CN109582713A - A kind of recognition methods of motion state, device and terminal - Google Patents
A kind of recognition methods of motion state, device and terminal Download PDFInfo
- Publication number
- CN109582713A CN109582713A CN201811455450.4A CN201811455450A CN109582713A CN 109582713 A CN109582713 A CN 109582713A CN 201811455450 A CN201811455450 A CN 201811455450A CN 109582713 A CN109582713 A CN 109582713A
- Authority
- CN
- China
- Prior art keywords
- data
- motion state
- section
- relative coefficient
- state
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
Abstract
The invention discloses a kind of recognition methods of motion state, applied to technical field of data processing, solve the problems, such as that walking line state and riding condition in the prior art cannot accurately distinguish, this method comprises: obtaining the data of motion sensor acquisition, and when determining that motion state is to walk or ride fringe according to data, relative coefficient is calculated using the expectation of data and preset delay time；When relative coefficient is located at the first relative coefficient section, then motion state is to walk line state；When relative coefficient is located at the second relative coefficient section, then motion state is riding condition；This method can accurately distinguish away line state and riding condition using relative coefficient, and then improve moving state identification accuracy；The invention also discloses a kind of identification device of motion state, terminal and computer readable storage mediums, have above-mentioned beneficial effect.
Description
Technical field
The present invention relates to technical field of data processing, in particular to a kind of recognition methods of motion state, device, terminal and
Computer readable storage medium.
Background technique
The identification of motion state generally uses Fourier transformation, wavelet transformation and mean value, standard deviation etc. in the prior art
Feature.But these are only applicable to theoretical research, and in practical applications, the quality of moving state identification result depends on completely
In the quality of collected data, especially in the fuzzy border regions of motion state two-by-two, the recognition result of motion state is not
Accurately, therefore, it is necessary to improve the accuracy of moving state identification.
Summary of the invention
The object of the present invention is to provide a kind of recognition methods of motion state, device, terminal and computer-readable storage mediums
Matter can accurately distinguish away line state and riding condition, and then improve moving state identification accuracy.
In order to solve the above technical problems, the present invention provides a kind of recognition methods of motion state, comprising:
The data of motion sensor acquisition are obtained, and works as and determines that motion state is to walk or ride fuzzy according to the data
When state, relative coefficient is calculated using the expectation of the data and preset delay time；
When the relative coefficient is located at the first relative coefficient section, then motion state is to walk line state；
When the relative coefficient is located at the second relative coefficient section, then motion state is riding condition.
Optionally, before the expectation using the data and preset delay time calculate relative coefficient,
Further include:
First sample entropy is calculated using the data；
When the first sample entropy is located at first sample entropy section, then motion state is to walk line state；
When the first sample entropy is located at the second Sample Entropy section, then motion state is riding condition；
When the first sample entropy is located at third Sample Entropy section, then execute the expectation using the data and
Preset delay time calculates the step of relative coefficient.
Optionally, before the calculating first sample entropy using the data, further includes:
The first kurtosis value is calculated using the data；
When first kurtosis value is located at the first kurtosis value section, then motion state is to walk line state；
When first kurtosis value is located at the second kurtosis value section, then motion state is riding condition；
When first kurtosis value is located at third kurtosis value section, then executes the utilization data and calculate the first sample
The step of this entropy.
Optionally, before first kurtosis value of calculating using the data, further includes:
Stationarity parameter is calculated using the data；
When the stationarity parameter is located at the first stationarity parameter section, then executes and described calculate the using the data
The step of one kurtosis value；
When the stationarity parameter is located at the second stationarity parameter section, the second Sample Entropy is calculated using the data；
When second Sample Entropy is greater than sample entropy threshold, then motion state is riding condition；When second Sample Entropy is less than institute
When stating sample entropy threshold, then motion state is to walk line state.
It is optionally, described to calculate stationarity parameter using the data, comprising:
The resultant acceleration for obtaining the corresponding 3-axis acceleration of preset window, determines the minimum resultant acceleration of the preset window
Min1；
Utilize formulaCalculate total resultant acceleration average value Average1；
Determine the minimum resultant acceleration Min2q that wicket is respectively slided in the preset window；
Utilize formulaCalculate the resultant acceleration average value of each sliding wicket
Average2q；
From the resultant acceleration average value of total resultant acceleration average value Average1 and each sliding wicket
Maximum value Max and minimum M in is determined in Average2q；
Using the ratio Max/Min of Max and Min as stationarity parameter AS；
Wherein, n is the corresponding resultant acceleration number of preset window, and X (i) is i-th of resultant acceleration, and Min2q is default window
The minimum resultant acceleration of q-th of sliding wicket in mouth, Xq (m) are that m-th of conjunction of q-th of sliding wicket in preset window adds
Speed, l are the resultant acceleration number for sliding wicket.
Optionally, when determining that motion state is to run or ride fringe according to the data, further includes:
The second kurtosis value is calculated using the data；
When second kurtosis value is less than the first kurtosis threshold value, then motion state is running state；
When second kurtosis value is greater than the first kurtosis threshold value, then motion state is riding condition.
Optionally, it is described according to the data determine motion state be walk or fringe of riding, comprising:
The resultant acceleration of the 3-axis acceleration obtained is calculated, and calculates the standard deviation of the resultant acceleration；
When the standard deviation is located at the first standard deviation section, then motion state be walk or fringe of riding.
It is optionally, described to calculate relative coefficient using the expectation of the data and preset delay time, comprising:
Utilize formula A (τ)=E [(Xt-μ)(Xt+τ- μ)] calculate separately the relevance parameter A that preset delay time is 0
(0) and preset delay time be n relevance parameter A (n)；
It regard ratio A (0)/A (n) of A (0) and A (n) as relative coefficient A；
Wherein, τ is delay time, XtFor the resultant acceleration obtained in preset time period t, Xt+τFor in preset time period t
The resultant acceleration that delay τ is obtained, μ XtExpectation, E be seek expectation computing.
The present invention also provides a kind of identification devices of motion state, comprising:
Relative coefficient computing module is determined for obtaining the data of motion sensor acquisition, and when according to the data
Motion state is to calculate correlation using the expectation of the data and preset delay time when walking or ride fringe
Coefficient；
First categorization module is used for when the relative coefficient is located at the first relative coefficient section, then motion state
To walk line state；When the relative coefficient is located at the second relative coefficient section, then motion state is riding condition.
The present invention also provides a kind of terminals, comprising:
Motion sensor, for acquiring data；
Memory, for storing computer program；
Processor, the step of recognition methods of motion state described above is realized when for executing the computer program.
The present invention also provides a kind of computer readable storage medium, calculating is stored on the computer readable storage medium
The step of machine program, the computer program realizes the recognition methods of motion state described above when being executed by processor.
A kind of recognition methods of motion state provided by the present invention, comprising: the data of motion sensor acquisition are obtained, and
When determining that motion state is to walk or ride fringe according to data, expectation and preset delay time using data
Calculate relative coefficient；When relative coefficient is located at the first relative coefficient section, then motion state is to walk line state；Work as phase
When pass property coefficient is located at the second relative coefficient section, then motion state is riding condition.
As it can be seen that this method distinguishes away line state and riding condition using relative coefficient, since relative coefficient characterizes
Degree of correlation of the same event between two different times, is able to reflect out the relevance of itself different moments, and walks
The motion process of the two is different with riding, i.e., event is different, and then the relative coefficient being calculated will not be identical, therefore benefit
Relative coefficient is calculated with the expectation of collected data and preset delay time, to obtain same event in two differences
Degree of correlation between period can accurately distinguish away line state and riding condition, and then improve moving state identification
Accuracy；The present invention also provides a kind of identification device of motion state, terminal and computer readable storage mediums, have above-mentioned
Beneficial effect, details are not described herein.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is the flow chart of the recognition methods of motion state provided by the embodiment of the present invention；
Fig. 2 is the schematic diagram of stationarity coefficients statistics result provided by the embodiment of the present invention；
Fig. 3 is the schematic diagram that sliding wicket divides provided by the embodiment of the present invention；
Fig. 4 is the structural block diagram of the identification device of motion state provided by the embodiment of the present invention.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
Every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
At present under the application scenarios for carrying out movement identification, all moved mostly by the feature of accelerometer data
Identification, such as the features such as Fourier transformation (FFT), wavelet transformation and mean value, standard deviation.But theoretical research is applied
When in practical application, the quality of moving state identification result depends entirely on the quality of collected data, especially two-by-two
The fuzzy border regions of motion state, the recognition result inaccuracy of motion state.The present embodiment to solve the above-mentioned problems, into
Relative coefficient is increased during row moving state identification, and then improves the accuracy of moving state identification.Specifically ask
It is the flow chart of the recognition methods of motion state provided by the embodiment of the present invention with reference to Fig. 1, Fig. 1；This method may include:
S110, obtain motion sensor acquisition data, and when according to data determine motion state be walk or mould of riding
When paste state, relative coefficient is calculated using the expectation of data and preset delay time.
Specifically, the present embodiment do not limit how according to motion sensor acquisition data determine current state be on foot
Or the mode for fringe of riding, do not limit the type of the data of motion sensor acquisition accordingly yet.Such as data can be
Acceleration either resultant acceleration, the corresponding average value or mean value that can use data, then either standard deviation with accordingly set
Fixed threshold interval is compared, and then determines that the corresponding motion state of data of acquisition may be to walk line state or may be
Riding condition, that is, walk or fringe of riding.It is to be understood that as long as user can determine its correspondence according to the data
Motion state may be to walk line state, it is also possible to be riding condition.It can fast and accurately be determined by subsequent step
The corresponding motion state of the data is specially to walk line state or riding condition, the final accurate identification realized to motion state.
Further, for the accuracy and computational efficiency of the fringe identification that guarantees on foot or ride.The present embodiment is logical
It crosses resultant acceleration and standard deviation and corresponds to specification area come the fringe that determines on foot or ride.It is obtained in specific the present embodiment
The data of motion sensor acquisition, and work as and determine that motion state is on foot or fringe of riding may include: according to data
The resultant acceleration of the 3-axis acceleration obtained is calculated, and calculates the standard deviation of resultant acceleration；
When standard deviation is located at the first standard deviation section, then motion state be walk or fringe of riding.
Specifically, not limiting the quantity of the 3-axis acceleration of acquisition in the present embodiment, user can be according to actual meter
Calculate accuracy requirement, setting moving state identification calculatings frequency and motion sensor acquire frequency etc. of data to determine.
Such as call within user 8 seconds the corresponding algorithm of recognition methods of primary motion state provided in this embodiment, then corresponding acquisition
The data of motion sensor acquisition are exactly collected data in this 8 seconds.It is specific amount of how much also to be acquired with motion sensor
The frequency dependence of data.Such as when the frequency of motion sensor acquisition data is 26Hz, i.e. 26 data of acquisition per second, corresponding 8
Second obtains 208 data altogether.After getting corresponding data, the standard deviation of corresponding resultant acceleration and resultant acceleration is calculated.
The calculating process is not defined in the present embodiment, user can add with reference to calculating resultant acceleration in the related technology and closing
The concrete mode of the standard deviation of speed.
Wherein, the present embodiment is not defined the specific data interval in the first standard deviation section, and user can basis
Practical application scene is determined.When standard deviation be located at the first standard deviation section be determined motion state at this time be walk or
It rides fringe.Certainly, user can also with reference in the related technology walk or the corresponding standard deviation section of fringe of riding
The concrete mode of setting.
The present embodiment is determining that standard deviation is located at the first standard deviation section, that is, determines that motion state is to walk or ride fuzzy
When state, need to calculate the relative coefficient that can will be walked or fringe of riding accurately distinguishes.In the present embodiment simultaneously
The concrete mode for calculating relative coefficient is not defined.User can calculate energy according to practical computational accuracy and hardware
Power selects corresponding calculation method.Such as this implementation is calculated using pearson correlation (Pearson correlation) algorithm
Relative coefficient in example.Wherein, auto-correlation is defined as the pearson correlation in random process between the numerical value of different moments.
It specifically can use correlation functionTo calculate relative coefficient, wherein correlation function R
(τ) can be expressed as the function of delay time T, XtFor the resultant acceleration obtained in preset time period t, Xt+τFor when default
Between the obtained resultant acceleration of section t delay τ, μ XtExpectation, σ XtStandard deviation, E be seek expectation computing.
Certainly, user can also modify to the calculation method of existing relative coefficient, obtain being more suitable for movement shape
State identifies the calculation method of the customized relevance parameter in field.Since relative coefficient characterizes same event at two not
Degree of correlation between same time is able to reflect out the relevance of itself different moments.Therefore, by preset in the present embodiment
Delay time obtains two different times of same event, and the expectations of the data acquired in conjunction with motion sensor calculates phase
Close property coefficient.Further, answering for above-mentioned relative coefficient calculating is reduced in the present embodiment using customized relative coefficient
Polygamy improves computational efficiency.Preferably, correlation is calculated using the expectation of data and preset delay time in the present embodiment
Coefficient may include:
Utilize formula A (τ)=E [(Xt-μ)(Xt+τ- μ)] calculate separately the relevance parameter A that preset delay time is 0
(0) and preset delay time be n relevance parameter A (n)；By A (0) to the ratio A (0) of A (n)/A (n) as related
Property coefficient A.
It should be noted that A (0) is exactly to be brought into τ=0 in A (τ) to obtain A (0)=E [(Xt-μ)(Xt- μ)], similarly A
It (n) is exactly that τ=n is brought into A (τ) to obtain A (n)=E [(Xt-μ)(Xt+n-μ)].Wherein, τ is delay time, XtFor pre-
If the resultant acceleration obtained in time period t, Xt+τFor the resultant acceleration obtained in preset time period t delay τ, μ XtExpectation, E
To ask expectation computing, Xt+nFor the resultant acceleration obtained in preset time period t delay n.It is understood that in the present embodiment simultaneously
The specific time numerical value that preset delay time is n is not defined, can be set according to actual use scene.Such as
N can be 0.5 second either 1 second etc..The present embodiment is not also defined the numerical value of preset time period t, such as can be tune
It is 8 seconds with the time of the corresponding algorithm of recognition methods of primary motion state provided in this embodiment, such as preset time period t.When
So, moving state identification accuracy and computational efficiency are considered in the present embodiment, it is preferred that the value of n can take 1, and unit is
Second.Corresponding relative coefficient A at this time specifically: A (0)/A (1), wherein A (1) is exactly to be brought into τ=1 in A (τ) to obtain A
(1)=E [(Xt-μ)(Xt+1- μ)], Xt+1For in preset time period t 1 second obtained resultant acceleration of delay.
S120, when relative coefficient is located at the first relative coefficient section, then motion state is to walk line state.
S130, when relative coefficient is located at the second relative coefficient section, then motion state be riding condition.
Specifically, not to the specific of the first relative coefficient section and the second relative coefficient section in the present embodiment
Data interval is defined, and user can be determined according to practical application scene.Due to walking the corresponding correlation system of line state
Number is less than the corresponding relative coefficient of riding condition, therefore the upper limit value in the first relative coefficient section should be the second correlation
The lower limit value in coefficient section.Such as when the first relative coefficient section be less than or equal to the first relative coefficient range, second
Relative coefficient section is the range greater than the first relative coefficient.Such as when the first relative coefficient is A1, the first correlation
Coefficient section is the range less than or equal to A1, and the second relative coefficient section is the range greater than A1.It is of course also possible to be first
Relative coefficient section is the range less than A1, and the second relative coefficient section is the range more than or equal to A1.That is the present embodiment
In do not limit which section is the corresponding value of interval endpoint particularly belong to, user can be determined according to the actual situation.
The shortcomings that in order to overcome in the related technology, improves the accuracy of moving state identification, and the present embodiment is utilized and calculated
To relative coefficient to walk or fringe of riding accurately distinguished.Detailed process can be, it is first determined current
Motion state, which is in, walks or fringe of riding (such as can quickly be determined whether by calculating simple standard deviation
In walk or fringe of riding), then calculate relative coefficient, and the relative coefficient area according to locating for relative coefficient
Between, accurately determine corresponding motion state (being currently to walk line state or riding condition).
Based on the above-mentioned technical proposal, the recognition methods of motion state provided in an embodiment of the present invention, this method utilize correlation
Property coefficient distinguishes away line state and riding condition, avoids moving state identification confounding issues, can accurately distinguish and leave
Line state and riding condition, and then improve moving state identification accuracy.
Based on the above embodiment, determining that motion state is that can calculate on foot or directly correlation after fringe of riding
Coefficient, using relative coefficient come to walk or fringe of riding accurately distinguished, naturally it is also possible to be determine fortune
Dynamic state is to carry out movement knowledge first with the feature of the data of other motion sensors acquisition on foot or after fringe of riding
Not, from walk or fringe of riding in distinguish a part accurately walk line state and riding condition, other features cannot
To the part walked or fringe of riding is accurately distinguished, this feature of relative coefficient is recycled to distinguish.In this way
The calculation amount of relative coefficient can be reduced on the basis of guaranteeing moving state identification accuracy.Specifically, in the present embodiment
Can also include: before calculating relative coefficient using the expectation of data and preset delay time
First sample entropy is calculated using data；
When first sample entropy is located at first sample entropy section, then motion state is to walk line state；
When first sample entropy is located at the second Sample Entropy section, then motion state is riding condition；
When first sample entropy is located at third Sample Entropy section, then when executing the expectation and preset delay using data
Between calculate relative coefficient the step of.
Specifically, Sample Entropy is the detection method of time series complexity, Sample Entropy is bigger, and sequence is more complicated, periodically
It is poorer.The acceleration signal of human motion has some cycles, it is possible to carry out human motion state knowledge using Sample Entropy
Not.The present embodiment does not limit the concrete form that Sample Entropy is calculated using data, and user can be according to the tool of the data of its acquisition
The calculating of body type progress respective sample entropy.
Not to the specific of first sample entropy section, the second Sample Entropy section and third Sample Entropy section in the present embodiment
Data interval is defined, and user can be determined according to practical application scene.It is corresponding due under normal circumstances, walking line state
Sample Entropy be less than the corresponding Sample Entropy of riding condition, therefore the upper limit value in first sample entropy section should be third Sample Entropy area
Between lower limit value, the upper limit value in third Sample Entropy section should be the lower limit value in the second Sample Entropy section.Such as work as first sample
Entropy section is less than the range of the first setting Sample Entropy, and third Sample Entropy section is more than or equal to the first setting Sample Entropy and to be less than
Equal to the range of the second setting Sample Entropy, the second Sample Entropy section is greater than the range of the second setting Sample Entropy, i.e., the first setting
Sample Entropy is less than the second setting Sample Entropy.Such as when first sets Sample Entropy as S1, second when setting Sample Entropy as S2, the first sample
This entropy section is the range less than S1, and third Sample Entropy section is the second sample more than or equal to S1 and in the range of being less than or equal to S2
This entropy section is the range greater than S2.It is of course also possible to be first sample entropy section be range less than or equal to S1, third sample
Entropy section is greater than S1 and in the range of being less than S2, and the second Sample Entropy section is the range more than or equal to S2.I.e. in the present embodiment
Do not limit which section is the corresponding value of each interval endpoint particularly belong to, user can be determined according to the actual situation.
In the present embodiment when first sample entropy is located at first sample entropy section, then it can accurately determine that motion state is
Walk line state；When first sample entropy is located at the second Sample Entropy section, then it can accurately determine that motion state is riding condition；
Only when first sample entropy is located at third Sample Entropy section, by Sample Entropy, this feature can accurately not distinguished currently
Motion state walk line state or riding condition on earth, just need to calculate relative coefficient at this time, utilize relevance parameter
(specific process of distinguishing can refer to last embodiment) is accurately distinguished to the state in this obscure portions region.
Further, can also include: before calculating first sample entropy using data in the above-described embodiments
The first kurtosis value is calculated using data；
When the first kurtosis value is located at the first kurtosis value section, then motion state is to walk line state；
When the first kurtosis value is located at the second kurtosis value section, then motion state is riding condition；
When the first kurtosis value is located at third kurtosis value section, then the step of calculating first sample entropy using data is executed.
Specifically, in the present embodiment according to collected data determine motion state be in walk or fringe of riding
When, first with this feature of kurtosis value to walk or fringe of riding distinguish, kurtosis value cannot to walk or ride
The part that row fringe is accurately distinguished reuses Sample Entropy this feature and distinguishes, and detailed process can refer to upper
State embodiment.The present embodiment does not limit the concrete form that the first kurtosis value is calculated using data, and user can be with reference to related skill
Art.
In the present embodiment select this feature of kurtosis value come distinguish on foot or ride fringe the reason of be kurtosis value meter
It calculates simply, calculating speed is fast.Several features are combined in the present embodiment, it can be in the accuracy for guaranteeing moving state identification
On the basis of in turn ensure computational efficiency.First by setting multiple kurtosis value sections, realize to be easy to happen obscure region and
Line state can be accurately identified away and riding condition region distinguishes.Next obscures region use in being easy to happen for marking off
Accurate high Sample Entropy is calculated come the identification for the fringe that executes on foot or ride, same pass through sets multiple Sample Entropy sections,
It realizes and obscures region and line state can be accurately identified away and riding condition region is distinguished to being easy to happen.Then it is dividing
Being easy to happen out obscures region using the accurate higher relative coefficient of calculating come the knowledge for the fringe that executes on foot or ride
Not, finally by multiple relative coefficient sections realize to walk or fringe of riding accurate identification.To by several
The combination of feature complete to walk or fringe of riding accurately identify, guarantee moving state identification accuracy base
Computational efficiency is in turn ensured on plinth.
Not to the specific of the first kurtosis value section, the second kurtosis value section and third kurtosis value section in the present embodiment
Data interval is defined, and user can be determined according to practical application scene.It is corresponding due under normal circumstances, walking line state
Kurtosis value be less than the corresponding kurtosis value of riding condition, therefore the upper limit value in the first kurtosis value section should be third kurtosis value area
Between lower limit value, the upper limit value in third kurtosis value section should be the lower limit value in the second kurtosis value section.Such as when the first kurtosis
Being worth section is less than the range of the first setting kurtosis value, and third kurtosis value section is more than or equal to the first setting kurtosis value and to be less than
Equal to the range of the second setting kurtosis value, the second kurtosis value section is greater than the range of the second setting kurtosis value, i.e., the first setting
Kurtosis value is less than the second setting kurtosis value.Such as when first sets kurtosis value as K1, second when setting kurtosis value as K2, first peak
Angle value section is the range less than K1, and third kurtosis value section is the second peak more than or equal to K1 and in the range of being less than or equal to K2
Angle value section is the range greater than K2.It is of course also possible to be the first kurtosis value section be range less than or equal to K1, third kurtosis
Being worth section is greater than K1 and in the range of being less than K2, and the second kurtosis value section is the range more than or equal to K2.I.e. in the present embodiment
Do not limit which section is the corresponding value of each interval endpoint particularly belong to, user can be determined according to the actual situation.
In the present embodiment when the first kurtosis value is located at the first kurtosis value section, then it can accurately determine that motion state is
Walk line state；When the first kurtosis value is located at the second kurtosis value section, then it can accurately determine that motion state is riding condition；
Only when the first kurtosis value is located at third kurtosis value section, by kurtosis value, this feature can accurately not distinguished currently
Motion state walk line state or riding condition on earth, just need to calculate Sample Entropy at this time, using Sample Entropy to be in this
The state in obscure portions region is accurately distinguished；It, then can be accurate when first sample entropy is located at first sample entropy section
Determination motion state be to walk line state；When first sample entropy is located at the second Sample Entropy section, then fortune can be accurately determined
Dynamic state is riding condition；Only when first sample entropy is located at third Sample Entropy section, simultaneously by Sample Entropy this feature
Current motion state can not accurately be distinguished and walk line state or riding condition on earth, just need to calculate correlation system at this time
Number accurately distinguishes the state in this obscure portions region using relevance parameter.
It is, of course, understood that user can not also select this feature of Sample Entropy, that is to say, that when the first kurtosis value
When positioned at the first kurtosis value section, then it can accurately determine that motion state is to walk line state；When the first kurtosis value is located at second
When kurtosis value section, then it can accurately determine that motion state is riding condition；Only it is being located at third peak when the first kurtosis value
When angle value section, by kurtosis value this feature can accurately not distinguish current motion state walk on earth line state or
This feature of relative coefficient can be directly used also at this time to carry out subsequent identification in riding condition.Area is unable in kurtosis value
The region divided directly calculates relative coefficient, using relative coefficient to the state for being in the undistinguishable fuzzy region of kurtosis value
Accurately distinguished.
Based on the above-mentioned technical proposal, the recognition methods of motion state provided in an embodiment of the present invention, this method utilize kurtosis
Value, Sample Entropy and relative coefficient distinguish away line state and riding condition, avoid moving state identification confounding issues, energy
It is enough accurately to distinguish away line state and riding condition, and then improve moving state identification accuracy.And pass through multiple features
Combination, played the advantage of every kind of feature, avoid each feature disadvantage (it is small to have played kurtosis value calculation amount, calculate
Fireballing advantage, recognition result is bigger than normal by calculation amount in the disadvantage of some regions inaccuracy, but it is high to calculate accuracy rate
Sample Entropy and relative coefficient solve, and have played Sample Entropy and relative coefficient and calculated the high advantage of accuracy rate), therefore subtract
Lack calculation amount, improves computational efficiency.
Based on the above embodiment, in order to further increase the arithmetic speed that movement identifies, operand can be greatly reduced, with section
About code power consumption, the present embodiment can also include: before calculating the first kurtosis value using data
Stationarity parameter is calculated using data；
When stationarity parameter is located at the first stationarity parameter section, then the step that the first kurtosis value is calculated using data is executed
Suddenly；
When stationarity parameter is located at the second stationarity parameter section, the second Sample Entropy is calculated using data；When the second sample
When this entropy is greater than sample entropy threshold, then motion state is riding condition；When the second Sample Entropy is less than sample entropy threshold, then move
State is to walk line state.
In the present embodiment when determining on foot or riding fringe, in order to improve the efficiency of subsequent calculating, pass through first
For stationarity parameter by subsequent to walking or the identification of fringe of riding is divided into two parts, the two parts select different knowledges
Not strategy to walk or fringe of riding accurately distinguished, and then realize meet walk or ride fringe identify
On the basis of accuracy, identification arithmetic speed is greatly improved, operand can be greatly reduced, saves code power consumption.
Fig. 2 can specifically be referred to, wherein * is to walk, namely walks line state ,+to ride namely riding condition.In Fig. 2
Mass data is tested, the longitudinal axis is that stationarity parameter is AS, and horizontal axis indicates test quantity, determines that stationarity parameter is located at
Second stationarity parameter section, that is, the corresponding data of line state of walking of black line area above are only 7 in figure.That is
What stationarity parameter was located at the second stationarity parameter section is essentially all the corresponding data of riding condition.The shape of walking of only a few
The stationarity parameter that the corresponding data of state are calculated is bigger than normal, walks the stationarity ginseng that the corresponding data of line state are calculated substantially
Number is all less than normal.Therefore, for data more than black line in the present embodiment, it is only necessary to by Sample Entropy distinguish in the way of execute
?.For black line data below, using distinguished first with kurtosis value (though cannot all accurately distinguish, can be with
Greatly reduce calculation amount), recycle Sample Entropy to distinguish, the mode for finally relative coefficient being recycled to distinguish, which executes, (to be passed through
Sample Entropy and relative coefficient can accurately complete to walk or fringe of riding identification).It can certainly be direct elder generation
It is distinguished using Sample Entropy, the mode for then relative coefficient being recycled to distinguish executes, or merely with relative coefficient
The mode of differentiation executes, these modes can realize to walk or fringe of riding accurately identify, difference be only that meter
Calculation amount is different and calculating parameter quantity is different.Wherein, various executive modes can be with reference to the tool in above-mentioned each embodiment
Body process.
It is to be understood that walking acceleration signal of line state itself has some cycles, it is also opposite that there is regularity, institute
To utilize stationarity numerical value, it can greatly reduce to the calculation amount walked or fringe of riding is accurately distinguished, improve fortune
Calculate speed.In addition, relevance parameter can accurately identify motion state again.So being guaranteed in the present embodiment by steady property coefficient
The calculating speed of moving state identification guarantees the accurate of moving state identification by kurtosis value, Sample Entropy and relative coefficient
Degree realizes faster more accurately identification motion state.
The concrete mode for calculating stationarity parameter is not defined in the present embodiment.User can be according to practical calculating
Precision and hardware computing capability select corresponding calculation method.In order to improve the reliability of stationarity parameter calculating.This implementation
May include: using the process that data calculate stationarity parameter in example
The resultant acceleration for obtaining the corresponding 3-axis acceleration of preset window, determines the minimum resultant acceleration of preset window
Min1；
Utilize formulaCalculate total resultant acceleration average value Average1；
Determine the minimum resultant acceleration Min2q that wicket is respectively slided in preset window；
Utilize formulaCalculate the resultant acceleration average value of each sliding wicket
Average2q；
From total resultant acceleration average value Average1 and each resultant acceleration average value Average2q for sliding wicket really
Determine maximum value Max and minimum M in；
Using the ratio Max/Min of Max and Min as stationarity parameter AS；
Wherein, n is the corresponding resultant acceleration number of preset window, and X (i) is i-th of resultant acceleration, and Min2q is default window
The minimum resultant acceleration of q-th of sliding wicket in mouth, Xq (m) are that m-th of conjunction of q-th of sliding wicket in preset window adds
Speed, l are the resultant acceleration number for sliding wicket.
Specifically, not being defined to the size of preset window in the present embodiment, can be passed according to the movement of acquisition
The data corresponding period of sensor acquisition is set, it is understood that call primary motion state provided in this embodiment
The period of the corresponding algorithm of recognition methods.Such as when the data for participating in moving state identification calculating are the data obtained in 8 seconds,
Corresponding preset window is sized for 8 seconds corresponding windows.The minimum resultant acceleration Min1 of corresponding preset window at this time, i.e.,
It can be understood as the minimum resultant acceleration Min1 in resultant acceleration whole in 8 seconds windows, wherein if motion sensor acquires number
According to frequency be 26Hz, i.e. 26 data of acquisition per second, corresponding 8 seconds 208 data of acquisition altogether.Specifically seek the process of Min1
May is that Min1=min (X (1), X (2), X (3) ... X (208)), wherein min () is the operator minimized.
Need to calculate total resultant acceleration average value Average1 after determining minimum resultant acceleration Min1.Wherein,
Average1 is exactly after each resultant acceleration in preset window is cut minimum resultant acceleration, and it is right in preset window at this time to calculate
The average value of data is answered, and is referred to as total resultant acceleration average value Average1.I.e. using the formulae express above process:Wherein, X (i) is i-th of resultant acceleration in preset window, and n is that preset window is corresponding
Resultant acceleration number.Specifically, corresponding formula is exactly when the quantity of resultant acceleration in preset window is 208
The quantity for sliding wicket included in preset window is not limited in the present embodiment, and slides wicket yet
Generating mode.Such as when preset window is 8 seconds windows, it can successively slide and be slided corresponding window of default second as one
Wicket.The present embodiment is not defined the numerical value of default second, such as can be 2 seconds.When each sliding wicket is all corresponding
When for 2 seconds windows, the preset window of 8 seconds windows can correspond to 7 sliding wickets, and respectively the 1st second and the 2nd second is the 1st
A sliding wicket, the 2nd second and the 3rd second is the 2nd sliding wicket, and the 3rd second and the 4th second is the 3rd sliding wicket ...,
7th second and the 8th second is the 7th sliding wicket, can specifically refer to Fig. 3.Just need to obtain each sliding wicket at this time
Minimum resultant acceleration Min2q.Minimum resultant acceleration Min2q in i.e. each sliding wicket in whole resultant accelerations.Namely
Say 7 sliding wickets if it exists at this time, then the range of corresponding q is exactly 1 to 7, can get 7 sliding wickets most
Small resultant acceleration.If the frequency that motion sensor acquires data is 26Hz, i.e. 26 data of acquisition per second are then each to slide small window
Mouthfuls of corresponding 2 seconds totally 52 data, that is, each sliding wicket need to determine the smallest conjunction acceleration from 52 resultant accelerations
Degree.
The detailed process that the resultant acceleration average value Average2q of each sliding wicket is calculated in the present embodiment can refer to
The calculating process of resultant acceleration average value Average1.Each resultant acceleration in i.e. q-th sliding wicket cuts the sliding
After the corresponding minimum resultant acceleration of wicket, the average value of corresponding data in q-th of sliding wicket at this time is calculated, and is claimed
The resultant acceleration average value Average2q of wicket is slided for q-th.Wherein, the value range of q is 1 to sliding wicket
The corresponding numerical value of total quantity.I.e. using the formulae express above process:Wherein, Xq
It (m) is m-th of resultant acceleration of q-th of sliding wicket in preset window, l is the resultant acceleration number for sliding wicket,
Min2q is the minimum resultant acceleration of q-th of sliding wicket in preset window.Specifically, when resultant acceleration in preset window
When quantity is 52, corresponding formula is exactlyAnd when the quantity of sliding wicket
When being 7, the value range of q is 1,2,3,4,5,6,7.
From total resultant acceleration average value Average1 and each resultant acceleration average value Average2q for sliding wicket really
Determine maximum value Max and minimum M in.It is namely corresponding from obtained total resultant acceleration average value and each sliding wicket
Maximum resultant acceleration numerical value Max and the smallest resultant acceleration numerical value Min is determined in resultant acceleration average value.When the small window of sliding
When the quantity of mouth is 7, then the process is exactly the maximizing Max and minimum M in from 8 data.Finally by Max with
The ratio Max/Min of Min is as stationarity parameter AS.
Specifically, not to the specific of the first stationarity parameter section and the second stationarity parameter section in the present embodiment
Data interval is defined, and user can be determined according to practical application scene.Since the first stationarity parameter section is corresponding
Be black line region below in Fig. 2, therefore the upper limit value in the first stationarity parameter section should be the second stationarity parameter area
Between lower limit value.Such as when the first stationarity parameter section is the range less than or equal to the first stationarity parameter, the second stationarity
Parameter section is the range greater than the first stationarity parameter.Such as when the first stationarity parameter is AS1, the first stationarity parameter area
Between for less than or equal to AS1, the second stationarity parameter section is greater than AS1.It is of course also possible to be that the first stationarity parameter section is
Less than AS1, the second stationarity parameter section is more than or equal to AS1.The corresponding value of interval endpoint is not limited in the present embodiment
Which section is particularly belonged to, user can be determined according to the actual situation.
Based on the above-mentioned technical proposal, the recognition methods of motion state provided in an embodiment of the present invention, this method is using steadily
Property coefficient and relative coefficient can effectively be easy to appear the situation obscured to motion state and carry out accurately to distinguish motion state
Differentiation, and reduce calculation amount, improve the computational efficiency of moving state identification, and improve hardware utilization.
Based on above-mentioned any embodiment, in the present embodiment when according to data determine motion state for running or fuzzy shape of riding
Can also include: when state
The second kurtosis value is calculated using data；
When the second kurtosis value is less than the first kurtosis threshold value, then motion state is running state；
When the second kurtosis value is greater than the first kurtosis threshold value, then motion state is riding condition.
Specifically, the present embodiment do not limit how according to motion sensor acquisition data determine current state be run
Or the mode for fringe of riding, do not limit the type of the data of motion sensor acquisition accordingly yet.Such as data can be
Acceleration either resultant acceleration, the corresponding average value or mean value that can use data, then either standard deviation with accordingly set
Fixed threshold interval is compared, and then determines that the corresponding motion state of data of acquisition may be running state or may be
Riding condition, that is, fringe of running or ride.It is to be understood that as long as user can determine its correspondence according to the data
Motion state may be running state, it is also possible to be riding condition.It can fast and accurately be determined by subsequent step
The corresponding motion state of the data is specially running state or riding condition, the final accurate identification realized to motion state.
Further, for the accuracy and computational efficiency of the fringe identification that guarantees to run or ride.The present embodiment is logical
Resultant acceleration and standard deviation are crossed to determine that running or fringe of riding correspond to specification area.It is obtained in specific the present embodiment
The data of motion sensor acquisition, and work as and determine that motion state may include: for running or fringe of riding according to data
The resultant acceleration of the 3-axis acceleration obtained is calculated, and calculates the standard deviation of resultant acceleration；
When standard deviation is located at the second standard deviation section, then motion state is running or fringe of riding.
Specifically, the quantity of the 3-axis acceleration of acquisition is not limited according to the explanation of above-described embodiment in the present embodiment,
User can adopt according to actual computational accuracy demand, the calculating frequency of the moving state identification of setting and motion sensor
Collect frequency of data etc. to determine.After getting corresponding data, the standard of corresponding resultant acceleration and resultant acceleration is calculated
Difference.The calculating process is not defined in the present embodiment, user can with reference in the related technology calculate resultant acceleration and
The concrete mode of the standard deviation of resultant acceleration.
Wherein, the present embodiment is not defined the specific data interval in the second standard deviation section, and user can basis
Practical application scene is determined.When standard deviation be located at the second standard deviation section be determined motion state at this time for running or
It rides fringe.Certainly, user can also be with reference to the corresponding standard deviation section of fringe of running or ride in the related technology
The concrete mode of setting.Under normal conditions, the corresponding standard deviation of fringe of running or ride be greater than walk or fuzzy shape of riding
State.Therefore the lower limit value in the second standard deviation section is more than or equal to the upper limit value in the first standard deviation section.
It can be by running state and riding condition area by single kurtosis threshold value i.e. the first kurtosis threshold value in the present embodiment
It separates, compared in such a way that Sample Entropy distinguishes running state and riding condition, kurtosis value calculates simpler, it is contemplated that calculates speed
Memory is spent and saves, this feature comes subregion running state and riding condition using kurtosis for selection in the present embodiment.Kurtosis and sample
This entropy is all basic data statistical characteristics amount, its principle that details are not described herein again.Certainly, first peak is not limited in the present embodiment
When spending the specific value of threshold value, and not limiting the second kurtosis value equal to the first kurtosis threshold value, corresponding motion state is specific
For running state or riding condition.For that can be set according to the actual situation.
Based on the above-mentioned technical proposal, Sample Entropy is replaced using kurtosis value in the embodiment of the present invention, is not influencing running state
With riding condition identify accuracy on the basis of, reduce data calculation amount, improve running state and riding condition identification
The utilization rate of efficiency and memory space.
Below to the identification device of motion state provided in an embodiment of the present invention, terminal and computer readable storage medium into
Row is introduced, identification device, terminal and the computer readable storage medium of motion state described below and above-described movement
The recognition methods of state can correspond to each other reference.
Referring to FIG. 4, Fig. 4 is the system block diagram of the identification device of motion state provided by the embodiment of the present invention, the dress
It sets and may include:
Relative coefficient computing module 100 is transported for obtaining the data of motion sensor acquisition, and when being determined according to data
Dynamic state is to calculate relative coefficient using the expectation of data and preset delay time when walking or ride fringe；
First categorization module 200, for when relative coefficient is located at the first relative coefficient section, then motion state to be
Walk line state；When relative coefficient is located at the second relative coefficient section, then motion state is riding condition.
Based on the above embodiment, which can also include:
Sample Entropy computing module, for calculating first sample entropy using data；
Second categorization module, for when first sample entropy is located at first sample entropy section, then motion state to be shape of walking
State；When first sample entropy is located at the second Sample Entropy section, then motion state is riding condition；When first sample entropy is located at third
When Sample Entropy section, then relative coefficient computing module 100 is triggered.
Based on the above embodiment, which can also include:
First kurtosis value computing module, for calculating the first kurtosis value using data；
Third categorization module, for when the first kurtosis value is located at the first kurtosis value section, then motion state to be shape of walking
State；When the first kurtosis value is located at the second kurtosis value section, then motion state is riding condition；When the first kurtosis value is located at third
When kurtosis value section, then Sample Entropy computing module is triggered.
Based on the above embodiment, which can also include:
Stationarity parameter computing module, for calculating stationarity parameter using data；
4th categorization module, for when stationarity parameter is located at the first stationarity parameter section, then triggering the first kurtosis
It is worth computing module；When stationarity parameter is located at the second stationarity parameter section, the second Sample Entropy is calculated using data；When second
When Sample Entropy is greater than sample entropy threshold, then motion state is riding condition；When the second Sample Entropy is less than sample entropy threshold, then transport
Dynamic state is to walk line state.
Based on the above embodiment, stationarity parameter computing module may include:
First computing unit determines preset window for obtaining the resultant acceleration of the corresponding 3-axis acceleration of preset window
Minimum resultant acceleration Min1；
Second computing unit, for utilizing formulaCalculate total resultant acceleration average value
Average1；
Third computing unit, for determining the minimum resultant acceleration Min2q for respectively sliding wicket in preset window；
4th computing unit, for utilizing formulaCalculate each sliding wicket
Resultant acceleration average value Average2q；
5th computing unit is flat for the resultant acceleration from total resultant acceleration average value Average1 and each sliding wicket
Maximum value Max and minimum M in is determined in mean value Average2q, and is joined the ratio Max/Min of Max and Min as stationarity
Number AS；
Wherein, n is the corresponding resultant acceleration number of preset window, and X (i) is i-th of resultant acceleration, and Min2q is default window
The minimum resultant acceleration of q-th of sliding wicket in mouth, Xq (m) are that m-th of conjunction of q-th of sliding wicket in preset window adds
Speed, l are the resultant acceleration number for sliding wicket.
Based on above-mentioned any embodiment, which can also include:
Second kurtosis value computing module, for calculating the second kurtosis value using data；
5th categorization module, for when the second kurtosis value is less than the first kurtosis threshold value, then motion state to be running state；
When the second kurtosis value is greater than the first kurtosis threshold value, then motion state is riding condition.
Based on above-mentioned any embodiment, relative coefficient computing module 100 may include:
On foot or fringe determination unit of riding, it for calculating the resultant acceleration of the 3-axis acceleration obtained, and calculates
The standard deviation of resultant acceleration；When standard deviation is located at the first standard deviation section, then motion state be walk or fringe of riding；
Relative coefficient computing unit, for utilizing formula A (τ)=E [(Xt-μ)(Xt+τ- μ)] calculate separately preset prolong
When the time be 0 relevance parameter A (0) and preset delay time be n relevance parameter A (n)；By A (0) with A's (n)
Ratio A (0)/A (n) is used as relative coefficient A；
Wherein, τ is delay time, XtFor the resultant acceleration obtained in preset time period t, Xt+τFor in preset time period t
The resultant acceleration that delay τ is obtained, μ XtExpectation, E be seek expectation computing.
It should be noted that being based on above-mentioned any embodiment, device be can be based on programmable logic device realization, can
Programmed logic device includes FPGA, CPLD, single-chip microcontroller, processor etc..
The present embodiment also provides a kind of terminal, comprising: motion sensor, for acquiring data；Memory, based on storing
Calculation machine program；Processor realizes the identification of the motion state as described in above-mentioned any embodiment when for executing computer program
The step of method.The data for obtaining motion sensor and acquiring are realized when executing computer program such as processor, and when according to data
Determine that motion state is to calculate correlation using the expectation of data and preset delay time when walking or ride fringe
Coefficient；When relative coefficient is located at the first relative coefficient section, then motion state is to walk line state；When relative coefficient position
When the second relative coefficient section, then motion state is riding condition.
Wherein, the present embodiment is not defined motion sensor, can be specific according to the selection of user's calculating parameter
It determines.Optionally, motion sensor is specially 3-axis acceleration sensor.
Wherein, the present embodiment is not defined terminal, which can be smartwatch, Intelligent bracelet, tracker
Etc. intelligent wearable devices, be also possible to the mobile terminals such as mobile phone.It is positioned when using terminal, such as using tracker to old
When people, children or valuables position, which, which classifies, can be used as the basis on location of terminal, by accurately transporting
Dynamic state recognition result provides foundation for the positioning strategy of tracker.For example, when being continuously in stationary state, without frequently fixed
Position, positioning in 1 hour is primary, and when walking line state, positioning in 4 minutes is primary, and when state of running, positioning in 2 minutes is primary, other states
There is corresponding positioning strategy, does not repeat herein.In this way, the function of terminal can be greatly reduced by accurate moving state identification
Consumption improves the standby of terminal and using duration, promotes user experience.It is of course also possible to which motion state is classified as and terminal
Other connected positioning devices provide basis on location.
The present embodiment also provides a kind of computer readable storage medium, is stored with computer on computer readable storage medium
Program, when computer program is executed by processor the step of the realization such as recognition methods of above-mentioned any embodiment motion state.Such as
The data for obtaining motion sensor acquisition are realized when computer program is executed by processor, and are worked as and determined motion state according to data
For walk or ride fringe when, utilize the expectation of data and preset delay time to calculate relative coefficient；Work as correlation
When property coefficient is located at the first relative coefficient section, then motion state is to walk line state；When relative coefficient is located at the second correlation
When property coefficient section, then motion state is riding condition.
The computer readable storage medium may include: USB flash disk, mobile hard disk, read-only memory (Read-Only
Memory, ROM), random access memory (Random Access Memory, RAM), magnetic or disk etc. is various to deposit
Store up the medium of program code.
Each embodiment is described in a progressive manner in specification, the highlights of each of the examples are with other realities
The difference of example is applied, the same or similar parts in each embodiment may refer to each other.For device disclosed in embodiment
Speech, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place is referring to method part illustration
?.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software generally describes each exemplary composition and step according to function in the above description.These
Function is implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Profession
Technical staff can use different methods to achieve the described function each specific application, but this realization is not answered
Think beyond the scope of this invention.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
Above to a kind of recognition methods of motion state provided by the present invention, device, terminal and computer-readable storage
Medium is described in detail.It is used herein that a specific example illustrates the principle and implementation of the invention, with
The explanation of upper embodiment is merely used to help understand method and its core concept of the invention.It should be pointed out that being led for this technology
For the those of ordinary skill in domain, without departing from the principle of the present invention, can also to the present invention carry out it is several improvement and
Modification, these improvements and modifications also fall within the scope of protection of the claims of the present invention.
Claims (11)
1. a kind of recognition methods of motion state characterized by comprising
Obtain motion sensor acquisition data, and when according to the data determine motion state be walk or fringe of riding
When, relative coefficient is calculated using the expectation of the data and preset delay time；
When the relative coefficient is located at the first relative coefficient section, then motion state is to walk line state；
When the relative coefficient is located at the second relative coefficient section, then motion state is riding condition.
2. the recognition methods of motion state according to claim 1, which is characterized in that in the phase using the data
Prestige and preset delay time calculate before relative coefficient, further includes:
First sample entropy is calculated using the data；
When the first sample entropy is located at first sample entropy section, then motion state is to walk line state；
When the first sample entropy is located at the second Sample Entropy section, then motion state is riding condition；
When the first sample entropy is located at third Sample Entropy section, then executes the expectation using the data and preset
Delay time calculate relative coefficient the step of.
3. the recognition methods of motion state according to claim 2, which is characterized in that calculated described using the data
Before first sample entropy, further includes:
The first kurtosis value is calculated using the data；
When first kurtosis value is located at the first kurtosis value section, then motion state is to walk line state；
When first kurtosis value is located at the second kurtosis value section, then motion state is riding condition；
When first kurtosis value is located at third kurtosis value section, then executes the utilization data and calculate first sample entropy
The step of.
4. the recognition methods of motion state according to claim 3, which is characterized in that calculated described using the data
Before first kurtosis value, further includes:
Stationarity parameter is calculated using the data；
When the stationarity parameter is located at the first stationarity parameter section, then executes the utilization data and calculate first peak
The step of angle value；
When the stationarity parameter is located at the second stationarity parameter section, the second Sample Entropy is calculated using the data；Work as institute
When stating the second Sample Entropy greater than sample entropy threshold, then motion state is riding condition；When second Sample Entropy is less than the sample
When this entropy threshold, then motion state is to walk line state.
5. the recognition methods of motion state according to claim 4, which is characterized in that described calculated using the data is put down
Stability parameter, comprising:
The resultant acceleration for obtaining the corresponding 3-axis acceleration of preset window, determines the minimum resultant acceleration of the preset window
Min1；
Utilize formulaCalculate total resultant acceleration average value Average1；
Determine the minimum resultant acceleration Min2q that wicket is respectively slided in the preset window；
Utilize formulaCalculate the resultant acceleration average value of each sliding wicket
Average2q；
From the resultant acceleration average value Average2q of total resultant acceleration average value Average1 and each sliding wicket
Middle determining maximum value Max and minimum M in；
Using the ratio Max/Min of Max and Min as stationarity parameter AS；
Wherein, n is the corresponding resultant acceleration number of preset window, and X (i) is i-th of resultant acceleration, and Min2q is in preset window
The minimum resultant acceleration of q-th of sliding wicket, Xq (m) are that m-th of conjunction of q-th of sliding wicket in preset window accelerates
Degree, l are the resultant acceleration number for sliding wicket.
6. the recognition methods of motion state according to claim 1, which is characterized in that moved when being determined according to the data
State is when running or ride fringe, further includes:
The second kurtosis value is calculated using the data；
When second kurtosis value is less than the first kurtosis threshold value, then motion state is running state；
When second kurtosis value is greater than the first kurtosis threshold value, then motion state is riding condition.
7. the recognition methods of motion state according to claim 1, which is characterized in that described determined according to the data is transported
Dynamic state is to walk or fringe of riding, comprising:
The resultant acceleration of the 3-axis acceleration obtained is calculated, and calculates the standard deviation of the resultant acceleration；
When the standard deviation is located at the first standard deviation section, then motion state be walk or fringe of riding.
8. the recognition methods of motion state according to claim 1-7, which is characterized in that described to utilize the number
According to expectation and preset delay time calculate relative coefficient, comprising:
Utilize formula A (τ)=E [(Xt-μ)(Xt+τ- μ)] calculate separately preset delay time be 0 relevance parameter A (0) with
And preset delay time is the relevance parameter A (n) of n；
It regard ratio A (0)/A (n) of A (0) and A (n) as relative coefficient A；
Wherein, τ is delay time, XtFor the resultant acceleration obtained in preset time period t, Xt+τFor in preset time period t delay τ
Obtained resultant acceleration, μ XtExpectation, E be seek expectation computing.
9. a kind of identification device of motion state characterized by comprising
Relative coefficient computing module is moved for obtaining the data of motion sensor acquisition, and when being determined according to the data
State is to calculate correlation system using the expectation of the data and preset delay time when walking or ride fringe
Number；
First categorization module, for when the relative coefficient is located at the first relative coefficient section, then motion state to be to walk
Line state；When the relative coefficient is located at the second relative coefficient section, then motion state is riding condition.
10. a kind of terminal characterized by comprising
Motion sensor, for acquiring data；
Memory, for storing computer program；
Processor realizes the identification of the motion state as described in any one of claim 1 to 8 when for executing the computer program
The step of method.
11. a kind of computer readable storage medium, which is characterized in that be stored with computer on the computer readable storage medium
Program realizes the identification side of the motion state as described in any one of claim 1 to 8 when the computer program is executed by processor
The step of method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811455450.4A CN109582713B (en) | 2018-11-30 | 2018-11-30 | Motion state identification method, motion state identification device and terminal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811455450.4A CN109582713B (en) | 2018-11-30 | 2018-11-30 | Motion state identification method, motion state identification device and terminal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109582713A true CN109582713A (en) | 2019-04-05 |
| CN109582713B CN109582713B (en) | 2023-05-19 |
Family
ID=65925874
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811455450.4A Active CN109582713B (en) | 2018-11-30 | 2018-11-30 | Motion state identification method, motion state identification device and terminal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109582713B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110109153A (en) * | 2019-05-10 | 2019-08-09 | 北京百度网讯科技有限公司 | Navigation processing method, navigation terminal, equipment and storage medium |
| CN113074724A (en) * | 2021-03-26 | 2021-07-06 | 歌尔股份有限公司 | Motion time calculation method, device, equipment and computer readable storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130041619A1 (en) * | 2011-08-12 | 2013-02-14 | Fujitsu Limited | Electronic device and motion state judgment method |
| CN103340634A (en) * | 2013-06-17 | 2013-10-09 | 无锡市中安捷联科技有限公司 | Method for detecting human motion state based on accelerated speed changes |
| JP2014033739A (en) * | 2012-08-07 | 2014-02-24 | Nippon Telegr & Teleph Corp <Ntt> | Gait measuring apparatus, method and program |
| CN104713568A (en) * | 2015-03-31 | 2015-06-17 | 上海帝仪科技有限公司 | Gait recognition method and corresponding pedometer |
| CN106237604A (en) * | 2016-08-31 | 2016-12-21 | 歌尔股份有限公司 | Wearable device and the method utilizing its monitoring kinestate |
| CN107145834A (en) * | 2017-04-12 | 2017-09-08 | 浙江工业大学 | A kind of adaptive behavior recognition methods based on physical attribute |
| CN107144291A (en) * | 2017-05-23 | 2017-09-08 | 维沃移动通信有限公司 | A kind of data processing method and mobile terminal |
| CN107530004A (en) * | 2015-02-20 | 2018-01-02 | Mc10股份有限公司 | The automatic detection and construction of wearable device based on personal situation, position and/or orientation |
| CN108814618A (en) * | 2018-04-27 | 2018-11-16 | 歌尔科技有限公司 | A kind of recognition methods of motion state, device and terminal device |
-
2018
- 2018-11-30 CN CN201811455450.4A patent/CN109582713B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130041619A1 (en) * | 2011-08-12 | 2013-02-14 | Fujitsu Limited | Electronic device and motion state judgment method |
| JP2014033739A (en) * | 2012-08-07 | 2014-02-24 | Nippon Telegr & Teleph Corp <Ntt> | Gait measuring apparatus, method and program |
| CN103340634A (en) * | 2013-06-17 | 2013-10-09 | 无锡市中安捷联科技有限公司 | Method for detecting human motion state based on accelerated speed changes |
| CN107530004A (en) * | 2015-02-20 | 2018-01-02 | Mc10股份有限公司 | The automatic detection and construction of wearable device based on personal situation, position and/or orientation |
| CN104713568A (en) * | 2015-03-31 | 2015-06-17 | 上海帝仪科技有限公司 | Gait recognition method and corresponding pedometer |
| CN106237604A (en) * | 2016-08-31 | 2016-12-21 | 歌尔股份有限公司 | Wearable device and the method utilizing its monitoring kinestate |
| CN107145834A (en) * | 2017-04-12 | 2017-09-08 | 浙江工业大学 | A kind of adaptive behavior recognition methods based on physical attribute |
| CN107144291A (en) * | 2017-05-23 | 2017-09-08 | 维沃移动通信有限公司 | A kind of data processing method and mobile terminal |
| CN108814618A (en) * | 2018-04-27 | 2018-11-16 | 歌尔科技有限公司 | A kind of recognition methods of motion state, device and terminal device |
Non-Patent Citations (2)
| Title |
|---|
| EKACHAI THAMMASAT: "The statistical recognition of walking, jogging, and running using smartphone accelerometers", 《THE 6TH 2013 BIOMEDICAL ENGINEERING INTERNATIONAL CONFERENCE》 * |
| 基于加速度传感器的人体运动状态识别研究: "彭际群", 《中国优秀硕士学位论文全文数据库 (信息科技辑)》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110109153A (en) * | 2019-05-10 | 2019-08-09 | 北京百度网讯科技有限公司 | Navigation processing method, navigation terminal, equipment and storage medium |
| CN110109153B (en) * | 2019-05-10 | 2021-03-23 | 北京百度网讯科技有限公司 | Navigation processing method, navigation terminal, device and storage medium |
| CN113074724A (en) * | 2021-03-26 | 2021-07-06 | 歌尔股份有限公司 | Motion time calculation method, device, equipment and computer readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109582713B (en) | 2023-05-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106289309B (en) | Step-recording method and device based on 3-axis acceleration sensor | |
| CN105912142B (en) | A kind of note step and Activity recognition method based on acceleration sensor | |
| Kose et al. | Online human activity recognition on smart phones | |
| CN104197952B (en) | A kind of user's walking step-recording method, device and mobile terminal | |
| CN103892840B (en) | The feature extracting method of a kind of Intelligent worn device and human motion | |
| CN104089624B (en) | Step-recording method and terminal device | |
| CN106017502B (en) | A kind of step-recording method and electronic equipment | |
| Susi et al. | Accelerometer signal features and classification algorithms for positioning applications | |
| CN106123911A (en) | A kind of based on acceleration sensor with the step recording method of angular-rate sensor | |
| CN106510658B (en) | A kind of human body emotion judgment method based on bracelet | |
| CN109558841B (en) | Motion state identification method, motion state identification device and terminal | |
| Shao et al. | DePedo: Anti periodic negative-step movement pedometer with deep convolutional neural networks | |
| CN108814618A (en) | A kind of recognition methods of motion state, device and terminal device | |
| CN109582713A (en) | A kind of recognition methods of motion state, device and terminal | |
| Biancat et al. | Review of transportation mode detection techniques | |
| CN105589977B (en) | A kind of times of exercise monitoring method and device | |
| Aiolli et al. | ClimbTheWorld: Real-time stairstep counting to increase physical activity | |
| CN103344249A (en) | Step counting method based on short-time spectral analysis of inertial sensor | |
| Zhao et al. | Healthy: A diary system based on activity recognition using smartphone | |
| CN106570479B (en) | A kind of pet motions recognition methods of Embedded platform | |
| CN109033995A (en) | Identify the method, apparatus and intelligence wearable device of user behavior | |
| CN107392106B (en) | Human activity endpoint detection method based on double thresholds | |
| Chen | An LSTM recurrent network for step counting | |
| CN105760646A (en) | Method and device for activity classification | |
| CN106994238A (en) | Data processing method and the electronic equipment for performing the data processing method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |