JP6516256B2 - Proximity test method and proximity test apparatus - Google Patents

Description

  The present invention relates to a technique for identifying proximity between terminals (proximity).

  In recent years, portable information terminals such as smartphones and tablets are widely spread, and the demand for services based on position information is increasing. One of the technologies for realizing a service based on location information is a proximity test (Proximity). Proximity test is a technology to identify whether users are close to each other. When performing a proximity test, it is necessary to protect the privacy of communicating users and to prevent attacks such as eavesdropping and impersonation of location information by others, location spoofing, and the like. Therefore, researches for performing a proximity test without exchanging actual position information obtained by GPS or the like are attracting attention.

  For example, using a correlation between the signal strength of a radio wave received by one user and the received signal strength of a radio wave received by another user around this user, a proxy that determines whether the users are close to each other Miti tests are known (see, for example, Non-Patent Document 1). This test method takes advantage of the high correlation of Received Signal Strength Indicator (RSSI) between terminals in close proximity. There is also known a method of using the amplitude and phase of a television signal or an FM signal to determine that the users are close to each other when the correlation between the amplitude and the phase is high (see, for example, Non-Patent Document 2). .

  In addition, the method of performing position estimation using a magnetic field strength vector of residual magnetism as a land mark (for example, refer to nonpatent literature 3) or indoor event is detected from the temporal change of the eigenvector of the received signal received by the array antenna. A method (see, for example, Non-Patent Document 4) is known.

A. Kalamandeen, A. Scannell, ED Lara, A. Sheth and A. Lamarca, "Ensemble: Cooperative Proximity-based Authentication," in Proc. ACM Mobile systems applications and services, pp. 271-284, San Francisco, June 2010 , S. Mathur, R. Miller, A. Varshavsky and W. Trappe, "ProxiMate: Proximity-based secure pairing using ambient wireless signals," in Proc. AMC MobySys, pp. 211-224, Washington, June 2011. P. Mirowski, H. Tin, Y. Saehoon and M. Macdonald, "Signal SLAM: Simultaneous localization and mapping with mixed WiFi, Bluetooth, LTE and magnetic signals," in Proc. IEEE IPIN, pp. 1-10, Montbeliard-Belfort , Oct. 2013 S. Ikeda, H. Tsuji and T. Ohtsuki, "Indoor Event Detection with Eigen vector Spanning Signal Subspace for Home or Office Security", IEEE Vehicle Tech. Conf., September, 2008

  In the method using the RSSI correlation, there is a problem that the spatial correlation of the RSSI hardly changes when the distance between the users is 2 meters or more apart, and the proximity can not be identified. The range in which the proximity test based on the RSSI correlation is effectively performed is limited to a narrow range. The method using the amplitude and phase of the television signal or the FM signal extends the identification range as compared to the method using the correlation of the RSSI, but the identification accuracy deteriorates in an environment where the surrounding environment changes.

  Therefore, it is an object of the present invention to provide a secure proximity test method that is resistant to changes in the surrounding propagation environment and that has high identification accuracy.

  In order to solve the above problems, in the present invention, actual position information can be obtained by using two or more of a plurality of feature quantities represented by spatial correlation of radio wave propagation and spatial correlation of magnetic field strength vector. Provide a method of secure proximity testing that does not exchange.

The proximity test method according to one aspect of the present invention is
Obtaining a measurement report including at least a received signal voltage and a residual magnetic field strength vector from the first terminal device and the second terminal device,
From the received signal voltage, a first evaluation function representing a change over time of an eigenvector of the signal and a second evaluation function representing a change over time of an eigenvalue of the signal are calculated;
The correlation of the received signal strength between the first terminal apparatus and the second terminal apparatus, the correlation of the first evaluation function, the correlation of the second evaluation function, and the magnetic field based on the measurement report and the calculation result At least two of the intensity vector correlations are used to determine whether the first terminal and the second terminal are located in the same space of a predetermined range.

One example is
Thresholds are individually set for the correlation of the received signal strength, the correlation of the first evaluation function, the correlation of the second evaluation function, and the correlation of the magnetic field strength vector, respectively.
The first terminal when two or more of the correlation of the received signal strength, the correlation of the first evaluation function, the correlation of the second evaluation function, and the correlation of the magnetic field strength vector exceed the corresponding threshold It is determined that the device and the second terminal are located in the same space.

  A secure proximity test with high identification accuracy and resistance to changes in the surrounding propagation environment is realized.

It is a figure which shows the correlation of the RSSI between users. It is a figure which shows the cumulative distribution function (CDF: Cumulative Distribution Function) of the correlation of RSSI when 1 antenna is used. It is a figure which shows CDF of the correlation of RSSI at the time of using the maximum value in all the combinations of an array antenna. It is a figure which shows CDF of the correlation of the eigenvector P (t) of a signal. It is a figure which shows CDF of the correlation of the eigen value Q (t) of a signal. It is a figure which shows the correlation of a magnetic field intensity vector. It is a figure which shows an identification rate when only the correlation of the RSSI of the largest value in all the combinations of an array antenna is used. It is a figure which shows an identification rate when only the correlation of eigenvector P (t) is used. It is a figure which shows the discrimination rate when only the correlation of the eigenvalue | characteristic value Q (t) is used. It is a figure which shows the discrimination rate when only the correlation of a magnetic field intensity vector is used. It is a figure which shows the combination of four feature-values, and the identification rate in each combination. FIG. 2 is a diagram showing an experimental environment I. It is a figure which shows experiment environment II. It is a figure which shows the identification accuracy in environment I. It is a figure which shows the identification accuracy in environment II. It is a figure which shows the comparison result of the identification precision at the time of using only the method of embodiment of Environment I, and RSSI correlation. It is a figure which shows the comparison result of the identification precision at the time of using only the method of embodiment of Environment II, and RSSI correlation. It is a figure which shows the application example of the method of embodiment. It is a flowchart which shows the process performed by a server. It is a schematic block diagram of a server.

  FIG. 1 is a diagram showing the correlation of RSSI when terminals are positioned close to each other. The terminal A and the terminal B receive radio waves from the same radio source such as an access point. The position of the terminal A is used as a reference position, and the position of the terminal B is changed to acquire the temporal change of the RSSI. As the distance between terminals approaches 2 m, 1 m and 10 cm, the correlation of RSSI increases. However, there is almost no change in the RSSI correlation when the distance between terminals is 2 m and 4 m, and it can be seen that the RSSI correlation can be used only within the range of 2 m.

In the embodiment, the fact that there is spatial correlation in radio wave propagation is used, and in addition to the received intensity (RSSI), the correlation of the eigenvalues of the signal and the correlation of the eigenvectors are used. Furthermore, the spatial correlation of the magnetic field strength vector is used to increase the accuracy of discrimination of proximity tests.
<Spatial correlation of radio wave propagation>
Introduce the time function P (t) of the eigenvectors of the signal and the time function Q (t) of the eigen value as feature quantities of radio wave propagation, and use P (t) and Q (t) as the evaluation function of the proximity test . P (t) represents the time-dependent change in the direction of arrival of the signal, and Q (t) represents the time-dependent change in the magnitude of the signal. The correlation of P (t) between terminals and the correlation of Q (t) are used as feature quantities of the proximity test. In the embodiment, the first terminal apparatus and the second terminal apparatus observe the propagation of radio waves using an array sensor, and the correlation of the RSSI which is the largest among all the combinations of antennas between the terminal apparatuses is a proximity test. Used as a feature amount. By using an array sensor having two or more antennas, spatial diversity effect is obtained to reduce the influence of fading.

  The evaluation function P (t) is expressed by equation (1).

ここで、ｖrefは参照時刻での信号の固有ベクトル、Ｈは複素共役転置、ｖob(t)は時刻ｔでの信号の固有ベクトルである。式（１）では、|Ｖref|＝|Ｖob(t)|＝１に正規化されている。Ｐ（ｔ）は参照時刻から電波伝搬方向がどのように変化したかを表わす。

Here, vref is the eigenvector of the signal at the reference time, H is the complex conjugate transpose, and vob (t) is the eigenvector of the signal at time t. In equation (1), normalization is performed to | Vref | = | Vob (t) | = 1. P (t) represents how the radio wave propagation direction has changed from the reference time.

  The evaluation function Q (t) is expressed by equation (2).

ここで、λrefは参照時刻での信号の固有値、λob(t)は時刻ｔでの信号の固有値である。Ｑ（ｔ）は、参照時刻から到来信号の大きさがどのように変化したかを表わす。

Here, λ ref is a characteristic value of the signal at the reference time, and λ ob (t) is a characteristic value of the signal at time t. Q (t) represents how the magnitude of the incoming signal has changed since the reference time.

  The use of the correlation of P (t) and the correlation of Q (t) means that the correlation of the change of the signal propagation state itself is used as a criterion of the proximity test. By using the correlation of P (t) and the correlation of Q (t) as the feature amount of the proximity test, it becomes possible to effectively perform the proximity test even in a place where the surrounding environment is likely to change.

  First, as a preliminary experiment, the correlation of RSSI, the correlation of P (t), and the correlation of Q (t) are individually evaluated. As the correlation of RSSI, the correlation when there is one antenna and the maximum correlation value in all combinations (4 × 4 = 16 ways) between array antennas having four antenna elements are evaluated.

  FIG. 2 is a CDF (Cumulative Distribution Function) of the correlation of RSSI when one antenna is present, and FIG. 3 is a CDF of the maximum RSSI correlation among combinations of 16 antennas. FIG. 4 shows the CDF of the correlation of P (t) and FIG. 5 shows the CDF of the correlation of Q (t).

  In the evaluation method, two rooms adjacent to each other across the corridor are prepared, and the positional relationship between the receiver 1 and the receiver 2 is changed to calculate the correlation. Each room is divided by concrete wall, and the size of the room is 5m x 7m and 7m x 7m. In each room, 2 to 3 people are doing normal work using a PC, etc., and create an environment in which people are present. (A) Distance of 2 meters away in the same room, (b) Distance of 4 meters away in the same room, (c) Distance of 2 meters away in separate rooms, (a) d) Vary the distance 4 meters apart, in separate rooms.

  Observe the signal for 90 seconds and calculate the correlation with a window width of 30 seconds. If the window width is short, it is difficult to see the difference in correlation due to the environment. Comparing FIG. 2 (one antenna) with FIG. 3 (four array antennas), in FIG. 2, the CDF in each positional relationship crosses each other, and the receiver 1 and the receiver 2 are necessarily close distances or in the same room The RSSI correlation does not always increase at a certain time. Therefore, the correlation of RSSI using one antenna is difficult to use as an index for determining whether the user is in proximity or in the same space.

  On the other hand, in FIG. 3, the correlation of the RSSI is high when the two receivers are in the same room, and the correlation of the RSSI is low in another room. In the case of the same room and in the case of another room, the degree of correlation is clearly distinguished. Further, in the same room, the correlation is higher when the distance is short, but the correlation is low regardless of the distance between terminals located in different rooms. By using the array antenna, the reliability of the RSSI correlation in the proximity test is improved.

The correlation between the correlation of the eigenvectors P (t) of FIG. 4 and the correlation of the eigenvalues Q (t) of FIG. 5 is also high when the two receivers are in the same room, and the correlation in the same room And the degree of correlation in the separate room are distinguished from one another. Thus, in addition to the correlation of the RSSI by the array antenna, the correlation of the eigenvector evaluation function P (t) of the signal and the correlation of the evaluation function Q (t) of the eigenvalues of the signal can be used as a criterion for the proximity test It was confirmed that there is.
<Spatial correlation of magnetic field strength vector>
In the embodiment, in addition to the correlation of RSSI at the array antenna, the correlation of P (t) and the correlation of Q (t), the spatial correlation of the magnetic intensity vector is also used for the proximity test.

  Environmental structures have unique residual magnetism different from geomagnetism, and magnetic field strength vectors differ depending on the location. So far, the spatial correlation of the magnetic intensity vector has not been considered, and in the present invention, the spatial correlation of the magnetic intensity vector of the remanence is observed to confirm that it can be used for the proximity test.

FIG. 6 shows the inner product of the direction of the magnetic field strength vector (two-dimensional) measured at each of the indoor and outdoor points and the reference point. The direction of the magnetic field strength vector at each point is indicated by an arrow. The inner product with the reference point, that is, the correlation of the magnetic field strength vector is shown as a normalized numerical value. It is understood from FIG. 6 that the correlation of the magnetic field strength vector is larger as the position is closer to the reference point, and the correlation of the magnetic field strength vector is higher when in the same room. The magnetic field strength vector of the remanent magnetism is unique depending on the place and has a small temporal variation. Therefore, the correlation of the magnetic field strength vector can be appropriately used as a determination criterion to determine whether two terminals are close or not in the same room.
<Method of proximity test>
As described above, it has been confirmed that the correlation of RSSI, correlation of P (t), correlation of Q (t), and correlation of magnetic field strength vector are all appropriate as feature amounts used for a proximity test.

  Next, how these four feature quantities are used in a proximity test will be described. In the embodiment, individual threshold values are set for the RSSI correlation, P (t) correlation, Q (t) correlation, and magnetic field strength vector correlation, and each correlation value obtained by measurement is compared with the threshold. To determine the proximity (proximity). The threshold value is determined in advance by performing a proximity test using each feature amount alone, and determining the value that the discrimination rate (identification accuracy) is the highest for each feature amount.

  FIG. 7 shows the identification rate when using only the RSSI correlation for four antennas. FIG. 8 shows the discrimination rate when only the correlation of P (t) is used, and FIG. 9 shows the discrimination rate when only the correlation of Q (t) is used. FIG. 10 shows the discrimination rate when only the correlation of the magnetic field strength vector is used. 7 to 10, the horizontal axis indicates the threshold value, and the vertical axis indicates the identification accuracy as the identification rate. The identification rate (identification accuracy) is indicated by the ratio of the number of times the two terminals could be correctly identified to be in the same room to the total number of times of identification.

Identification rate = (number of correct identifications) / (total number of identifications)
From FIG. 7, the threshold value of the RSSI correlation for four antennas is set to 0.35. From FIG. 8, the correlation threshold of P (t) is set to 0.2. From FIG. 9, the correlation threshold of Q (t) is set to 0.2. From FIG. 10, the threshold value of the correlation (inner product) of the normalized magnetic field strength vector is set to 0.96.

  FIG. 11 is a diagram showing how to use the threshold comparison results of the four feature quantities to increase the accuracy of discrimination of proximity tests. Pattern A is in the same room when two or more of the four feature quantities of RSSI correlation, P (t) correlation, Q (t) correlation, and magnetic field strength vector correlation exceed the corresponding threshold It is a method of determining The identification rate at this time is 0.893.

  Pattern B is a method of determining that the user is in the same room when two or more of the three feature quantities of the RSSI correlation, the P (t) correlation, and the magnetic field strength vector correlation exceed the corresponding threshold. The discrimination rate at this time is 0.848. Pattern C is a method of determining that the user is in the same room when one or more of the correlation of the RSSI and the correlation of the magnetic field strength vector exceeds the corresponding threshold. The identification rate at this time is 0.818. Pattern D is in the same room when three or more of four feature quantities of RSSI correlation, P (t) correlation, Q (t) correlation, and magnetic field strength vector correlation exceed the corresponding threshold It is a method of determining The identification rate at this time is 0.807.

In any of the patterns A to D, at least one of the three feature quantities representing spatial correlation of radio wave propagation and the correlation of the magnetic field strength vector are used in combination to achieve an identification rate of 80% or more. ing. However, it is understood that it is not necessary that the feature quantities simply exceed the threshold, and the proportion that can be correctly identified is largest when two or more of the four feature quantities exceed the threshold. In addition, even when two or more of the combination of the correlation of the magnetic field strength vector and the two feature quantities representing the spatial correlation of radio wave propagation exceed the threshold, the discrimination rate is close to 85%.
<Characteristics evaluation>
A method to determine that the user is in the same room when two or more of four feature quantities such as RSSI correlation, P (t) correlation, Q (t) correlation, and magnetic field strength vector correlation exceed a threshold (pattern Perform the actual characterization using A). Two different environments are set as the experiment environment. These two experimental environments are different from the measurement environments of the preliminary experiments described above.

  FIG. 12 shows the experimental environment I. Room 1 of the experimental environment I is surrounded by a concrete wall, and the door is closed. The first receiver (Rx1) is fixed, the position of the second receiver (Rx2) is 2 meters from the receiver Rx1 in the same room (position (1)), 4 meters in the same room Change the position (2), the distance of 2 meters outside (position (3)), and the distance of 4 meters outside (position (4)).

  FIG. 13 shows the experimental environment II. There is a partition between room 2 and room 3, but there is a gap at the top and bottom. The wall separating room 2 and the corridor is made of glass and the other walls are made of concrete. The door is closed. The first receiver (Rx1) is fixed in the room 2, and the position of the second receiver (Rx2) is 2 meters from the receiver Rx1 in the same room (position (1)), 4 meters in the same room The distance (position (2)), the distance of 2 meters outside (position (3)), and the distance of 8 meters outside (position (4)).

  In both experimental environments I and II, the transmitters (Tx) are separated by partitions to provide an NLOS (non-line-of-sight) environment. In Room 1 of Environment I and Room 2 of Environment II, two people are performing normal work using a PC or the like, and people are walking freely in the corridor. The observation is made for 90 seconds at each position of environment I and environment II, the correlation of RSSI is calculated with a window width of 30 seconds, and the window is slid every second. Table 1 shows the specifications of the receiver and the transmitter at this time.

図１４は、環境Ｉにおけるアンテナ４本のときのＲＳＳＩ相関による識別率を、アンテナ１本のときのＲＳＳＩ相関の識別率と比較して示す図である。図１５は、環境ＩＩにおけるアンテナ４本のときのＲＳＳＩ相関による識別率を、アンテナ１本のときのＲＳＳＩ相関の識別率を比較して示す図である。識別率は、ＲＳＳＩ相関値が閾値を超えたときに同室にいると判定する方法で、全識別回数に対する正しく識別できた回数の比（識別率＝（正しい識別回数）／（全識別回数））である。なお、アンテナが１本のときの閾値はアンテナが４本のときと同様の手順であらかじめ決めておく。

FIG. 14 is a diagram showing the identification rate by RSSI correlation in the case of four antennas in environment I in comparison with the identification rate of RSSI correlation in the case of one antenna. FIG. 15 is a diagram showing the identification rate by RSSI correlation when there are four antennas in environment II by comparing the identification rate of RSSI correlation when there is one antenna. The identification rate is the ratio of the number of correct identifications to the total number of identifications by the method to determine that the user is in the same room when the RSSI correlation value exceeds the threshold (identification rate = (number of correct identifications) / (total number of identifications)) It is. The threshold value for one antenna is determined in advance in the same procedure as for four antennas.

  In both the environment I and the environment II, the identification rate in the case of using four antennas is higher than in the case of using one antenna. It is considered that this is because changes in RSSI are more accurately captured due to the diversity effect. However, the correlation of the RSSI of the four antennas alone makes the discrimination rate slightly inferior when the distance is long even in the same room or when the distance is short even in another room. As a result, the discrimination rate that can be correctly determined to be in the same room regardless of the distance falls to 80% or less. Therefore, as described above, it is determined that two or more of the three feature quantities representing spatial correlation of radio wave propagation and the feature quantities representing the correlation of the magnetic field strength vector are in the same room when two or more of them exceed the threshold (see FIG. The identification rate is improved by the 11 patterns A).

  FIG. 16 is a diagram comparing discrimination rates in the case of using only the correlation of RSSI of four antennas in environment I and the case of using the method of the embodiment. FIG. 17 is a diagram comparing the correlation of only the RSSI of four antennas in environment II with the identification rate when the method of the embodiment is used. In environment I, a discrimination rate of 90% or more is achieved in any positional relationship. Even in an environment in which room 2 and room 3 are separated by a partition with a gap between them as in environment II, the identification rate that is judged to be in the same room at a distance between terminals reaches 85%. The discrimination rate in other positional relationships is 90% or more. By using the method of the embodiment, the identification rate is particularly improved when the distance is long in the same room or when the distance is short in another room.

  Since the correlation of the RSSI largely changes depending on the distance between users, the identification rate decreases when the distance is long even in the same room or when the distance is short in another room. On the other hand, since the correlation of P (t) and the correlation of Q (t) look at the correlation of the radio wave propagation state, the correlation becomes small when the users are close to each other but in different rooms. In addition, since the correlation between the signal arrival direction and the change in magnitude itself is observed, it is not easily affected by the change in the surrounding environment. The magnetic field strength vector of the residual magnetism is also less temporally variable, and the influence of being in the same room is larger than the influence of the distance. Therefore, it can be accurately determined whether the user is in the same room regardless of the distance. This is also supported by the fact that all of the patterns A to D in FIG. 11 realize an identification rate of 80% or more.

  FIG. 18 is a diagram showing an application example of the proximity test method of the embodiment. For example, service information such as coupons is distributed only to users in a certain store or users in a specific area. The use of a position detection function such as GPS can identify a user present in a specific building or in a predetermined area, but transmission and reception of position information is not desirable from the viewpoints of privacy protection and prevention of disguising position information. Therefore, the proximity between users is determined without transmission and reception of position information of the user terminal.

  In the room 100, a server 10 as a proximity test device and a user terminal 2-1 exist. Another user terminal 2-2 is moving into the room 100. Each of the user terminals 2-1 and 2-2 has an array antenna including a plurality of antenna elements. The server 10 transmits a measurement request message, for example, by broadcast (procedure (1)). The user terminal 2-1 having received the measurement request message measures the received signal voltage of the measurement request message received by the array antenna or any reference signal from the server 10, the RSSI, and the magnetic field strength vector of the residual magnetism (procedure (2 ), The measurement result A is transmitted to the server 10 in a measurement report message (procedure (3)). Similarly, the user terminal 2-2 receiving the measurement request message measures the received signal voltage of the measurement request message received by the array antenna or the received signal voltage of any signal from the server 10, the RSSI, and the magnetic field strength vector of the residual magnetism (procedure (2)) The measurement result A ′ is transmitted to the server 10 as a measurement report message (procedure (3)).

  The reception signal voltage mentioned here is a voltage value output from each antenna element of the array antenna, and represents a space electric field received by the antenna element. The RSSI is a power value received by each antenna element or its logarithm, and includes a time average. In this example, the user terminals 2-1 and 2-2 calculate the received signal voltage and the RSSI, but at least the received signal voltage and the magnetic field strength vector may be included in the measurement report message.

  The server 10 calculates P (t) and Q (t) in the user terminal 2-1 from the received signal voltage included in the measurement report message received from the user terminal 2-1. Similarly, P (t) and Q (t) in the user terminal 2-2 are calculated from the received signal voltage included in the measurement report message received from the user terminal 2-2 (procedure (4)).

  The server 10 further, based on the measurement result, the correlation of the RSSI between the user terminal 2-1 and the user terminal 2-2, the correlation of P (t), the correlation of Q (t), and the correlation of the magnetic field strength vector. It calculates and compares each feature quantity with the threshold value stored beforehand, and determines the proximity (proximity) of the user terminal 2-1 and the user terminal 2-2 (procedure (5)). Regarding the correlation of the RSSI, the maximum correlation among all the antenna element combinations is determined from the RSSI received from the user terminal 2-1 and the RSSI received from the user terminal 2-2. When the measurement report message from the user terminals 2-1 and 2-2 does not include the RSSI, the server 10 may calculate the RSSI from the received signal voltage received and calculate the correlation of the RSSI between the terminals. . The server 10 determines that the user terminal 2-1 and the user terminal 2-2 are inside the room 100 when two or more of the four feature amounts exceed the corresponding threshold.

  It is assumed that the server 10 knows that the user terminal 2-1 is in the room 100 by a conventional proximity test performed between the user terminal 2-1 and another user terminal not shown. . If it is determined by threshold determination that the user terminal 2-2 is in the same room 100 as the user terminal 2-1, for example, service information such as a coupon is transmitted from the server 10 to the address of the user terminal 2-2 .

  The server 10 may transmit a known pilot signal instead of the measurement request message. In that case, the user terminals 2-1 and 2-2 use the reception of the pilot signal as a trigger to measure the received signal voltage of the pilot signal, the RSSI, and the magnetic field strength vector of the residual magnetism, and include the measurement result in the measurement report message. It may be sent to the server 10.

  The measurement request message and the pilot signal may not necessarily be transmitted from the server 10. For example, a transmitter for transmitting a pilot signal is separately disposed in the room 100, and measurement results of received signal strength, RSSI, and magnetic field strength vector from user terminals 2-1 and 2-2 which have received the pilot signal from the transmitter May be received.

  FIG. 19 is a flowchart showing the operation of the server 10. The server 10 broadcasts a measurement request message, for example, at regular intervals (S11). This step is not essential, and a measurement request message or pilot signal may be transmitted from a separately installed transmitter.

  The server obtains the measurement results of the received signal voltage, the RSSI, and the residual magnetic field strength vector from the two or more user terminals from the measurement report message (S12). Calculate the eigenvector vob (t) and the eigenvalue λob (t) of the signal from the received signal voltage included in the measurement report message, and calculate the evaluation functions P (t) and Q (t) of each user terminal (S13) . The eigenvector vob (t) and the eigenvalue λob (t) are calculated by the method described in Non-Patent Document 4, for example. The server 10 holds, for each of the user terminals 2-1 and 2-2, the eigenvector vref and the eigenvalue λref acquired at the reference time, and P (t) and Q (Q t) calculate. The reference time may be, for example, the time when the measurement result is first received from each of the user terminals 2-1, 2-2.

  The server 10 correlates the RSSI correlation, the P (t) correlation, and the Q (t) based on the measurement results acquired from the user terminals 2-1 and 2-2 and the calculation results of P (t) and Q (t). ) And the correlation of the magnetic field strength vector, and each correlation is compared with the corresponding threshold (S14). The correlation of the RSSI is the largest correlation among the combinations of all antenna pairs of the user terminals 2-1 and 2-2 of the array antenna. The correlation of the magnetic field strength vector is represented, for example, by taking the position of the user terminal 2-1 as a reference position and normalizing the inner product of the magnetic field strength vector at the reference position and the magnetic field strength vector measured by the user terminal 2-2.

  The server 10 determines whether two or more of the four correlations exceed the corresponding threshold (S15). If two or more of the four correlations exceed the threshold (YES in S15), it is determined that the user terminals 2-1 and 2-2 are in the same room (S16). If two or more correlations that exceed the threshold value are not present (NO in S15), it is determined that they do not exist in the same room (S17). This flow is repeated while the server 10 is in operation, and ends when the server 10 is powered off or when distribution of service information is completed.

  In steps S14 to S16, all of the four correlations may not be used depending on the installation location of the server 10 and the time zone. For example, the pattern B, the pattern C, etc. of FIG.

  FIG. 20 is a schematic block diagram of the server 10. The server 10 includes a processor 11, a memory 15, and a wireless communication unit 17. The wireless communication unit 17 may perform wireless communication with the user terminal 2 via an access point of a wireless LAN such as WiFi, for example, or performs P2P (Peer-to-Peer) communication in which devices are directly connected. May be In the latter case, the wireless communication unit 17 broadcasts the measurement request message, and receives the measurement results of the RSSI and the magnetic field strength vector from two or more user terminals 2.

  The memory 15 has a threshold holding unit 16 and stores a threshold determined in advance for at least one of correlation of RSSI, correlation of P (t), correlation of Q (t), and correlation of residual magnetic field intensity vector . The processor 11 has a feature quantity calculation unit 12 and a comparison determination unit 13. The feature quantity calculation unit 12 calculates P (t) and Q (t) from the received signal voltage, and RSSI, P (t), Q (t) and residual magnetic field strength vector are each calculated between user terminals. Calculate the correlation. When the RSSI is included in the measurement report, the feature quantity calculation unit 12 uses the RSSI, and when the RSSI is not included in the measurement report, the RSSI is calculated from the received signal voltage to calculate the RSSI between the user terminals. The correlation of may be calculated. The comparison determination unit 13 uses each of the thresholds stored in the threshold storage unit 16 to determine whether or not a predetermined number or more of correlation values among the used correlation values exceed the threshold. For example, in the case of using all four correlations (feature amounts), it is determined that the user terminals are in the same room when the threshold value is exceeded by two or more feature amounts. The processor 11 may distribute predetermined service information through the wireless communication unit 17 when it is determined that the user terminals 2 are in the same room.

  The server 10 is a general-purpose computer in which a central processing unit (CPU), a read only memory (ROM), a random access memory (RAM), a communication interface (I / F), a user interface, and the like are interconnected by a bus. It is feasible. Further, one user terminal 2-1 may be, for example, a fixed terminal fixed at the reference position.

  By using the method of the embodiment, it is possible to perform a proximity test that is high in identification accuracy and resistant to changes in the surrounding environment without transmitting and receiving position information.

2-1, 2-2 User terminal (terminal device)
10 server (proximity test equipment)
11 processor 12 feature amount calculation unit 13 comparison determination unit 15 memory 16 threshold value holding unit 17 wireless communication unit

Claims (10)

Obtaining a measurement report including at least a received signal voltage and a residual magnetic field strength vector from the first terminal device and the second terminal device,
From the received signal voltage, a first evaluation function representing a change over time of an eigenvector of the signal and a second evaluation function representing a change over time of an eigenvalue of the signal are calculated;
Based on the measurement report and the calculation result, correlation of received signal strength between the first terminal apparatus and the second terminal apparatus, correlation of the first evaluation function, correlation of the second evaluation function, and the magnetic field strength At least two of vector correlations are used to determine whether the first terminal device and the second terminal device are located in the same space within a predetermined range.
Proximity test method characterized by that. Thresholds are individually set for the correlation of the received signal strength, the correlation of the first evaluation function, the correlation of the second evaluation function, and the correlation of the magnetic field strength vector, respectively.
The first terminal apparatus when two or more of the received signal strength correlation, the first evaluation function correlation, the second evaluation function, and the magnetic field strength vector correlation exceed the corresponding threshold value. And determine that the second terminal device is located in the same space,
The proximity test method according to claim 1, wherein   The correlation of the received signal strength is determined by combining all of a plurality of antenna elements between the signal received by the array antenna of the first terminal apparatus and the signal received by the array antenna of the second terminal apparatus. The proximity test method according to claim 1 or 2, characterized in that the correlation is maximum.   The correlation of the said magnetic field strength vector is obtained by normalizing the inner product of the two-dimensional magnetic field strength vector measured by each of the said 1st terminal device and the said 2nd terminal device, It is characterized by the above-mentioned. Proximity test method.   When the received signal strength is included in the measurement report, the received signal strength is extracted from the measured report to obtain the correlation of the received signal strength, and the received signal strength is not included in the measurement report The proximity test method according to any one of claims 1 to 4, wherein the correlation of the received signal strength is calculated by calculating the received signal strength from the received signal voltage. A communication unit that receives, from the first terminal device and the second terminal device, a measurement report including at least a received signal voltage and a magnetic field strength vector of residual magnetism;
A calculation unit for calculating, from the received signal voltage, a first evaluation function representing a change over time of an eigenvector of a signal and a second evaluation function representing a change over time of an eigenvalue of the signal;
Based on the measurement report and the calculation result, correlation of received signal strength between the first terminal device and the terminal device, correlation of the first evaluation function, correlation of the second evaluation function, and of the magnetic field strength vector A determination unit that determines whether the first terminal device and the second terminal device are located in the same space within a predetermined range using at least two of the correlations;
And a proximity test device characterized by having: A storage unit storing threshold values individually set for the correlation of the received signal strength, the correlation of the first evaluation function, the correlation of the second evaluation function, and the correlation of the magnetic field strength vector,
And have
The determination unit is configured to determine whether the received signal strength correlation, the correlation of the first evaluation function, the correlation of the second evaluation function, and the correlation of the magnetic field strength vector exceed the corresponding threshold by at least two. Determining that the first terminal device and the second terminal device are located in the same space,
7. The proximity test apparatus according to claim 6, wherein:   The calculation unit is the largest of all combinations of a plurality of antenna elements between the signal received by the array antenna of the first terminal apparatus and the signal received by the array antenna of the second terminal apparatus. 8. The proximity test apparatus according to claim 6, wherein a correlation value is calculated as the correlation of the received signal strength.   The said calculation part normalizes the inner product of the two-dimensional magnetic field strength vector measured by each of the said 1st terminal device and the said 2nd terminal device, and calculates the correlation of the said magnetic field strength vector, The proximity test device according to 6 or 7.   When the measurement report includes the received signal strength, the calculation unit calculates the correlation of the received signal strength using the received signal strength included in the measurement report, and the reception of the measurement report is performed. 10. If the signal strength is not included, the received signal strength is calculated from the received signal voltage, and the correlation of the received signal strength is calculated according to any one of claims 6 to 9. Proximity test equipment.

Images