CN106331362A - Mobile phone theft prevention method based on built-in acceleration sensor - Google Patents

Abstract

The invention discloses a mobile phone theft prevention method based on a built-in acceleration sensor, which belongs to the technical field of safety. Firstly, the intelligent mobile phone of a user is trained, the user puts the mobile phone in a pocket or other place, after a period of time of movement, the mobile phone acceleration sensor acquires personal feature information, and the mobile phone can recognize the identity of the owner through the feature information; and if the mobile phone is stolen, an illegal person carries the mobile phone to move, the mobile phone can recognize the mobile phone is stolen, an alarm is given instantly, the user is reminded that the mobile phone is stolen, and the mobile phone sends the geographical position of the stolen mobile phone to a bound mobile phone. Thus, when the mobile phone is stolen, an alarm can be given for positioning, and a great role is played in subsequent mobile phone tracking.

Description

A kind of anti-theft method of mobile phone based on built-in acceleration sensor

Technical field

The invention belongs to technical field of burglary prevention, be specifically related to a kind of antitheft mobile phone side based on built-in acceleration sensor Method.

Background technology

Along with the development of mobile Internet, mobile phone has become as that people are daily just live in a requisite part, with Time hands also store substantial amounts of individual privacy data, such as note, photo etc..Along with the population in city increases, many public arenas Crowded to capacity, hand-set from stolen also happens occasionally, and once mobile phone private data leak, and brings sternly to the Working Life of client Ghost image rings.

For this social reality, the various electronic products of all kinds of antitheft mobile phones and theft preventing method arise at the historic moment, wherein A part is the portable security alarm device being associated with mobile phone, although this anti-theft device has instantaneity, but not very convenient, Because when mobile phone is away from the certain scope of device, and device will be reported to the police.Device is once omitted, and does not just have antitheft effect, and And buy the expense that device needs are extra.Another part is that mobile phone self arranges password, and once code error will be reported to the police, to association Mobile phone sends positional information, although can recover mobile phone, but be a lack of ageing, when mobile phone is recovered, information may be let out already Dew.

Along with the progress in epoch, the development of science and technology, smart mobile phone increasingly " intelligent ".Interior of mobile phone of today is self-contained Various sensors, such as acceleration transducer, gyroscope, Magnetic Sensor etc..It is engraved in during these sensors and records cellphone subscriber Individual behavior, everyone individual behavior is unique, can carry out the identification of personnel according to these data.

Summary of the invention

For this social phenomenon of hand-set from stolen, the present invention utilizes smart mobile phone that cellphone subscriber is carried out the identification of identity, Can identify immediately when the carrier of mobile phone changes, and send alarm.Additionally smart mobile phone itself can containing GPS module To position accurately, it is possible to send the position of stolen mobile phone in real time to association mobile phone.

Realize technical scheme as follows:

A kind of anti-theft method of mobile phone based on built-in acceleration sensor, is first trained the smart mobile phone of user, Mobile phone is placed on pocket, or other positions by user, and a period of time of moving allows mobile phone acceleration sensor obtain individual Characteristic information so that mobile phone can identify mastership by these characteristic informations；If hand-set from stolen is stolen, lawbreaker takes Band mobile phone is walked about, and mobile phone i.e. may recognize that hand-set from stolen, sends alarm immediately, reminds user mobile phone stolen, and mobile phone starts to tying up Determine mobile phone and send the geographical position of stolen mobile phone.

Further, the concrete steps of described method include:

(1) firstly the need of user's carrying mobile phone segment distance that moves as slowly walked, record what acceleration transducer recorded The data of user；

(2) user data utilizing Kalman filtering to obtain step (1) processes；Filtering relational expression is as follows:

X (k | k-1)=A*X (k-1 | k-1)+B*U (k)

P (k | k-1)=A*P (k-1 | k-1) * A'+Q

Kg (k)=P (k | k-1) * H'/(H*P (k | k-1) H'+R)

X (k | k-1)=X (k | k-1)+Kg (k) * (Z (k)-H*X (k | k-1))

P (k | k)=(I-Kg (k) * H) * P (k | k-1)

Wherein, X (k | k-1) is the result utilizing laststate to predict, X (k-1 | k-1) is the result that laststate is optimum, U (k) be the controlled quentity controlled variable controlled quentity controlled variable here of status praesens be 0；A, B are the parameter of Multi-model System.

P (k | k-1) is the covariance that X (k | k-1) is corresponding, and P (k-1 | k-1) is the covariance that X (k-1 | k-1) is corresponding, A ' Representing the transposed matrix of A, Q is the covariance of systematic procedure.

K_gK () is the gain of Kalman, R, for measuring variance, reacts the certainty of measurement of acceleration transducer；

Z (k) represents the measured value in k moment；

(3) utilizing temporal signatures to be identified, the parameter of use includes cycle maximum averageDegree of bias S, quartile Away from IQR, correlation coefficient C, concrete calculation expression is as follows.

The meansigma methods of cycle maximumExpression formula:

$\stackrel{\‾}{X}=\frac{\underset{1}{\overset{n}{\Σ}}{I}_{max}}{n}$

Wherein I_maxFor the peak value in single cycle, n is the number of peak value；

Degree of bias S is data skew direction and degree, its expression formula:

$S=\frac{1}{N}\underset{1}{\overset{N}{\Σ}}{\left(\frac{a_i-a}{\mathrm{\σ}}\right)}^3$

a_iThe data that sensor records, i=1,2,3 ..., N, N are data amount check；A is the meansigma methods of N number of data measured； σ is the standard deviation of data measured；

Interquartile-range IQR IQR expression formula:

$Q_j=b_{k_j}+(b_{k_j+1}-b_{k_j})r_j$

IQR=Q₃-Q₁

Data aj are descending is ordered as b_j, j=1,2 ..., N；The position of quartile is P_j=1+ (N-1) j/4, j (=1,2,3) are a point position number, k_jFor P_jInteger part, rj is fractional part；

Cross-correlation coefficient C expression formula:

C_xyz=COV (x, y, z)/(σ_xσ_yσ_z)

In formula: (x, y, z) be x to COV, y, the covariance of z-axis acceleration；σ_x,σ_y, σ_zFor x, y, the standard of z-axis acceleration Difference.

(4) use decision tree that eigenvalue carries out classification to judge；After decision tree T builds, estimate predictive value, identify user Identity；

(5) if hand-set from stolen is stolen, lawbreaker's carrying mobile phone is walked about, and mobile phone i.e. may recognize that the identity letter of lawbreaker Ceasing inconsistent with mobile phone mastership information, it is judged that hand-set from stolen, mobile phone starts to send the ground of stolen mobile phone to binding mobile phone Reason position, and send alarm, remind user mobile phone stolen.

Further, decision tree described in described step (4) is to be built by calculating entropy, and described entropy expression formula is:

$Entroy=-{\mathrm{\Σ}}_{i=1}^{n}p\left(i\right)*{\log }_2^{p\left(i\right)}$

Wherein p (i)=the i-th class number/total number.

Further, described step (4) being estimated, predictive value is by estimating that the normal distribution of an accuracy rate is interval, it is thus achieved that Data are reasonably classified behind interval by normal distribution, distinguish mobile phone carrier identity information.

Further, the confidence interval of described normal distribution solves as follows:

1) by acc standardization, i.e.Acc=X/N；N is just predicting for record data, X True record number；

2) select confidence level a, obtain statistic Z of accurate normal distribution corresponding for a/2 with 1-a/2_a/2,Z_1-a/2Take off Face, about the inequality of p, obtains the confidence interval of p:

$-Z_{a/2}\≤(acc-p)/\sqrt{p*(1-p)/N}\≤Z_{1-a/2}.$

Beneficial effects of the present invention:

1, the present invention is cheap need not extra sensor device, it is only necessary to the acceleration sensing that smart mobile phone is built-in Device just can complete above function.

2, the present invention combines the identification of identity mobile phone can be allowed to have the function of autonomous entity identification, antitheft relative to other Measure has certain instantaneity, and once mobile phone carrier is changed, and mobile phone sends alarm immediately, and positions, to follow-up Effect greatly is played in the tracking of mobile phone.

Accompanying drawing explanation

Fig. 1 is the original user data schema that the present invention extracts；

The user data that Fig. 2 is original be filtered after figure.

Detailed description of the invention

General user can be placed on mobile phone in pocket, the acceleration number of degrees of the user that mobile phone acceleration sensor collects According to, obtain the data of user, then logarithm carries out the extraction of feature, if average, maximum, variance etc. are as the characteristic point of user, By decision tree, characteristic is sorted out, to identify user identity.Then the GPS module utilizing smart mobile phone integrated is entered Row location.

First being trained the smart mobile phone of user, mobile phone can be placed on pocket, or other positions by user Put, and a period of time that needs to move can allow mobile phone acceleration sensor obtain enough personal characteristic information so that hands Machine can identify mastership by these characteristic informations, if stolen, lawbreaker's carrying mobile phone is walked about, mobile phone Identifying hand-set from stolen, send alarm immediately, remind user mobile phone stolen, mobile phone starts to send stolen mobile phone to binding mobile phone Geographical position.Specifically comprise the following steps that

(1) firstly the need of user's carrying mobile phone segment distance that moves as slowly walked, record what acceleration transducer recorded The data of user.

(2) obtaining data from previous step is primary data, and the interference of data is more, and there is substantial amounts of burr, such as Fig. 1 institute Showing, needing after filtering, filtered signal is rounder and more smooth, effect such as Fig. 2.

The filtering method that the present invention uses is Kalman filtering.Filtering relational expression is as follows:

X (k | k-1)=A*X (k-1 | k-1)+B*U (k) (1)

P (k | k-1)=A*P (k-1 | k-1) * A'+Q (2)

Kg (k)=P (k | k-1) * H'/(H*P (k | k-1) H'+R) (3)

X (k | k-1)=X (k | k-1)+Kg (k) * (Z (k)-H*X (k | k-1)) (4)

P (k | k)=(I-Kg (k) * H) * P (k | k-1) (5)

In formula (1), X (k | k-1) is the result utilizing laststate to predict, X (k-1 | k-1) is the knot that laststate is optimum Really, U (k) be the controlled quentity controlled variable controlled quentity controlled variable here of status praesens be 0；A and B is Multi-model System parameter.

In formula (2), P (k | k-1) is the covariance that X (k | k-1) is corresponding, and P (k-1 | k-1) is the association that X (k-1 | k-1) is corresponding Variance, A ' represents the transposed matrix of A, and Q is the covariance of systematic procedure.

In formula (3), K_g(k): for the gain of Kalman, R, for measuring variance, reacts the certainty of measurement of acceleration transducer.

In formula (4), Z (k) expression is the measured value in k moment.

Formula (1), (2) are the first two in the middle of 5 formula of Kalman filter, the namely prediction to system.Formula (3) (4) in (5), I, H are the matrix of 1.

(3) filtered figure has the obvious cycle, and the present invention runs in view of cell phone processor computing energy at interior of mobile phone Power is more weak, so mainly using temporal signatures to be identified, has used cycle maximum averageDegree of bias S, interquartile-range IQR IQR, correlation coefficient C, concrete calculation expression is as follows.

The meansigma methods of cycle maximumExpression formula:

$\stackrel{\‾}{X}=\frac{\underset{1}{\overset{n}{\Σ}}{I}_{max}}{n}$

Wherein I_maxFor the peak value in single cycle, n is the number of peak value.

Degree of bias S is data skew direction and degree, its expression formula:

$S=\frac{1}{N}\underset{1}{\overset{N}{\Σ}}{\left(\frac{a_i-a}{\mathrm{\σ}}\right)}^3$

a_iThe data that sensor records, i=1,2,3 ..., N, N are data amount check；A is the meansigma methods of N number of data measured； σ is the standard deviation of data measured.

Interquartile-range IQR IQR expression formula:

$Q_j=b_{k_j}+(b_{k_j+1}-b_{k_j})r_j$

IQR=Q₃-Q₁

Data a_jDescending it is ordered as b_j, j=1,2 ..., N.The position of quartile is P_j=1+ (N-1) j/4, j (=1,2,3) are a point position number, k_jFor P_jInteger part, rj is fractional part.

Cross-correlation coefficient C expression formula:

C_xyz=COV (x, y, z)/(σ_xσ_yσ_z)

In formula: (x, y, z) be x to COV, y, the covariance of z-axis acceleration；σ_x,σ_y, σ_zFor x, y, the standard of z-axis acceleration Difference.

(4) use decision tree that eigenvalue carries out classification to judge, identify identity.Decision tree builds according to " purity ", this Invent and built by calculating entropy, entropy expression formula:

$Entroy=-{\mathrm{\Σ}}_{i=1}^{n}p\left(i\right)*{\log }_2^{p\left(i\right)}$

Wherein p (i)=the i-th class number/total number.

After decision tree T builds, estimation predictive value, the accuracy rate estimated by sample, but the deviation of there will be likely some, So the present invention uses the method comparing science, it is the interval estimating an accuracy rate, it is thus achieved that just data are closed behind interval The classification of reason.

The confidence interval being distributed the most very much solves as follows:

1) by acc standardization, i.e.Acc=X/N.N is just predicting for record data, X True record number

2) select confidence level a, obtain statistic Z of accurate normal distribution corresponding for a/2 with 1-a/2_a/2,Z_1-a/2Take off Face, about the inequality of p, obtains the confidence interval of p.

$-Z_{a/2}\≤(acc-p)/\sqrt{p*1-p/N}\≤Z_{1-a/2}$

After obtaining interval, it is possible to reasonably classify data, the classification results of last synthetic data, differentiation is sold Machine carrier identity information.

(5) if hand-set from stolen is stolen, lawbreaker's carrying mobile phone is walked about, and mobile phone i.e. may recognize that the identity letter of lawbreaker Ceasing inconsistent with mobile phone mastership information, it is judged that hand-set from stolen, mobile phone starts to send the ground of stolen mobile phone to binding mobile phone Reason position, and send alarm, remind user mobile phone stolen.

The a series of detailed description of those listed above is only for the feasibility embodiment of the present invention specifically Bright, they also are not used to limit the scope of the invention, all equivalent implementations made without departing from skill of the present invention spirit Or change should be included within the scope of the present invention.

Claims (5)

1. an anti-theft method of mobile phone based on built-in acceleration sensor, it is characterised in that the first smart mobile phone to user Being trained, mobile phone is placed on pocket, or other positions by user, and a period of time of moving allows mobile phone acceleration sensor Obtain personal characteristic information so that mobile phone can identify mastership by these characteristic informations；If hand-set from stolen is stolen, disobey Legal person person's carrying mobile phone is walked about, and mobile phone i.e. may recognize that hand-set from stolen, sends alarm immediately, reminds user mobile phone stolen, mobile phone Start to send the geographical position of stolen mobile phone to binding mobile phone.

A kind of anti-theft method of mobile phone based on built-in acceleration sensor the most according to claim 1, it is characterised in that institute The concrete steps stating method include:

(1) firstly the need of user's carrying mobile phone segment distance that moves as slowly walked, the user that acceleration transducer records is recorded Data；

(2) user data utilizing Kalman filtering to obtain step (1) processes；Filtering relational expression is as follows:

X (k | k-1)=A*X (k-1 | k-1)+B*U (k)

P (k | k-1)=A*P (k-1 | k-1) * A'+Q

Kg (k)=P (k | k-1) * H'/(H*P (k | k-1) H'+R)

X (k | k-1)=X (k | k-1)+Kg (k) * (Z (k)-H*X (k | k-1))

P (k | k)=(I-Kg (k) * H) * P (k | k-1)

Wherein, X (k | k-1) is the result utilizing laststate to predict, X (k-1 | k-1) is the result that laststate is optimum, U (k) Controlled quentity controlled variable controlled quentity controlled variable here for status praesens is 0；A, B are the parameter of Multi-model System.

P (k | k-1) is the covariance that X (k | k-1) is corresponding, and P (k-1 | k-1) is the covariance that X (k-1 | k-1) is corresponding, and A ' represents A Transposed matrix, Q is the covariance of systematic procedure.

K_gK () is the gain of Kalman, R, for measuring variance, reacts the certainty of measurement of acceleration transducer；

Z (k) represents the measured value in k moment；

(3) utilizing temporal signatures to be identified, the parameter of use includes cycle maximum averageDegree of bias S, interquartile-range IQR IQR, correlation coefficient C, concrete calculation expression is as follows.

The meansigma methods of cycle maximumExpression formula:

$\stackrel{\‾}{X}=\frac{\underset{1}{\overset{n}{\Σ}}{I}_{max}}{n}$

Wherein I_maxFor the peak value in single cycle, n is the number of peak value；

Degree of bias S is data skew direction and degree, its expression formula:

$S=\frac{1}{N}\underset{1}{\overset{N}{\Σ}}{\left(\frac{a_i-a}{\mathrm{\σ}}\right)}^3$

a_iThe data that sensor records, i=1,2,3 ..., N, N are data amount check；A is the meansigma methods of N number of data measured；σ is for surveying Obtain the standard deviation of data；

Interquartile-range IQR IQR expression formula:

Q_j=b_kj+(b_kj+1-b_kj)r_j

IQR=Q₃-Q₁

Data a_jDescending it is ordered as b_j, j=1,2 ..., N；The position of quartile is P_j=1+ (N-1) j/4, j (=1, 2,3) for a point position number, k_jFor P_jInteger part, r_jFor fractional part；

Cross-correlation coefficient C expression formula:

C_xyz=COV (x, y, z)/(σ_xσ_yσ_z)

In formula: (x, y, z) be x to COV, y, the covariance of z-axis acceleration；σ_x,σ_y,σ_zFor x, y, the standard deviation of z-axis acceleration.

(4) use decision tree that eigenvalue carries out classification to judge；After decision tree T builds, estimate predictive value, identify user's body Part；

(5) if hand-set from stolen is stolen, lawbreaker's carrying mobile phone is walked about, mobile phone i.e. may recognize that the identity information of lawbreaker with Mobile phone mastership information is inconsistent, it is judged that hand-set from stolen, and mobile phone starts to send the geographical position of stolen mobile phone to binding mobile phone Put, and send alarm, remind user mobile phone stolen.

A kind of anti-theft method of mobile phone based on built-in acceleration sensor the most according to claim 2, it is characterised in that institute Stating decision tree described in step (4) is to be built by calculating entropy, and described entropy expression formula is:

$Entroy=-{\mathrm{\Σ}}_{i=1}^{n}p\left(i\right)*{\log }_2^{p\left(i\right)}$

Wherein p (i)=the i-th class number/total number.

A kind of anti-theft method of mobile phone based on built-in acceleration sensor the most according to claim 2, it is characterised in that institute State and step (4) being estimated, predictive value is by estimating that the normal distribution of an accuracy rate is interval, it is thus achieved that be right behind normal distribution interval Data are reasonably classified, and distinguish mobile phone carrier identity information.

A kind of anti-theft method of mobile phone based on built-in acceleration sensor the most according to claim 4, it is characterised in that institute The confidence interval stating normal distribution solves as follows:

1) by acc standardization, i.e.Acc=X/N；N is record data, and X prediction is correct Record number；

2) select confidence level a, obtain statistic Z of accurate normal distribution corresponding for a/2 with 1-a/2_a/2,Z_1-a/2Solve and close below In the inequality of p, obtain the confidence interval of p:

$-Z_{a/2}\≤(acc-p)/\sqrt{p*(1-p)/N}\≤Z_{1-a/2}.$