by an accelerometerfixed to the body of the user. In
`this connection, it is emphasized that “vertical acceleration”
`meansherein an acceleration directed along the vertical ofthe
`user’s body. In particular, the occurrence of a step is deter-
`mined by identifying acceleration peaks that appear in the
`acceleration signal, and said peaksare detected by comparing
`the acceleration signal with a given reference threshold, hav-
`ing a pre-set value.
`However, even thoughthe acceleration signal hasa profile
`that is repeatableat eachstep, its pattern (and, in particular,its
`amplitude and temporal extension) has a wide variability
`according to a numberoffactors that affect the gait, such as
`the resting surface, the type of shoe worn (rigid sole or flex-
`ible sole, etc.), and the speed of the gait (slow walking, fast
`walking, running,etc.). Furthermore, each individualuser has
`given characteristics and peculiarities that affect the gait,
`differentiating it from that of other users.
`It follows that a step detection based upon the comparison
`ofthe value ofthe acceleration signal with a reference thresh-
`old having a pre-set value for the detection of acceleration
`peaks, involves the occurrence of errors that may even be
`considerable in counting ofthe steps, and in the measurement
`of the distance traveled. In particular, if the threshold is too
`low, spurious signals, rebounds, or noise in general, may be
`
`For a better understanding of the present invention, pre-
`ferred embodiments thereof are now described, purely by
`way of non-limiting example and with reference to the
`attached drawings, wherein:
`FIG. 1 showsa block diagram of a pedometer device;
`FIG. 2 shows a graph corresponding to the pattern of an
`acceleration signal during a step;
`FIG. 3 shows a flowchart corresponding to operations of
`detection and counting of steps, executed by a processing unit
`of the pedometer device of FIG. 1;
`FIG. 4 shows a flowchart corresponding to operations of
`self-adaptive modification of acceleration thresholds,
`executed by the processing unit of the pedometer device of
`FIG.1;
`FIGS. 5-6 are graphs corresponding to the pattern of an
`acceleration signal during a step and of reference thresholds
`associated to the algorithm of FIG.3;
`FIG. 7 showsa possible variant of the flowchart of FIG.3;
`and
`
`FIG. 8 is a partially exploded schematic view ofa portable
`device,
`in particular a mobile phone,
`incorporating the
`pedometer device of FIG. 1.
`
`DETAILED DESCRIPTION OF THE INVENTION
`
`FIG. 1 is a schematic illustration of a pedometer device 1,
`comprising an accelerometer2, of a linear type and having a
`vertical detection axis z, and a processing unit 3, connected to
`the accelerometer 2. Advantageously, the accelerometer 2
`and the processing unit 3 are mounted on the sameprinted
`circuit, housed inside a casing of the pedometer device 1 (not
`illustrated). The pedometer device 1 is carried by a user, for
`example on his belt or on his shoulder, so asto be fixed to the
`
`HTC v. Uniloc Luxembourg
`
`Page 7 of 12
`
`HTC Ex. 1005
`
`HTC v. Uniloc Luxembourg
`
`Page 7 of 12
`
`HTC Ex. 1005
`
`

`

`US 7,463,997 B2
`
`3
`4
`rence of a step), counting ofthe steps and measurementofthe
`body of the user and be able to sense vertical accelerations
`total distance traveled are updated; otherwise, the algorithm
`that occur during the step, caused by the impactof the feet on
`returns to the initial condition of looking for a new positive
`the ground.
`phaseofthe step. In particular, the positive acceleration peaks
`The pedometer device 1 further comprises a display screen
`that occur within the pre-set time interval are ignored by the
`4, connected at an output of the processing unit 3, and an
`algorithm (in so far as they can be ascribed to phenomena of
`interface 5, connected at an input ofthe processing unit 3. The
`noise, such as impact, anomalous rebounds, etc.).
`display screen 4 displays information at output from the
`In detail, the algorithm starts with initialization, block 10,
`pedometer device 1, such as the numberofsteps, the distance
`ofthe values ofthe positive and negative reference thresholds
`traveled, etc. The interface 5, for example, including push-
`Stand S", respectively, at a positive minimum value S, and at
`buttons, an alphanumeric keypad, communicationports, etc.,
`a negative minimum value S,, the latter being smaller, in
`allows the user to communicate with the processing unit 3 (for
`absolute value, than the positive minimum value S,. As will
`example, by entering data).
`be clarified, said minimum values represent limit values
`The accelerometer 2 is advantageously an integrated sen-
`below which the reference thresholds are not allowedto drop.
`sor of semiconductor material, made using the MEMStech-
`In addition, the values of a positive envelope Env* and of a
`nology, of a known type and thus not described in detail
`negative envelope Env” of the acceleration signal A (which
`herein. In use, the accelerometer 2 detects the component
`will subsequently be used for modification of the reference
`along the detection axis z of the vertical acceleration gener-
`thresholds)are initialized, respectively, at the positive mini-
`ated during the step, and produces a corresponding accelera-
`mum value S, andat the negative minimum value S,.
`tion signal A.
`20
`As shownin FIG.2,the pattern ofthe acceleration signal A Next, block 11, the processing unit 3 determinesafirst
`
`(with the d.c. componentfiltered out) in time t has a given
`acceleration datum CalAcc, and consequently modifies the
`acceleration profile which repeats at each step (indicated by
`values of the reference thresholds (as will be described in
`the dashed rectangle). In detail, the acceleration profile com-
`detail hereinafter with reference to FIGS. 4 and 5).
`prises in succession: a positive phase, in which a positive-
`The algorithm then proceeds, block 12, with the search for
`acceleration peak occurs(i.e., directed upwards), due to con-
`the positive phase of the step, by comparing the value of the
`tact and consequent impactof the foot with the ground; and a
`acceleration datum CalAcc with thepositive reference thresh-
`old S*, to detect a positive acceleration peak of the accelera-
`negative phase in which a negative-acceleration peak occurs
`(i.e., directed downwards) due to rebound, having an absolute
`tion signal A.
`value smaller than that of the positive-acceleration peak.
`Until a positive phase of the step is found, block 13, the
`The processing unit 3, comprising a microprocessorcircuit
`algorithm proceeds with acquisition of a new acceleration
`(for example, a microcontroller or DSP), acquires at pre-set
`datum CalAcc in block 11 (and corresponding modification
`intervals samples of the acceleration signal A generated by
`of the reference thresholds), and with the comparisonof said
`the accelerometer 2, and executes appropriate processing
`new acceleration datum with the positive reference threshold
`s*.
`operations for counting the numberof steps and measuring
`the distance traveled. As will be described in detail hereinaf-
`Thepositive phase is detected whenthe acceleration datum
`exceeds the positive reference threshold S* and then drops
`below the positive reference threshold, the instant of detec-
`tion of the positive phase corresponding to the instant in
`whichthe acceleration datum drops again below the positive
`reference threshold S*. At this instant, the processing unit 3
`stores the value assumedby the positive reference threshold
`S*, which is a maximum value S*,,,_...
`After the positive phase detection, the algorithm proceeds
`with the search for the negative phase of the step, block 14,
`i.e., of a negative acceleration peak, by comparing the value
`ofthe acceleration datum CalAcc with the negative reference
`threshold S~. In particular, the search for the negative phase of
`the step is executed within a certain time interval Mask, the
`value of which must be lower than a maximum interval Max-
`
`40
`
`45
`
`ter, the processing unit 3 compares the value ofthe accelera-
`tion signal A (with the d.c. componentfiltered out) with a
`positive reference threshold S* and with a negative reference
`threshold S-, for identifying, respectively, the positive phase
`(positive acceleration peak) and the negative phase (negative
`acceleration peak) of the step.
`According to one embodimentofthe present invention, the
`values ofthe positive and negative reference thresholds S*, S~
`are not fixed and equal to a given pre-set value, but are
`calculated in a self-adaptive way(i.e., in a way that adapts
`without any external intervention from a user) by the process-
`ing unit 3, based on the values assumed by the detected
`acceleration.In particular, as will be clarified hereinafter, the
`values ofthe positive and negative reference thresholds St, S~
`are modified at each acquisition of a new sample of the
`acceleration signal A, as a function of the value ofa positive
`and negative amplitude envelopeofthe acceleration signal, in
`such a mannerthat the reference thresholds vary with time
`approximately following said envelopes. The pedometer
`device 1 thus adapts to variations in the detection conditions
`(and, in particular, to different profiles of the acceleration
`signal, in terms of amplitude and duration), due, for example,
`to a different type ofterrain, or to an increase in the speed of
`the gait.
`The algorithm implemented by the processing unit 3 for
`performing, amongotherthings, the operations of step count-
`ing andoftraveled distance measurementis now described,
`with reference to FIG. 3. Said algorithm envisages the analy-
`sis of the acceleration signal A in orderto look for a positive
`phase ofthe step followed by anegative phase within a pre-set
`time interval from the occurrenceof the positive phase. In the
`case where said sequence occurs (which indicates the occur-
`
`_Maskfrom detection ofthe positive phase (corresponding to
`acertain numberof samples, the value ofwhich is determined
`also as a function of the sampling rate of the acceleration
`data).
`Until a negative acceleration peak is detected, block 15,
`and as long as the time interval Mask is shorter than the
`maximum interval Max_Mask, block 16, the algorithm pro-
`ceeds with the search for the negative phase ofthe step. In
`detail, the time interval Mask is incremented, block 17, a new
`acceleration datum CalAcc is acquired (and the values of the
`reference thresholds are modified accordingly), block 18
`(which is equivalent to block 11), and the algorithm returnsto
`block 14. Ifno negative phase of the step has been identified
`after expiry of the maximum interval Max_Mask,block 16,
`the algorithm returns to block 11 in order to look for a new
`potential positive phase ofthe step.
`On the contrary, if the negative phase is identified within
`the maximum interval Max_Mask (ie.,
`the acceleration
`
`HTC v. Uniloc Luxembourg
`
`Page 8 of 12
`
`HTC Ex. 1005
`
`HTC v. Uniloc Luxembourg
`
`Page 8 of 12
`
`HTC Ex. 1005
`
`

`

`US 7,463,997 B2
`
`5
`datum drops below the negative reference threshold S-), the
`processing unit 3 determines the occurrenceof a step, and,
`block 20, increments the count of the detected steps. Further-
`more, the estimate of the distance traveled is updated by
`adding to the previous value an estimate of the length of the
`current step LPS.
`In detail, according to one embodiment of the present
`invention, block 21, the processing unit 3 calculates the esti-
`mate of the length of the current step LPS as a function of the
`maximum value S*,,,,, reached by the positive reference
`threshold S* during the positive phase of the step, which is
`indicatory of the valueofthe positive acceleration peak. The
`actual length ofthe step varies with respect to a standard value
`determined on the basis of the physical characteristics of the
`user, according to the speedofthe step, or, equivalently, to the
`amplitude of the generated acceleration. Consequently, the
`estimate ofthe length ofthe current step LPSis calculated via
`the formula:
`
`EPS=LPf(S"pax)
`
`where LPis a standard length ofthe step, correspondingto 0.4
`to 0.5 timesthe heightof the user, and f(S*,,,,,) is a corrective
`function, for example a linear one, based upon the maximum
`value S*,,,... Lhe corrective function f(S",,,,,.) can be tabulated
`on the basis ofexperimentaltests, which enable association to
`a given maximum value S”*,,,,.. of an appropriate correction to
`be madeto the standard length ofthe step LP. In particular, the
`function f(S",,,,,.) is conveniently stored in the processing unit
`3.
`
`The algorithm then proceeds, block 22, by increasing the
`distance traveled on the basis of the estimate of the length of
`the current step LPS, previously calculated. Furthermore,
`block 23, further variables supplied at output from the
`pedometer device 1 can be updated, such as the number of
`calories (also in this case, the previous count is updated by
`adding an average consumption ofcalories per step), and the
`average and instantaneous speed of travel, which are calcu-
`lated in a known way notdescribed in detail herein.
`Next, the algorithm returns to block 11 in order to detect a
`new acceleration profile indicating occurrenceofa step.
`There will now be described in detail, with reference to
`FIG.4, the algorithm implemented by the processing unit 3
`for determination of a new acceleration datum CalAcc and
`consequent updating ofthe valuesofthe positive and negative
`reference thresholds St and S~, in such a manner that the
`aforesaid values will follow approximately the positive and
`negative envelope of the acceleration signal.
`In brief, said algorithm envisages calculation, for each new
`acceleration datum CalAcc, ofthe values ofthe positive enve-
`lope Env* and negative envelope Env’, and modification of
`the value of the positive and negative reference thresholds St
`and S~ as a function of the positive envelope Env* and nega-
`tive envelope Env’, respectively.
`In detail,
`in an initial block 30, the processing unit 3
`acquires from the accelerometer 2 a new acceleration sample
`Accofthe acceleration A. Then, block 31, the d.c. component
`of said acceleration value (due substantially to the accelera-
`tion of gravity) is eliminated so as to determinethe accelera-
`tion datum CalAcc, with zero mean value, which will be used
`subsequentlyin the algorithm.In detail, the mean value Accm
`of the acceleration sample Accis calculated with the expres-
`sion:
`
`Accm=y:Accm+(1-y):CalAce
`
`wherey is a constant between 0 and 1, for example equal to
`0.95, andAccm and CalAccarethe values, respectively, ofthe
`
`10
`
`15
`
`20
`
`30
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`6
`mean value andof the acceleration datum, which were cal-
`culated at the previous acquisition. The new acceleration
`datum CalAccis calculated by applyingtherelation:
`CalAcc=Acc-Accm
`
`Then,the algorithm proceeds with the determination ofthe
`new valuesofthe positive and negative envelopes Env*, Env”.
`In detail, block 32, if the value of the acceleration datum
`CalAccis higher than the value ofthe positive envelope Env*
`(as calculated at the previous acquisition), the new value of
`the positive envelope Env* is set equal to the value of the
`acceleration datum CalAcc, block 33. Otherwise, block 34,
`the valueofthe positive envelope Env*is set equalto a proper
`fraction of the previous value; i.e., the previous value is mul-
`tiplied by a first constant a,<1, for example, a,=0.9458. In
`this way, the value of the envelope coincides substantially
`with the value of the acceleration datum,if the acceleration
`datum is greater than the previous value of the envelope, and
`otherwise decreases(in particular, almost exponentially) with
`respectto the previous value.
`Likewise, block 35, if the value of the acceleration datum
`CalAccis smaller than the negative envelope Env” (as calcu-
`lated at the previous acquisition), the new value of the nega-
`tive envelope Env”is set equal to the value of the acceleration
`datum CalAcc, block 36. Otherwise, block 37, the value ofthe
`negative envelope Env” is set equal to a properfraction ofthe
`previous value of the envelope; i.e., it is multiplied by a
`second constant a,<1, for example, a,=0.9438. Note, in par-
`ticular, that the different value of the first and second con-
`stants ©, 21s due to the different value ofthe positive and
`negative accelerations, said negative accelerations being of
`smaller intensity, since the negative phase of the step is a
`reboundofthe positive phase.
`The algorithm then proceeds with updating ofthe values of
`the reference thresholds as a function of the envelope values
`previously calculated. In detail, block 38, the value of the
`positive reference threshold S* is set equal to a certain proper
`fraction of the positive envelope Env‘; in particular, it is set
`equalto the value of the positive envelope Env* multiplied by
`a third constant B<1, for example, B=0.65. However, if the
`value thus calculatedis less than the positive minimum value
`S,, block 39, the value ofthe positive reference threshold S*
`is set at said positive minimum valueS,, block 40.
`Likewise, block 41, the value of the negative reference
`threshold S~ is set equal to a certain proper fraction of the
`negative envelope; in particular, also this value is multiplied
`by thethird constant §. However, once again,ifthe value thus
`calculated is less than the negative minimum value S,, block
`42,the value ofthe negative reference threshold S~ is set at the
`negative minimum value S,, block 43.
`The values of the new reference thresholds thus calculated
`
`are then usedfor detection of the positive and negative phases
`of the step, as described previously.
`FIGS. 5 and 6 show the curves of the positive and negative
`reference thresholds S*, S~, and of the positive and negative
`envelopes Env*, Env’, calculated using the algorithm
`described previously, and the pattern of the acceleration sig-
`nal CalAcc (constituted by the sequence of the acceleration
`data CalAcc). It is evident that the reference thresholds sub-
`stantially follow the envelopes of the acceleration signal
`(which, in turn, follow the peaks of the acceleration signal).
`In detail, the value ofthe positive acceleration threshold S*
`is initially equal to the positive minimum value S, (see, in
`particular, FIG. 6), and remains constantas long as the accel-
`eration datum CalAcc remains less than the positive accel-
`eration threshold S*. Starting from the instant at which the
`acceleration datum CalAcc exceedsthe positive acceleration
`
`HTC v. Uniloc Luxembourg
`
`Page 9 of 12
`
`HTC Ex. 1005
`
`HTC v. Uniloc Luxembourg
`
`Page 9 of 12
`
`HTC Ex. 1005
`
`

`

`US 7,463,997 B2
`
`7
`threshold S:, and as long as the acceleration datum CalAcc
`increases,the positive acceleration threshold S* follows, ina
`“damped”way,the increase ofthe acceleration datum CalAcc
`(see,
`in particular, FIG. 5). Next, the acceleration datum
`CalAcc starts to decrease, and, along with it, the positive
`acceleration threshold S*, which, as long as the acceleration
`datum CalAcc decreases, assumesa decreasing pattern (with-
`out, however, dropping below the positive minimum value
`S,). In particular, at the end of the positive phase of the step,
`the maximum value S*,,,,,, is stored. The positive reference
`threshold S* returns to the positive minimum value S, when
`the user comesto a halt. A similar pattern (in absolute value)
`is showedbythe negative acceleration threshold S, with the
`difference thatthe decrease (in absolute value)ofthe negative
`acceleration threshold S~ is different, in particular faster. Said
`difference is due to the different conformation ofthe negative
`acceleration peak, which has a smaller amplitude anda longer
`duration as comparedto the positive acceleration peak,so that
`an excessively long decrease time could lead to the peak not
`being detected. The difference, in absolute value, ofthe posi-
`tive minimum value S, and ofthe negative minimum value S,
`is due to the samereason.
`According to one embodimentofthe present invention, the
`positive minimum value S, andthe negative minimum value
`S, can be modified from outside, for example through the
`interface 5 in order to modify the sensitivity of the pedometer
`device 1. In particular, if said minimum valuesare decreased,
`the sensitivity ofthe device increases, in so far as acceleration
`peaks of smaller amplitude (for example, due to a particularly
`slow gait or to a surface that is not very rigid) can be detected.
`At the same time, however, the number offalse positives
`detected increases, in so far as noise (external vibrations,
`bumps, fast movements made by the user) is more likely to
`cause erroneous detections assimilated to the phases of the
`step.
`The advantages of the pedometer device and ofthe corre-
`sponding step detection methodare clear from the foregoing
`description.
`In any case, it is emphasized that the pedometer device 1 is
`able to adapt to changes in the acceleration profile, for
`example due to an increase in the walking speed, and so
`external interventions for resetting the acceleration thresh-
`olds necessary for step detection are not needed.
`The fact that the acceleration thresholds follow the enve-
`lopes of the acceleration signal (analogously to an electronic
`peak detector) enables said changes to be followed rapidly,
`without any risk for any loss of steps and counting errors
`occurring, and at the same time enables a goodinsensitivity to
`noise to be achieved. In particular, when the accelerations
`increase (in absolute value), for example because the walking
`speed has increased, the reference thresholds increase rap-
`idly, so as to adapt rapidly to the new conditions. When,
`instead, the accelerations decrease, for example because the
`user is slowing down,the reference thresholds also decrease,
`but slowly, and always remaining above a minimum value. In
`this way, the device is able to follow closely a new increase in
`the acceleration values.
`
`Finally, it is clear that modifications and variations can be
`made to what is described and illustrated herein without
`
`thereby departing from the scopeofthe present invention, as
`defined in the appendedclaims.
`In particular, as shown in FIG.7, in which the samerefer-
`ence numbers are used for designating blocks similar to the
`ones previously described, according to an alternative
`embodimentof the present invention, the step detection algo-
`rithm can be simplified, and can be based exclusively upon
`the identification of the positive phase ofthe step (i.e., of the
`
`25
`
`30
`
`35
`
`40
`
`45
`
`55
`
`65
`
`8
`positive acceleration peak). In this case, the algorithm uses a
`single reference threshold, in particular the positive reference
`threshold S*, which is modified as a function of the value of
`the positive envelope Env*, in a way altogether similar to
`what has been described previously. Said simplified algo-
`rithm, although computationally less burdensomeforthe pro-
`cessing unit 3, has, however, the disadvantage of being more
`sensitive to noise. In fact, the lack of check on the presence of
`the negative phase, after the positive phase, renders false
`detection and counting errors more likely.
`The accelerometer 2 could be equipped with a number of
`axes of measurement, for example three mutually orthogonal
`axes of measurement, and be built, for example, as described
`in “3-axis Digital Output Accelerometer For Future Automo-
`tive Applications”, B. Vigna et al. AMAA 2004. In this case,
`according to one embodimentof the present invention, the
`algorithm implemented by the processing unit 3 envisages
`identifying the main vertical axis to be used forstep detection
`as the axis of detectionthat has the highest mean acceleration
`value Accm (on account of gravity). For example, the main
`vertical axis can be identified at each acquisition of a new
`acceleration sample, block 30 of FIG. 4, so as to take into
`account variations in the orientation of the pedometer device
`1, and consequently of the accelerometer 2 arrangedinsideit.
`Instead of being integrated in the pedometer device 1, the
`accelerometer 2 could be arranged outside the casing thereof,
`and connected, in a wired or wireless way, to the detection
`unit 3. In this case, the accelerometer 2 could advantageously
`be housed in a garment or accessory worn bythe user, for
`example a shoe, a belt, a watch, etc.
`As shown in FIG. 8, the pedometer device 1, due to its
`reduced dimensions, may advantageously be housedinside a
`portable device, in particular a mobile phone 50 (or else an
`Mp3reader, a camera, a PDA, a game console, etc.). In this
`case, the