TWI605391B - Inductive multidimensional intelligent identification device and method thereof - Google Patents

Description

Inductive multi-dimensional intelligent identity recognition device and method thereof

The present invention relates to a device for identifying an identity and a method thereof, and more particularly to an identity recognition device and method for sensing a spatial magnetic field change through an inductor.

In the general identity identification system, the three points based on multi-factor identity identification are used as the basis for identification, including: (1) some of their knowledge (such as passwords); (2) users own Something they have, such as a key or proximity card; (3) something they are, such as a fingerprint.

However, the above methods have obvious problems, such as the problem of leakage of the password, the key may be copied or lost, the fingerprint identification device is expensive, etc., so the development of the security management device is bound to have a new design requirement.

There are many examples of improvements to the above-mentioned shortcomings, for example, Patent Patent Publication No. 201543252, "Remote Control Method with Identity Verification Mechanism and Wearable Device Performing the Method" Patent Case, which is proposed to use a wearable device for detection. The fingerprint replaces the traditional identity authentication method to improve the user's fraudulent use and transmit the user identity identification result through a single wireless technology. However, this case also has corresponding shortcomings. For example, for the problem of ensuring that the fingerprint identification is not damaged, the power consumption of the wearable device also needs to be taken into consideration.

In addition, as in the patent case of the Republic of China Patent Publication No. 201602826 "Personalized Control System for Applied Physiological Characteristics and Its Operation Method", the case is proposed to be obtained by using a detection device. The user's physiological signal replaces the traditional identity identification method to improve the shortcomings of the information hacker's identity and to transmit the user's identity identification result through a single wireless technology. However, there are also shortcomings in which physiological signals are often difficult to extract. When detecting physiological signals, if there are uncertain factors such as signal interference around the device, it will have a considerable impact on the identification results. Furthermore, the user must be familiar with the use of the detection device to improve the quality of the physiological signal acquisition in order to obtain relatively good identification results.

It can be seen that there are still many shortcomings in the above-mentioned attempts to improve, which is not a good design, but needs to be improved.

In order to solve the problems disclosed above, the object of the present invention is to provide a multi-factor identification technology based on (1) user-defined decoding skill knowledge and (2) personal special exercise habits as the basis for identification. The identification technology solution with high recognition ability and reliability.

To achieve the above object, the present invention provides an inductive multi-dimensional intelligent identity recognition device comprising a complex array electromagnetic coil, a magnetic field strength collecting unit and an identification control unit. The multi-array electromagnetic coil is used for sensing the change of the multi-dimensional magnetic field strength caused when the external object moves along a specific trajectory; the magnetic field strength collecting unit is connected to the foregoing complex array electromagnetic coil for collecting the sensed by the electromagnetic coil The change of the magnetic field strength; and an identification control unit connected to the magnetic field strength collecting unit, the function of which is to calculate the characteristics of the specific trajectory by using the change of the magnetic field strength, and provide the identification result according to the feature.

In order to achieve the above object, the present invention also provides an inductive multi-dimensional intelligent identity identification method, which comprises the following steps: a subject draws a specific trajectory in a three-dimensional space as a basis for identifying a record; and a specific trajectory is obtained by inductive sensing. Magnetic field strength changes and magnetic Calculating the coordinates of the plurality of two-dimensional coordinate points included in the specific trajectory; calculating the eigenvalues of the plurality of two-dimensional coordinate points; and obtaining the eigenvectors by using the relationship between the eigenvalues and the time; and identifying the algorithm for the specific trajectory The plurality of feature vectors are compared with the identification model to identify the user identity.

In summary, the present invention is characterized in that it utilizes the difference in the habitual motion trajectory of each user as its identification basis; the identification control unit can instantly recognize the user's motion trajectory feature points; in addition, compared with the conventional capacitive type Sensing, the identification result will be affected by the user's physiological condition and the time-varying effect of the capacitor effect. The inductive detection system used in the present invention senses the magnetic field change as the identification basis in a non-contact manner, and has better reliability. Sex. In addition, since the identification operation can be completed using only a single microprocessor, it also has the advantage of low cost.

1‧‧‧Inductive multi-dimensional intelligent identity identification device

11‧‧‧Electromagnetic coil

12‧‧‧Magnetic strength collection unit

13‧‧‧ID control unit

20‧‧‧Two-dimensional plane

22‧‧‧Measured objects

31‧‧‧1 electromagnetic coil

32‧‧‧2nd electromagnetic coil

33‧‧‧3 electromagnetic coil

35‧‧‧Three-dimensional magnetic field strength change

S401~S405‧‧‧Steps

1 is a system block diagram of an inductive multi-dimensional intelligent identity recognition device of the present invention.

Figure 2 is a 3D cross-sectional view of the magnetic field detection intensity.

FIG. 3 is a schematic diagram of an inductor coil erection of the inductive multi-dimensional intelligent identity recognition device of the present invention.

4 is a flow chart of a method for inductive multi-dimensional intelligent identity identification method of the present invention.

The specific embodiments are described below to illustrate the embodiments of the invention, but are not intended to limit the scope of the invention.

Please refer to FIG. 1 , which is a system block diagram of an inductive multi-dimensional intelligent identity recognition device 1 according to a first embodiment of the present invention, comprising: three sets of electromagnetic coils 11 , a magnetic field strength collecting unit 12 and an identification control unit 13 . The aforementioned magnetic field strength collecting unit 12 is connected to three sets of electromagnetic coils 11 And the identification control unit 13. The foregoing three sets of electromagnetic coils 11 can be a hard circuit coil or a flexible coil of a PCB circuit board. The magnetic field strength collecting unit 12 includes an amplifier and a communication interface connected thereto, and the communication interface can be wired or wirelessly connected and controlled. The unit 13 communicates, and the identification control unit 13 is an electronic device having a computing capability, a microprocessor, or a control chip on which the operating system is mounted.

The foregoing three sets of electromagnetic coils 11 are used to sense changes in the three-dimensional magnetic field strength caused when an external object moves along a specific trajectory, and are placed in a specific manner to form a sensing space (for example, a triangle). When the user wants to perform the identity recognition work, the predetermined trajectory (for example, a zigzag shape) is drawn in the sensing space formed by the three sets of electromagnetic coils 11 by using a specified object, and the foregoing three sets of electromagnetic coils are used. The 11 are located at different locations in the space, and thus different magnetic field changes are sensed, each generating a different amount of induced current and transmitted to the magnetic field strength collecting unit 12.

Wherein, the foregoing specified object for depicting the trajectory may be a metal object that can be hand-held or worn or an article containing metal.

After receiving the induced current, the magnetic field strength collecting unit 12 amplifies the signal by using an amplifier to make the circuit easy to recognize, and transmits it to the identification control unit 13 by wired or wireless communication.

The identification control unit 13 calculates the characteristics of the specific trajectory by using the induced current generated by the change in the magnetic field strength, and provides the identity recognition result according to the feature. The calculation process in the identification control unit 13 will be described in detail below: Since the magnetic field strength sensed by the electromagnetic coil will be a non-linear behavior, for the nonlinear magnetic field strength, the problem of linearization of the magnetic field must be prioritized. In the present invention, an Ordinary Least Squares estimator (OLS estimator) is used as a linearization method, and the linearized magnetic field strength represents the relative distance between the measured object and the center point of the coil magnetic field, as shown in equation (1). -(3): ( xx _a ) ² +( yy _a ) ² = d _a (1)

( xx _b ) ² +( yy _b ) ² = d _b (2)

( xx _c ) ² +( yy _c ) ² = d _c (3)

Where x is the x-axis of the object under test in a two-dimensional plane coordinates of the coordinate point, y is the y-axis by the object to be measured in a two-dimensional plane coordinates of the coordinate _{point, x a, x b, x} c , respectively a first, second and third The x-axis coordinate points of the set inductor coils, y _a , y _b , and y _c are the y-axis coordinate points of the first, second, and third sets of inductor coil mounting positions , respectively, d _a , d _b , and d _c are the first The distance between the second and third sets of inductive coils and the object to be measured.

Then, the three sets of electromagnetic field strengths are similarly calculated using the OLS estimator to calculate the two-dimensional plane coordinate system, as shown in equation (4):

Transform the coordinate points into eigenvalues in vector form:

The trajectory drawn by the user in the sensing space will be recorded in time and become a record of a plurality of two-dimensional coordinate points corresponding to the individual time. The plurality of two-dimensional coordinate points will be converted into Multiple feature values.

Please refer to FIG. 2 , which is a 3D cross-sectional view of the magnetic field detection intensity, wherein the two-dimensional plane 20 is an XY plane, and the plurality of two-dimensional coordinate points calculated by the identification control unit 13 are also on the plane. As shown in the figure, the object to be tested 22 is an elongated rod (other shapes or forms). When the user holds the object 22 to be measured in the sensing space, the two-dimensional plane 20 is passed. The calculation result of the coordinates of the object to be measured on the two-dimensional plane 20 by the identification control unit 13 varies depending on the speed, the depth change, and the shape of the object to be measured when the object 22 is moved on the plane. Then use the Dubinin Algorithms to process the eigenvalues of the coordinates obtained at each time point, as shown in equation (6):

Through the Durbin Algorithms, the feature vectors a ₁ ~ a _n taking into account the time factor can be obtained, and the feature vector is input into the pre-trained recognition algorithm model for comparison, and the function of identity recognition can be achieved.

In addition, the foregoing pre-trained recognition algorithm model may be a conventional technique such as a neural network and a hidden Markov model, and the training method is: a user draws a specific in the sensing space of the device. The trajectory (for example, drawing a zigzag or other more complex trajectory), the recognition control unit 13 will obtain a set of eigenvectors, repeating this step a plurality of times, so that the recognition algorithm model can recognize the characteristic of the user to draw a specific trajectory. The training is completed.

Please refer to FIG. 3 , which is an inductive coil erection diagram of the inductive multi-dimensional intelligent identity recognition device 1 , in which three electromagnetic coils of the first electromagnetic coil 31 , the second electromagnetic coil 32 , and the third electromagnetic coil 33 are formed around A sensing space having a triangular or approximately triangular shape to sense a change in the three-dimensional magnetic field strength caused by the measured object moving within the sensing space. However, if the object to be measured moves outside the sensing space, the electromagnetic coil group also senses the magnetic field change. Chemical. In addition, the inductor coil group can also be embedded in various safety devices of different sizes according to the needs of the user.

In an embodiment, the device adopts a passive sensing method, and does not emit any signal when sensing the object to be measured, and directly senses the change of the surrounding magnetic field to obtain the relative position of the measured object.

In another embodiment, the magnetic field strength collecting unit 13 of the device generates an alternating current signal into the electromagnetic coil group to generate electromagnetic waves, which are emitted onto the object to be measured, and the electromagnetic coil group receives electromagnetic waves reflected from the object to be measured. , calculate the position of the measured object.

Please refer to FIG. 4 , which is a flowchart of a method for inductive multi-dimensional intelligent identity identification method according to a second embodiment of the present invention, which includes the following steps:

S401: The subject draws a specific trajectory (for example, a zigzag shape) in a three-dimensional space as a basis for identifying the record. In order to make the magnetic field of the object to be measured change more, an object with a conductive material may be used to draw a specific trajectory.

S402: Inductively sensing a change in a magnetic field strength caused by drawing a specific trajectory, and calculating a coordinate of the plurality of two-dimensional coordinate points included in the specific trajectory. The method for calculating the above coordinates is the least square estimation method (OLS estimator).

S403: Calculate individual characteristic values of the plurality of two-dimensional coordinate points. The method for calculating the feature value is as shown in the formula (5), and the coordinate points of the XY axes of the coordinate points are individually squared, and then added and the root number is obtained.

S404: Obtain a feature vector by using a relationship between the feature value and the time. Among them, the feature vector is obtained by using the Durbin algorithm to calculate the relationship between the feature value and time (as shown in equation (6)). Since a trajectory is composed of a plurality of coordinate points, after this calculation, a plurality of eigenvectors (i.e., a ₁ ~ a _n in the equation (6)) for describing the trajectory are obtained.

S405: The identification algorithm compares the plurality of feature vectors of the specific trajectory with the identification model to identify the user identity. The identification model is trained by using a neural network as the identification architecture. The training method is to let the user repeat the steps of S401~S404, so that the identification model can distinguish the feature vectors when different users draw the specified trajectory. difference.

In the following, Table 1 is taken as an actual application example of the inductive multi-dimensional intelligent identity recognition device 1 and the method thereof.

In Table 1, the trajectory patterns set by the users 1, 2, and 3 are respectively M, Z, and Z, which represent the trajectories that the user has to draw in the sensing space when using the device of the present invention. Among them, the user 1 and the users 2, 3 set different trajectories, and the users 2, 3 set the same trajectory, and the user 4 is an external user without setting.

Like the general identity recognition system, when the user draws a custom trajectory graphic in the sensing space, the identification result can be obtained by the identity recognition, so that the user 1, 2, 3 draws their own preset trajectory graphic in the device. Will be recognized by identity, while User 4 will not pass.

However, when a user knows the trajectory pattern set by another person, for example, when the user 2, 3, 4 draws the M-word trajectory preset by the user 1 in the sensing space, the time for each stroke is drawn by different people. The length and interval will be different, or the angle of each stroke will be different. The resulting feature vector will also be different, so it will not be recognized by identity after comparison.

In the same situation, even if the user sets the same track pattern, such as user 2 and user 3, the same track pattern is set, but for the foregoing reasons, the device still It can distinguish the feature vector difference between the user 2 and the user 3 to draw the trajectory, and correctly identify the identity.

The detailed description of the preferred embodiments of the present invention is intended to be limited to the scope of the invention, and is not intended to limit the scope of the invention. The patent scope of this case.

1‧‧‧Inductive multi-dimensional intelligent identity identification device

11‧‧‧Electromagnetic coil

12‧‧‧Magnetic strength collection unit

13‧‧‧ID control unit

Claims (10)

An inductive multi-dimensional intelligent identity recognition device comprises: a complex array electromagnetic coil, which is used for sensing a change of a multi-dimensional magnetic field intensity caused when an external object moves along a specific trajectory; a magnetic field strength collecting unit is connected to the electromagnetic coil, Collecting the change of the magnetic field strength sensed by the electromagnetic coils; and identifying an control unit connected to the magnetic field strength collecting unit, wherein the identification control unit calculates the characteristic of the specific trajectory by using the magnetic field intensity change, and provides according to the characteristic The identity identification result; wherein the identification control unit calculates coordinates of a plurality of two-dimensional coordinate points included in the specific trajectory, respective eigenvalues of the two-dimensional coordinate points, and obtains individual characteristics by using the relationship between the eigenvalues and time The vector is compared to the eigenvectors and the identification model to identify the subject's identity, and wherein the identification control unit uses a Doberman algorithm to calculate the relationship between the eigenvalue and time to provide the eigenvector. The inductive multi-dimensional intelligent identity recognition device of claim 1, wherein the electromagnetic coils surround a sensing space formed in a triangle to sense the multi-dimensional magnetic field intensity variation in the sensing space. The inductive multi-dimensional intelligent identification device according to claim 1, wherein the magnetic field strength collecting unit generates an alternating current signal to the electromagnetic coils, so that the electromagnetic coils generate electromagnetic waves to sense the distance of the object to be tested. The inductive multi-dimensional intelligent identity recognition device according to claim 1, wherein the identification control unit calculates the coordinates of the two-dimensional coordinate points by a least square estimation method. The inductive multi-dimensional intelligent identity recognition device according to claim 1, wherein the identification control unit After the two-axis coordinate value of the two-dimensional coordinate point is squared, the root number is added and the root value is calculated to calculate the characteristic value. An inductive multi-dimensional intelligent identity identification method includes the following steps: a subject draws a specific trajectory in a three-dimensional space as a basis for identifying a record; and inductively sensing a magnetic field intensity change caused by the specific trajectory, and using the magnetic field Calculating the coordinates of the plurality of two-dimensional coordinate points included in the specific trajectory; calculating individual eigenvalues of the plurality of two-dimensional coordinate points; calculating a eigenvector by using a Dubin algorithm to calculate the relationship between the eigenvalue and time; The identification algorithm compares the plurality of feature vectors of the specific trajectory with the identification model to identify the user identity. The inductive multi-dimensional intelligent identity identification method according to claim 6, wherein the method of drawing a specific trajectory comprises: using an item having a conductive material to move according to a specific trajectory customized. The inductive multi-dimensional intelligent identity identification method according to claim 6, wherein the method for calculating the two-dimensional coordinate point is a least square estimation method. The inductive multi-dimensional intelligent identity identification method according to claim 6, wherein the method for calculating the characteristic value is that the two-axis coordinates of the coordinate point are squared individually, and then the root number is added. The inductive multi-dimensional intelligent identity identification method according to claim 6, wherein the identification model is trained using a neural network as the identification architecture.