JP4634334B2 - Position estimation system and position estimation method - Google Patents

Description

  The present invention relates to a position estimation system, and more particularly, to a position estimation system that estimates the position of a position estimation target based on detection of a fixed tag and a movable tag.

Conventionally, a method for detecting the position of a user using the radio field strength of a wireless tag has been proposed. However, the radio field strength of a wireless tag is easily affected by the environment, and the radio field strength received by three or more wireless tag readers is used. Even if the position is measured by the simple triangulation used, the accuracy is low.
A technique for estimating the position of a mobile robot using a probabilistic approach has been developed by paying attention to the radio field intensity of the wireless tag (see Non-Patent Document 1). In addition, there is a method in which the position and orientation of the user acquired in advance and the radio wave intensity there are represented by a probability distribution, and the position is estimated from the radio wave intensity obtained at the time of position estimation. In this method, only the radio wave intensity is used for position estimation, and higher position estimation can be performed by using the past history of the user's position together (see Non-Patent Document 2).

D Hahnel, W. Burgard, D. Fox, K. Fishkin, M. Philipose, "Mapping and Location with RFID Technology", Proc. Of 2004 IEEE Int. Conf. On Robotics and Automation, pp. 1015-1020, 2004

Ito, Kawaguchi, "Location estimation system using wireless LAN in real environment and its application" IPSJ SIG, 2004-MBL-30, pp.33-40, 2004/9/17

  However, the method using the probabilistic approach focusing on the radio field intensity of the wireless tag as described in Non-Patent Document 1 is a method used in combination with the odometry of the mobile robot, and is applied to the user position estimation. I can't. Further, the method described in Non-Patent Document 2 has a problem that the estimation model has to be changed again when the user who is the position estimation target moves.

  The present invention has been made in view of the above circumstances, and an object thereof is to provide a position estimation system capable of estimating a user position without changing a probability model even when a position estimation target moves.

According to the first aspect of the present invention, the estimation region is equally divided into n in the X-axis direction and m in the Y-axis direction (m and n are integers), i-th in the X-axis direction, and j in the Y-axis direction. In a position estimation system for estimating the position of a tag reader in a coordinate system (i, j is an integer) in which the th region is represented by X (i, j), a fixed tag that has a fixed position read in advance by the tag reader, and Means for receiving radio wave intensity and tag identification information from a movable tag that is movable with no fixed position defined, and obtained from the radio wave intensity from the received fixed tag with the position of the fixed tag as a center A means for voting on a region through which the circumference of a circle having a radius as a radius passes and a region X (i, j) corresponding to the circle , and the current tag time and the previous tag reader centered on the previously estimated tag reader position. The position of And means for voting for a region present in the circle X (i, j) of the elapsed time and distance determined from moving speed previously defined for the reader between the time and the radius, Means for calculating a probability that a tag reader exists in the area X (i, j) based on a value of a vote result voted on the area X (i, j), and radio wave intensity and movement from the received movable tag Based on the identification information of the movable tag, the means for calculating the distance between the movable tag and the tag reader, and the calculated distance between the movable tag and the tag reader around the previously estimated position of the movable tag. Only movable tags that pass through a region X (i, j) in which the value of the vote result voted for the region through which the circumference of the circle as the radius passes and the region X (i, j) in the vicinity thereof are equal to or greater than the threshold value Using the previously estimated A means for voting on a region through which a circumference of a circle whose radius is the distance between the calculated movable tag and the tag reader and a region X (i, j) in the vicinity thereof centered on the position of the movable tag; A position estimation system comprising: means for estimating a weighted center of gravity of an area X (i, j) having a vote result value equal to or greater than a predetermined threshold as a tag reader position .
According to a second invention of the present invention, in the first invention, among the movable tags, identification of the movable tag that cannot be voted is based on the radio wave intensity received from the movable tag that cannot be voted. Means for calculating a distance between the movable tag indicated by the information and the tag reader, and the movable information indicated by the calculated identification information of the movable tag that cannot be voted about the position of the tag reader Means for voting on a region where the circumference of a circle having a radius between the tag and the tag reader passes and a region X (i, j) in the vicinity thereof, and an attribute of the movable tag that cannot be voted on Based on the above, the means for voting a voting value of a predetermined value to a predetermined area X (i, j), and the voting result voted for the movable tag that cannot be voted Based on the value, the means for calculating the probability that the movable tag that cannot be voted exists in the region X (i, j), and the value of the voting result voted for the movable tag that cannot be voted are threshold values. A position estimation system further comprising means for estimating the weighted centroid of the region X (i, j) as described above as the position of the movable tag that cannot be voted.
According to a third invention of the present invention, in the first invention, the value of voting for the region through which the circle circumference passes is different from the value of voting for the region in the vicinity of the circle circumference. A position estimation system characterized by that.

  According to the present invention, it is possible to estimate the user position without changing the probability model even if the position estimation target moves.

A position estimation system according to an embodiment of the present invention will be described below with reference to the drawings.
FIG. 1 is a diagram for explaining a use environment of a position estimation system according to an embodiment of the present invention.
In the figure, a user 1 has a tag reader 2. The tag reader 2 includes tag identification information and radio wave intensity from wireless tags 11-1 to 11-6 attached to non-moving objects (wall 3, desk 4, 5, stand 6) or moving objects (chairs 7, 8). Is output to a position calculation server (to be described later) that performs position estimation.

  In this embodiment mode, a wireless tag attached to an object that does not move, such as a wall or a desk, is a fixed tag, and a wireless tag attached to an object that can move, such as a chair, is a movable tag. Further, as a precondition, the initial positions of the wireless tags 11-1 to 11-6 are known on a map on the coordinate system A. The user moves with the tag reader 2 and detects identification information and radio wave intensity from the wireless tags 11-1 to 11-6.

In the embodiment of the present invention, the position of a user is estimated while updating the position of the moved movable tag even when the positions of some of the movable tags move.
FIG. 2 is a diagram for explaining a hardware environment of the position estimation system.

  As shown in the figure, an access point 22 and a position estimation server 23 are connected to a network 21 such as the Internet. The tag reader 2 transmits the detected tag identification information and radio wave intensity to the position estimation server 23 via the access point 22 and the network 21.

The position estimation server 23 performs position estimation processing according to the embodiment of the present invention based on the tag identification information and the radio wave intensity sent from the tag reader 2.
FIG. 3 is a diagram illustrating a configuration of the position estimation server.
As shown in the figure, a CPU 32, a communication I / F 33, a memory 34, and a storage device 35 are connected to the bus 31.

The CPU 32 cooperates with the position estimation program 41 stored in the storage device 35 to perform position estimation processing according to the embodiment of the present invention.
The communication I / F 33 controls communication via the network 21. The memory 34 is used as a work area when the position estimation program is executed.

  The storage device 35 stores a position estimation program 41 and a database 42. The position estimation program 41 is for realizing the position estimation process according to the embodiment of the present invention. The database 42 associates information necessary for executing the position estimation process in the embodiment of the present invention such as tag identification information, radio wave intensity, tag attribute information, detection time, position information of each tag, position information of the tag reader, etc. Store. Tag attribute information indicates a tag attribute, for example, a fixed tag, a movable tag type, a target to which a tag such as a desk or a chair is attached, and is grasped from a predetermined tag identification information definition It is.

FIG. 4 is a flowchart for explaining position estimation processing according to the embodiment of the present invention.
As shown in the figure, when the position estimation process is started, first, identification information and radio wave intensity are read from each wireless tag by the tag reader (S1). Next, the distance between the fixed tag and the tag reader is calculated from the radio field intensity and identification information of the fixed tag among the read wireless tags. Various methods for calculating the distance are conceivable. For example, the method disclosed in Japanese Patent Application Laid-Open No. 2004-53510, or the radio wave intensity and distance may be measured and used in advance.

Then, the position of the tag reader is voted using the radio wave intensity from the fixed tag whose calculated distance and position are known (S2). As shown in FIG. 5, this vote takes a coordinate system A in which the estimation area is equally divided into n in the X-axis direction and m in the Y-axis direction. For example, the position vector [X _No , Y _No "2" is added to a region X (i, j) through which a circle having a radius of the distance r _No to the tag reader obtained from the radio wave intensity passes.

In addition, a circle whose center is the position vector [X _No , Y _No ] of each fixed tag and whose radius is the distance r _No to the tag reader obtained from the radio wave intensity, and r _No ± dr is the radius on the same center. 1 is added to the region X (i, j) between the circles to be performed. FIG. 6 is a diagram illustrating a case where the position of the tag reader 2 is voted from the fixed tags 11-1 to 11-3 whose positions are already known.

  Next, the estimated position of the tag reader 2 at the present (time t) is voted from the position of the tag reader at the past (time t-1) (S3). In this voting, first, an area where a tag reader is estimated is obtained. Specifically, for the region estimated to have a tag reader, for example, the movement distance h is obtained from the elapsed time between time t and time t−1 and the walking speed of the person.

Then, the position of the tag reader at time t can be estimated to be within a circle with a radius h, centered on the tag reader position vector X _{user t-1} at time t-1. Next, vote at an area X (i, j) where it is estimated that there is a tag reader at time t. In voting, 1 is added to an area X (i, j) in a circle where it can be estimated that there is a tag reader. FIG. 7 is a diagram illustrating a case where the estimated position of the current tag reader is voted from the position of the tag reader at time t−1.

Thereafter, the probability p _ij of the position of the tag reader at the current time t is obtained from the voting result of the fixed tag and the voting result for the area where the tag reader is expected to be at time t based on the equation (1) (S4).
p _ij = q _ij / Σq _ij (1)
However, q _ij is the total value of the values voted for the region X (i, j). Next, the position of the tag reader is voted based on the radio wave intensity from the movable tag and the position of the movable tag on the map (S5). ).

Specifically, first, the distance r _{M_No} between the movable tag and the tag reader is calculated from the detected radio wave intensity e _{M_No} (M_No is the ID of the movable tag) from the movable tag. Various methods for calculating the distance are conceivable. For example, the method disclosed in Japanese Patent Application Laid-Open No. 2004-53510, or the radio wave intensity and distance may be measured and used in advance.

A vote is then taken. At this time, a movable tag that cannot vote in a region X (i, j) that has a probability p _ij > α (α: constant) that is voted with a fixed tag is determined to have a possibility of movement and is not voted. .
The voting is based on the position vector [X _{M — No} , Y _{M — No} ] of each movable tag, and 2 is given to a region X (i, j) through which a circle having a radius r r _{M — No} from the radio wave intensity to the tag reader passes. to add. Also, the position vector [X _{M — No} , Y _{M — No} ] of each movable tag is the center, the same center as the circle whose radius is the distance r _{M — No} from the radio wave intensity to the tag reader, and r _{M — No} ± dr is the radius. Add 1 to the area X (i, j) between the circles. FIG. 8 is a diagram for explaining voting by a movable tag that can vote for an area having a predetermined probability or more. Thereafter, the voting value of the region where the voting value <β of the obtained movable tag is set to 0. FIG. 9 is a diagram for explaining a case where the region where the vote value <β (= 3) in FIG. 8 is 0.

Then, to calculate the weight w _ij of the position of the tag reader from voting value (S6).
The weight calculation process in S6 is performed for an area where the vote value β ≧ 3. The calculation of the weight is performed based on the following formula (2).
w _ij = p _ij d _ij / Σp _kl d _kl (2)
(K = 1,2,3, ..., n, l = 1,2,3, ..., m)
Here, p _ij is the probability of the movable tag, and d _ij is the vote value for the region X (i, j) obtained by voting the movable tag. FIG. 10 is a diagram illustrating a region where the vote value β ≧ 3 for performing the weighting process.

Next, the position of the tag reader is calculated based on Equation (3) based on the weight calculated in S6.
p _t = Σw _kl x _kl (3)
(K = 1,2,3, ..., n, l = 1,2,3, ..., m)
Thereby, the position of the tag reader can be estimated. FIG. 11 is a diagram illustrating the estimated position of the tag reader.

Next, a vote of the position of the movable tag that may have moved is performed (S8).
This vote First, a moving tag candidate vectors m_x _it from the radio wave strength of the movable tags that are determined to have the movement possibilities.
m_x _it = {x _t , m _it ) (4)
However, m _it indicates field intensity of movable tags i that might have moved measured by the tag reader position xt of time t. Next, voting is performed using a past moving tag candidate vector of a movable tag that is determined to be movable.

For example, the position of the movement candidate tag i is voted with the movement tag candidate vector m_x _it (t = s0 to s1) from time t = s0 to time t = s1. In this embodiment, for example, the position vector [X _{t, Y t]} tag reader each time the centered, area through which circle the distance r _it to tag reader obtained from the radio wave strength m _it as a radius X ( Add 2 to i, j). The position vector _{[X t, Y} t] tag reader each time the center and a circle that the distance _{r it} to tag reader obtained from the radio wave strength _{m it} as a radius, the _{r it} ± dr same central radius 1 is added to the region X (i, j) between the circles. FIG. 12 is a diagram for explaining a voting result at each time of a movable tag that may have moved.

  Next, the position of the movement candidate tag is voted based on the attribute of the movable tag (S9). In this embodiment, it is assumed that the movement candidate tag is a chair, the tags 11-3 and 11-2 are desks, and the chair is in the vicinity of the desk. 1 is added to a region X (i, j) in a circle having a predetermined radius R. FIG. 13 is a diagram illustrating a case where voting is performed on a region within a circle having a radius R centered on the fixed tags 11-2 and 11-3 in the region illustrated in FIG.

Next, the existence probability m_p _imn of the movement candidate tag i is calculated based on the equation (5) (S10).
m_p _imn = m_q _imn / Σm_q _ikl (5)
(K = 1,2,3, ..., n, l = 1,2,3, ..., m)
However, m_q _imn is the vote value of the tag i at the position X (m, n).

After the process in S10, the weight of the movement candidate tag i is calculated (S11).
In calculating the weight, the weight m_w _imn is calculated based on the following formula (6) at the position X (m, n) that is equal to or higher than the probability m_p _imn > γ (γ: predetermined value). FIG. 14 is a diagram illustrating a region to which a weight is given in the region illustrated in FIG. 13.

m_w _imn = m_p _imn / Σm_p _ikl (6)
(K = 1,2,3, ..., n, l = 1,2,3, ..., m)
Thereafter, the position of the wireless tag i that may have moved is calculated based on the equation (7), and the position of the movable tag i is updated in the map (S12). Specifically, the calculated position information of the movable tag is registered in the database.

m_x _i = Σm_w _kl x _kl (7)
(K = 1,2,3, ..., n, l = 1,2,3, ..., m)
Thereby, the position of the moved tag can be automatically updated, and the position of the user can be continuously grasped in an environment where a fixed tag and a movable tag are mixed. FIG. 15 is a diagram illustrating an area in which the position of the movable tag is updated.

  According to the embodiment of the present invention, the presence area of the user holding the current tag reader from the radio wave intensity of the fixed tag and the presence area of the user estimated from the tag position given in advance and the past user position. Is expressed by a probability distribution. Then, the weight of the user existing area is calculated from the radio field intensity of the movable tag, the predetermined tag position determined in advance, and the probability distribution of the user existing area, and the gravity center position of the weight is determined by the user. Estimated position.

  According to the embodiment of the present invention, it is possible to estimate the position of the user with high accuracy by eliminating the radio wave intensity of the movable tag that cannot be voted for the user's existing area among the movable tags. Further, it is possible to estimate the position of a movable tag that cannot be voted for the user's presence area from the user's movement history and the radio wave intensity at each time point.

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

  In addition, the method described in the embodiment is, for example, a magnetic disk (floppy (registered trademark) disk, hard disk, etc.), optical disk (CD-ROM, DVD) as programs (software means) that can be executed by a computer (computer). , MO, etc.), semiconductor memory (ROM, RAM, flash memory, etc.), etc., and can also be transmitted and distributed by communication media. The program stored on the medium side includes a setting program that configures software means (including not only the execution program but also a table and data structure) in the computer. A computer that implements this apparatus reads a program recorded on a recording medium, constructs software means by a setting program as the case may be, and executes the above-described processing by controlling the operation by this software means. The recording medium referred to in this specification is not limited to distribution, but includes a storage medium such as a magnetic disk or a semiconductor memory provided in a computer or a device connected via a network.

It is a figure for demonstrating the use environment of the position estimation system concerning embodiment of this invention. It is a figure for demonstrating the hardware environment of a position estimation system. It is a figure which shows the structure of a position estimation server. It is a flowchart for demonstrating the position estimation process which concerns on embodiment of this invention. It is a figure for demonstrating the coordinate system of an estimation area | region. It is a figure which shows the case where the position of a tag reader is voted from the electromagnetic wave intensity from a fixed tag with a known position. It is a figure which shows the case where the position of the present tag reader is voted from the position of the tag reader at the time of time t-1. It is a figure for demonstrating the vote by the movable tag which can vote to the area | region more than predetermined probability. FIG. 9 is a diagram for explaining a case where an area where vote value <β (= 3) is 0 in FIG. 8. It is a figure which shows the area | region where voting value (beta)> = 3 for performing a weighting process. It is a figure which shows the position of the estimated tag reader. It is a figure for demonstrating the vote result in each time of the movable tag which may have moved. It is a figure which shows the case where voting is performed to the area | region in the circle | round | yen of a radius centering on a fixed tag in the area | region shown in FIG. It is a figure which shows the area | region where a weight is provided in the area | region shown in FIG. It is a figure which shows the area | region which updated the position of the movable tag.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... User, 2 ... Tag reader, 3 ... Wall 4, 5 ... Desk, 6 ... Stand, 7, 8 ... Chair, 11-1 to 11-6 ... Wireless tag, 21 ... Network, 22 ... Access point, 23 ... Position estimation server, 31 ... bus, 32 ... CPU, 33 ... communication interface, 34 ... memory, 35 ... storage device, 41 ... position estimation program, 42 ... database.

Claims (6)

The estimated area is equally divided into n in the X-axis direction and m in the Y-axis direction (m and n are integers), and the i-th area in the X-axis direction and the j-th area in the Y-axis direction are X (i, j). In a position estimation system that estimates the position of a tag reader in a coordinate system represented (i and j are integers),
Means for receiving radio wave intensity and tag identification information from a fixed tag that is read in advance by a tag reader and whose fixed position is fixed and a movable tag that is not fixed and movable;
Vote for a region passing through the circumference of a circle whose radius is a distance obtained from the received radio wave intensity from the fixed tag and a region X (i, j) corresponding to the vicinity of the fixed tag. Means for
Around the position of the previous estimated tag reader, the elapsed time and the circumference of the moving distance and the radius as determined from the moving speed defined in advance of the tag reader between time estimating the position of the current time and the previous tag reader Means for voting on a region X (i, j) existing in
Means for calculating a probability that a tag reader exists in the area X (i, j) based on a value of a voting result voted on the area X (i, j) ;
Means for calculating a distance between the movable tag and the tag reader based on the received radio wave intensity from the movable tag and the identification information of the movable tag;
Centering on the position of the previously estimated movable tag, with respect to a region through which the circumference of a circle whose radius is the distance between the calculated movable tag and the tag reader and a region X (i, j) in the vicinity thereof The calculated movable tag centering on the position of the movable tag estimated last time using only the movable tag passing through the region X (i, j) where the value of the voted voting result is equal to or greater than the threshold value; Means for voting on a region through which the circumference of a circle having a radius between the tag readers passes and a region X (i, j) in the vicinity thereof;
A position estimation system comprising: means for estimating a weighted center of gravity of an area X (i, j) where a value of a voting result is equal to or greater than a predetermined threshold value as a tag reader position . The distance between the movable tag indicated by the identification information of the movable tag that cannot be voted and the tag reader is calculated from the radio wave intensity received from the movable tag that cannot be voted among the movable tags. Means to
An area through which the circumference of a circle whose radius is the distance between the movable tag and the tag reader indicated by the calculated identification information of the movable tag that cannot be voted, centered on the position of the tag reader, and its Means for voting on a nearby region X (i, j);
A means for voting a predetermined voting value to a predetermined area X (i, j) based on an attribute of the movable tag that cannot be voted;
Means for calculating a probability that a movable tag that cannot be voted exists in the region X (i, j) based on a value of a vote result voted for the movable tag that cannot be voted ;
Means for estimating a weighted centroid of a region X (i, j) in which a value of a voted vote result for a movable tag that cannot be voted is equal to or greater than a threshold value as a position of the movable tag that cannot be voted; The position estimation system according to claim 1, further comprising:   The position estimation system according to claim 1, wherein a voting value for an area through which the circumference of the circle passes is different from a voting value for an area in the vicinity of the circumference of the circle. The estimated area is equally divided into n in the X-axis direction and m in the Y-axis direction (m and n are integers), and the i-th area in the X-axis direction and the j-th area in the Y-axis direction are X (i, j). In the position estimation method in the position estimation system for estimating the position of the tag reader in the coordinate system represented (i and j are integers),
Receiving radio wave intensity and tag identification information from a fixed tag that is read in advance by a tag reader and whose fixed position is fixed and a movable tag that is not fixed and movable;
Vote for an area X (i, j) corresponding to a circle passing through a circle whose radius is a distance obtained from the received radio wave intensity from the fixed tag centered on the position of the fixed tag. Do
Around the position of the previous estimated tag reader, the elapsed time and the circumference of the moving distance and the radius as determined from the moving speed defined in advance of the tag reader between time estimating the position of the current time and the previous tag reader Vote for the region X (i, j) existing in
Calculating a probability that a tag reader exists in the region X (i, j) based on a value of a vote result voted on the region X (i, j) ;
Based on the received radio wave intensity from the movable tag and the identification information of the movable tag, the distance between the movable tag and the tag reader is calculated,
Centering on the position of the previously estimated movable tag, the region passing through the circumference of the circle whose radius is the distance between the calculated movable tag and the tag reader and the region X (i, j) in the vicinity thereof The calculated movable tag centered on the position of the previously estimated movable tag using only the movable tag passing through the region X (i, j) where the value of the voted voting result is equal to or greater than the threshold value; Vote for the region through which the circumference of the circle whose radius is the distance between the tag readers and the region X (i, j) in the vicinity thereof,
A position estimation method in a position estimation system, wherein a weighted center of gravity of an area X (i, j) where a value of a voting result is equal to or greater than a predetermined threshold is estimated as a tag reader position . The distance between the movable tag indicated by the identification information of the movable tag that cannot be voted and the tag reader is calculated from the radio wave intensity received from the movable tag that cannot be voted among the movable tags. And
An area through which the circumference of a circle whose radius is the distance between the movable tag and the tag reader indicated by the calculated identification information of the movable tag that cannot be voted , centered on the position of the tag reader, and its Vote for nearby region X (i, j)
Voting a predetermined voting value to a predetermined region X (i, j) based on an attribute of the movable tag that cannot be voted;
Calculating a probability that a movable tag that cannot be voted exists in the region X (i, j) based on a value of a vote result voted for the movable tag that cannot be voted ;
Estimating the weighted centroid of the region X (i, j) where the value of the voted vote result for the movable tag that cannot be voted is equal to or greater than a threshold value as the position of the movable tag that cannot be voted. The position estimation method according to claim 4, further comprising:   5. The position estimation method according to claim 4, wherein a voting value for an area through which the circumference of the circle passes is different from a voting value for an area in the vicinity of the circumference of the circle.

Images