JP6396765B2 - Estimation method and estimation apparatus using the same - Google Patents

Estimation method and estimation apparatus using the same Download PDF

Info

Publication number
JP6396765B2
JP6396765B2 JP2014227152A JP2014227152A JP6396765B2 JP 6396765 B2 JP6396765 B2 JP 6396765B2 JP 2014227152 A JP2014227152 A JP 2014227152A JP 2014227152 A JP2014227152 A JP 2014227152A JP 6396765 B2 JP6396765 B2 JP 6396765B2
Authority
JP
Japan
Prior art keywords
node
value
processing unit
node processing
average value
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
JP2014227152A
Other languages
Japanese (ja)
Other versions
JP2016090471A (en
Inventor
レザ カハル アジズ ムハンマド
レザ カハル アジズ ムハンマド
アンワル コイルー
アンワル コイルー
松本 正
正 松本
Original Assignee
株式会社光電製作所
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 株式会社光電製作所 filed Critical 株式会社光電製作所
Priority to JP2014227152A priority Critical patent/JP6396765B2/en
Publication of JP2016090471A publication Critical patent/JP2016090471A/en
Application granted granted Critical
Publication of JP6396765B2 publication Critical patent/JP6396765B2/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Description

  The present invention relates to an estimation technique, and relates to an estimation method for estimating the position of an unknown source and an estimation apparatus using the estimation method.
  Wireless position estimation is an important technology in recent wireless systems, and for example, a factor graph is used. In a factor graph, messages are exchanged in the form of mean and variance values with a Gaussian distribution between multiple local functions factored into a generic function. The exchanged message is generated by TOA (Time Of Arrival) (see, for example, Non-Patent Document 1).
Jung-Chieh Chen, Yeong-Cheng Wang, Ching-Shyang Maa and Junn-Tsair Chen IRFN ETR 5, NO. 10, p. 2696-2704
  In position estimation based on TOA, synchronization between a wireless device to be estimated and a sensor is required. However, when the wireless device to be estimated is an illegal wireless station or an unknown transmission source, it is difficult to establish synchronization. On the other hand, position estimation based on TDOA (Time Difference Of Arrival) is used in order to eliminate synchronization with a wireless device to be estimated and only synchronization between sensors. Therefore, it is desired to improve the estimation accuracy of position estimation based on TDOA.
  The present invention has been made in view of such circumstances, and an object thereof is to provide a technique for improving position estimation accuracy.
  In order to solve the above-described problem, an estimation device according to an aspect of the present invention includes an acquisition unit that receives a signal from a target wireless device at each of a plurality of sensors and acquires a reception time difference between different sensors. And a first factor node process for deriving an average value and a variance value of a distance in TDOA (Time Difference of Arrival) assuming that the measurement value error follows a Gaussian distribution based on the reception time difference acquired by the acquisition unit Part, a second factor node processing unit for converting the average value and variance value of the distance in TDOA to the average value and variance value of the distance in TOA (Time Of Arrival), and the average value and variance of the distance in TOA A first variable node processing unit that executes a sum product algorithm on the value and a distance of the TOA A third factor node processing unit for performing conversion between the average value and the variance value, and the average value and the variance value of the position coordinates from the sensor; the average value and the variance value of the position coordinates from the sensor; A fourth factor node processing unit for performing conversion between the average value and the variance value of the position coordinates, and a second variable node process for executing the sum product algorithm on the average value and the variance value of the position coordinates of the wireless device And the processing performed in the order of the second factor node processing unit, the first variable node processing unit, the third factor node processing unit, the fourth factor node processing unit, the second variable node processing unit, and the order is reversed. A control unit that repeatedly executes the processes performed in the order of, and an output unit that outputs an average value of the position coordinates of the wireless device after the process repeatedly performed by the control unit.
  Another aspect of the present invention is an estimation method. In this method, a signal from a target wireless device is received by each of a plurality of sensors, an acquisition step of acquiring a reception time difference between different sensors, and a measurement value error based on the acquired reception time difference. Is calculated from the first factor node processing step for deriving the average value and the variance value of the distance in TDOA (Time Difference of Arrival), and the average value and the variance value of the distance in TDOA. Of Arriv), a second factor node processing step for performing conversion to an average value and a variance value of distances, and a first variable node processing step for executing a sum product algorithm on the average value and variance values of distances in TOA And the average value and dispersion value of the distance in TOA and the position from the sensor A third factor node processing step for performing conversion between the average value and the variance value of the coordinates, the average value and the variance value of the position coordinates from the sensor, and the average value and the variance value of the position coordinates of the wireless device A fourth factor node processing step for executing the transformation of the second factor node processing step, a second variable node processing step for executing the sum product algorithm on the average value and the variance value of the position coordinates of the wireless device, a second factor node processing step, The processing performed in the order of the one variable node processing step, the third factor node processing step, the fourth factor node processing step, and the second variable node processing step, and the processing performed in the reverse order are performed repeatedly. And outputting an average value of the position coordinates of the wireless device.
  It should be noted that any combination of the above-described constituent elements and a conversion of the expression of the present invention between a method, an apparatus, a system, a recording medium, a computer program, etc. are also effective as an aspect of the present invention.
  According to the present invention, position estimation accuracy can be improved.
It is a figure which shows the structure of the estimation apparatus which concerns on the Example of this invention. It is a flowchart which shows the procedure of the estimation process by the estimation apparatus of FIG. It is a flowchart which shows the procedure of the downward direction process of FIG. It is a flowchart which shows the procedure of the up direction process of FIG.
  Before describing the present invention specifically, an outline will be given first. Embodiments of the present invention relate to an estimation device that estimates the position of an illegal radio station or an unknown source (hereinafter referred to as “radio device”). A plurality of sensors are connected to the estimation device. The TDOA measurement is indicated by the difference in the TOA measurement between the two sensors. Here, an equivalent measured value of TOA is derived from the measured value of TDOA. As a result, TOA-based factor graph location estimation is extended to TDOA-based factor graph location estimation.
  FIG. 1 shows a configuration of an estimation apparatus 100 according to an embodiment of the present invention. The estimation apparatus 100 includes a first sensor 10a, a second sensor 10b, a third sensor 10c, a fourth sensor 10d, an acquisition unit 12, a first factor node processing unit 14, an external variable node processing unit 16, which are collectively referred to as a sensor 10. Second factor node processing unit 18, first variable node processing unit 20, third factor node processing unit 22, internal variable node processing unit 24, fourth factor node processing unit 26, second variable node processing unit 28, control unit 30 The output unit 32 is included. The third factor node processing unit 22 includes a zero value avoiding unit 34 and an imaginary value avoiding unit 36, and the first variable node processing unit 20 includes an infinity avoiding unit 38. Here, as an example, the number of sensors 10 is “4”, but the number of sensors 10 is not limited to “4”.
Timing synchronization is established between the plurality of sensors 10. Since a known technique may be used for the timing synchronization, the description is omitted here. Each sensor 10 receives a signal from a wireless device (not shown), and outputs the reception time of receiving the signal to the acquisition unit 12 as a measurement value. The acquisition unit 12 acquires the reception time from each sensor 10. The acquisition unit 12 calculates a reception time difference between different sensors 10. For example, the reception time difference between the reception time at the first sensor 10a and the reception time at the second sensor 10b is indicated as “Δτ 1,2 ”. When the number of sensors 10 is “4”, six types of reception time differences can be acquired by the acquisition unit 12.
The first factor node processing unit 14 includes E 1 , 2 node 50, E 1 , 3 node 52, E 1 , 4 node 54, E 2 , 3 node 56, E 2 , 4 node 58, E 3 , 4 node 60. including. E 1, 2 node 50, E 1 , 3 node 52, E 1 , 4 node 54, E 2 , 3 node 56, E 2 , 4 node 58, E 3 , 4 node 60 The reception time difference between the sensor 10b, the reception time difference between the first sensor 10a and the third sensor 10c, the reception time difference between the first sensor 10a and the fourth sensor 10d, the second sensor 10b and the third sensor 10c. Reception time difference between the second sensor 10b and the fourth sensor 10d, and reception time difference between the third sensor 10c and the fourth sensor 10d are received from the acquisition unit 12, respectively.
  Each node included in the first factor node processing unit 14 is based on the reception time difference acquired by the acquisition unit 12 and assumes that the measurement value error follows a Gaussian distribution. Is derived. In addition, since a well-known technique should just be used for derivation | leading-out, description is abbreviate | omitted here. The first factor node processing unit 14 outputs the derived average value and variance value to the external variable node processing unit 16.
The external variable node processing unit 16 includes d 1 , 2 node 62, d 1 , 3 node 64, d 1 , 4 node 66, d 2 , 3 node 68, d 2 , 4 node 70, d 3 , 4 node 72. Including. d 1, 2 node 62, d 1 , 3 node 64, d 1 , 4 node 66, d 2 , 3 node 68, d 2 , 4 node 70, d 3 , 4 node 72 are E 1 , 2 node 50, Connected to E 1,3 node 52, E 1,4 node 54, E 2,3 node 56, E 2,4 node 58, E 3,4 node 60, respectively, and receives an average value and a variance value. The external variable node processing unit 16 outputs the accepted average value and variance value to the second factor node processing unit 18. mean and variance values in each of the d 3, 4 node 72 from d 1, 2 node 62 is expressed as follows.
Here, “m” indicates an average value, and “σ 2 ” indicates a variance value. The subscript “d i, j ” indicates the d i, j node.
  From the second factor node processing unit 18 to the second variable node processing unit 28, an iterative process with a factor graph is performed. Here, after describing the outline of the process of each component requirement and the connection relationship, the control content of the iterative process will be described, and finally the details of the process of each component requirement will be described. First, the outline of the processing of each component and the connection relationship will be described. The second factor node processing unit 18 connects the external variable node processing unit 16 to one end side, and connects the first variable node processing unit 20 to the other end side. The second factor node processing unit 18 performs conversion from the average value and variance value of the distance in TDOA to the average value and variance value of the distance in TOA. The first variable node processing unit 20 connects the second factor node processing unit 18 to one end side, and connects the third factor node processing unit 22 to the other end side. The first variable node processing unit 20 executes the sum product algorithm on the average value and the variance value of the distance in TOA.
  The third factor node processing unit 22 connects the first variable node processing unit 20 to one end side, and connects the internal variable node processing unit 24 to the other end side. The third factor node processing unit 22 performs conversion between the average value and variance value of the distance in the TOA, and the average value and variance value of the position coordinates from the sensor 10. The internal variable node processing unit 24 connects the third factor node processing unit 22 to one end side and connects the fourth factor node processing unit 26 to the other end side. The internal variable node processing unit 24 delivers the average value and the variance value of the position coordinates from the sensor 10 between the third factor node processing unit 22 and the fourth factor node processing unit 26.
  The fourth factor node processing unit 26 connects the internal variable node processing unit 24 to one end side and connects the second variable node processing unit 28 to the other end side. The fourth factor node processing unit 26 performs conversion between the average value and the variance value of the position coordinates from the sensor 10 and the average value and the variance value of the position coordinates of the wireless device. The second variable node processing unit 28 connects the fourth factor node processing unit 26 to one end side and connects the output unit 32 to the other end side. The second variable node processing unit 28 executes a sum product algorithm on the average value and the variance value of the position coordinates of the wireless device.
  Next, the control content of the iterative process will be described. In the iterative processing, the second factor node processing unit 18, the first variable node processing unit 20, the third factor node processing unit 22, the internal variable node processing unit 24, the fourth factor node processing unit 26, and the second variable node processing unit 28 are used. (Hereinafter referred to as “downward direction processing”) and the reverse order, that is, the second variable node processing unit 28, the fourth factor node processing unit 26, the internal variable node processing unit 24, The processes performed in the order of the third factor node processing unit 22, the first variable node processing unit 20, and the second factor node processing unit 18 (hereinafter referred to as “upward direction processing”) are alternately and repeatedly executed. Such repeated processing is executed by the control unit 30. In the following, the downstream process is described, and then the upstream process is described.
The downstream processing will be described in detail. The second factor node processing unit 18 includes a D 1 , 2 node 74, a D 1 , 3 node 76, a D 1 , 4 node 78, a D 2 , 3 node 80, a D 2 , 4 node 82, and a D 3 , 4 node 84. including. D 1, 2 node 74, D 1 , 3 node 76, D 1 , 4 node 78, D 2 , 3 node 80, D 2 , 4 node 82, D 3 , 4 node 84 are d 1 , 2 node 62, d1,3 node 64, d1,4 node 66, d2,3 node 68, d2,4 node 70, d3,4 node 72 are connected respectively. Each node receives the average value and variance value of the distance in TDOA, and converts them into the average value and variance value of the distance in TOA as follows.
Here, the subscript “X → Y” indicates output from the X node to the Y node, the subscript “D i, j ” indicates the D i, j node, and the subscript “r”. “ i ” indicates an r i node described later. Moreover, Formula (2) and Formula (3) show the average value of the distance in TOA, and Formula (4) and Formula (5) show the dispersion value of the distance in TOA.
The first variable node processing unit 20 includes an r 1 node 86, an r 2 node 88, an r 3 node 90, and an r 4 node 92, and each of these four nodes corresponds to the sensor 10 on a one-to-one basis. r 1 node 86 is connected to D 1 , 2 node 74, D 1 , 3 node 76, D 1 , 4 node 78, and r 2 node 88 is connected to D 1 , 2 node 74, D 2 , 3 node 80, D 2,4 node 82 is connected, r 3 node 90 is connected to D 1,3 node 76, D 2,3 node 80, D 3,4 node 84, and r 4 node 92 is connected to D 1,4 Node 78, D 2 , 4 node 82, D 3 , 4 are connected to node 84. Each node executes the sum product algorithm on the average value and the variance value of the distance in TOA as follows.
  Here, “N” indicates the number of sensors 10 and is “4” in the case of FIG. Expression (6) is an execution result of the sum product algorithm for the average distance value in TOA, and Expression (7) is an execution result of the sum product algorithm for the distance dispersion value in TOA. In the following, the execution result of the sum product algorithm may also be referred to as an average value and a variance value of distances in TOA.
The third factor node processing unit 22 includes a C 1 node 110, a C 2 node 112, a C 3 node 114, and a C 4 node 116, and also includes a zero value avoiding unit 34. The zero value avoiding unit 34 executes the following processing when each node included in the first variable node processing unit 20 passes the average value and the variance value of the distance in the TOA to the third factor node processing unit 22. To do. The zero value avoiding unit 34 changes the average value of the distance in the TOA to a predetermined value if the average value of the distance in the TOA is smaller than the predetermined value as follows.
Here, “ε” represents a predetermined value. This is a process for preventing the distance from degenerating to zero.
The C 1 node 110, the C 2 node 112, the C 3 node 114, and the C 4 node 116 are transferred to the r 1 node 86, the r 2 node 88, the r 3 node 90, and the r 4 node 92 via the zero value avoidance unit 34. Each is connected. Each node receives the average value and the variance value of the distance in the TOA, and converts the average value and the variance value of the position coordinates from the sensor 10 as follows.
Expression (10) is an average value output from each node to the later-described Δx 1 node 118, Δx 2 node 120, Δx 3 node 122, and Δx 4 node 124, and the average of the position coordinates from the sensor 10 Indicates the value. Expression (11) is a variance value output from each node to the later-described Δx 1 node 118, Δx 2 node 120, Δx 3 node 122, and Δx 4 node 124, and the position coordinates from the sensor 10 The variance value of.
Expression (12) is an average value output from each node to the later-described Δy 1 node 126, Δy 2 node 128, Δy 3 node 130, and Δy 4 node 132, and the position coordinates from the sensor 10 The average value is shown. Expression (13) is a variance value output from each node to a later-described Δy 1 node 126, Δy 2 node 128, Δy 3 node 130, and Δy 4 node 132, and the position coordinates from the sensor 10 The variance value of.
The internal variable node processing unit 24 includes a Δx 1 node 118, a Δx 2 node 120, a Δx 3 node 122, a Δx 4 node 124, a Δy 1 node 126, a Δy 2 node 128, a Δy 3 node 130, and a Δy 4 node 132. Each node delivers the average value and the variance value of the position coordinates from the sensor 10 to the fourth factor node processing unit 26 from the third factor node processing unit 22 as follows.
Equation (14) represents the average value output from the Δx 1 node 118, Δx 2 node 120, Δx 3 node 122, and Δx 4 node 124, and Equation (15) represents the Δx 1 node 118, Δx 2 node 120, The variance values output from the Δx 3 node 122 and the Δx 4 node 124 are shown. Equation (16) represents the average value output from the Δy 1 node 126, Δy 2 node 128, Δy 3 node 130, and Δy 4 node 132, and Equation (17) represents the Δy 1 node 126 and Δy 2 node. 128, Δy 3 node 130, and Δy 4 node 132 indicate the variance value output.
The fourth factor node processing unit 26 includes an A 1 node 134, an A 2 node 136, an A 3 node 138, an A 4 node 140, a B 1 node 142, a B 2 node 144, a B 3 node 146, and a B 4 node 148. . Each node receives the average value and the variance value of the position coordinates from the sensor 10 and converts them into the average value and the variance value of the position coordinates of the wireless device as follows.
Expression (18) represents an average value output from the A 1 node 134, the A 2 node 136, the A 3 node 138, and the A 4 node 140, and “X i ” is a position where the i-th sensor 10 is installed. The X coordinate of is shown. The X coordinate is shown as longitude, for example. Expression (19) shows the distributed values output from the A 1 node 134, the A 2 node 136, the A 3 node 138, and the A 4 node 140.
Expression (20) represents an average value output from the B 1 node 142, B 2 node 144, B 3 node 146, and B 4 node 148, and “Y i ” is a position where the i-th sensor 10 is installed. Indicates the Y coordinate. The Y coordinate is indicated as latitude, for example. Equation (21) indicates the variance value output from the B 1 node 142, B 2 node 144, B 3 node 146, and B 4 node 148.
The second variable node processing unit 28 includes an x node 150 and a y node 152. The x node 150 is connected to the A 1 node 134, the A 2 node 136, the A 3 node 138, and the A 4 node 140, and the y node 152 is connected to the B 1 node 142, the B 2 node 144, the B 3 node 146, and the B 4. Connected to node 148. Each node executes a sum product algorithm on the average value and the variance value of the position coordinates of the wireless device. Note that the x node 150 executes two types of processing for each of the average value and the variance value. The first process (hereinafter referred to as “continuation process”) is indicated as follows.
Expression (22) indicates the distributed value output from the x node 150 to the A 1 node 134, the A 2 node 136, the A 3 node 138, and the A 4 node 140, and the expression (23) The average values output to the A 1 node 134, the A 2 node 136, the A 3 node 138, and the A 4 node 140 are shown.
The second process (hereinafter referred to as “termination process”) is indicated as follows.
Expression (24) indicates the variance value generated at the x node 150, and Expression (25) indicates the average value generated at the x node 150. These are not output to the A 1 node 134, the A 2 node 136, the A 3 node 138, and the A 4 node 140. How to use the values generated by these two types of processing will be described later.
Similarly to the x node 150, the y node 152 performs two types of processing. The continuation process is indicated as follows.
The termination process is shown as follows.
The uplink processing will be described in detail. Each node of the fourth factor node processing unit 26 receives an average value and a variance value of the position coordinates of the wireless device from the second variable node processing unit 28. These are the average value and the variance value in the continuation process described above. In particular, the A 1 node 134, the A 2 node 136, the A 3 node 138, and the A 4 node 140 accept the average value and the variance value from the x node 150, and the B 1 node 142, the B 2 node 144, the B 3 node 146, The B 4 node 148 receives the average value and the variance value from the y node 152. Each node converts the average value and variance value of the position coordinates of the wireless device into the average value and variance value of the position coordinates from the sensor 10 as follows.
Expression (30) represents an average value output from the A 1 node 134, the A 2 node 136, the A 3 node 138, and the A 4 node 140, and the expression (31) represents the A 1 node 134, the A 2 node 136, The variance values output from the A 3 node 138 and the A 4 node 140 are shown. Expression (32) represents an average value output from the B 1 node 142, B 2 node 144, B 3 node 146, and B 4 node 148, and Expression (33) represents the B 1 node 142 and B 2 node. 144, B 3 node 146, B shows the dispersion value output from the fourth node 148.
Each node in the internal variable node processing unit 24 delivers the average value and the variance value of the position coordinates from the sensor 10 to the third factor node processing unit 22 from the fourth factor node processing unit 26 as follows.
Equation (34) represents the average value output from the Δx 1 node 118, Δx 2 node 120, Δx 3 node 122, and Δx 4 node 124, and Equation (35) represents Δx 1 node 118, Δx 2 node 120, The variance values output from the Δx 3 node 122 and the Δx 4 node 124 are shown. Expression (36) represents an average value output from the Δy 1 node 126, Δy 2 node 128, Δy 3 node 130, and Δy 4 node 132, and Expression (37) represents the Δy 1 node 126 and Δy 2 node. 128, Δy 3 node 130, and Δy 4 node 132 indicate the variance value output.
The third factor node processing unit 22 includes an imaginary value avoiding unit 36 in addition to the above description. When the imaginary value avoiding unit 36 passes the average value of the position coordinates from the sensor 10 to the fourth factor node processing unit 26, more specifically, from the internal variable node processing unit 24 to the third factor node processing unit 22, The following processing is executed. The imaginary value avoiding unit 36 is as follows if the square value of the average value of the distance in the TOA from the first variable node processing unit 20 is smaller than the square value of the average value of the position coordinates from the sensor 10 as follows. The average value of the position coordinates from the sensor 10 is changed based on the average value of the distance in the TOA from the first variable node processing unit 20.
Here, the expressions (38) and (39) are processes for the average values output from the Δx 1 node 118, the Δx 2 node 120, the Δx 3 node 122, and the Δx 4 node 124. (41) is a process for the average value output from the Δy 1 node 126, the Δy 2 node 128, the Δy 3 node 130, and the Δy 4 node 132. This is a process for preventing an imaginary value from being generated by the process in the third factor node processing unit 22.
The C 1 node 110 is connected to the Δx 1 node 118 and the Δy 1 node 126 through the imaginary value avoiding unit 36, and the C 2 node 112 is connected to the Δx 2 node 120, Δy 2 through the imaginary value avoiding unit 36. The C 3 node 114 is connected to the Δx 3 node 122 and the Δy 3 node 130 via the imaginary value avoiding unit 36, and the C 4 node 116 is connected to the Δx 3 via the imaginary value avoiding unit 36. 4 nodes 124 and Δy are connected to 4 nodes 132. Each node receives the average value and the variance value of the position coordinates from the sensor 10, and converts them into the average value and the variance value of the distance in the TOA as follows.
The first variable node processing unit 20 includes an infinity avoiding unit 38 in addition to the above description. The infinity avoiding unit 38 executes the following processing when passing the dispersion value of the distance in the TOA from the third factor node processing unit 22 to the first variable node processing unit 20. The infinity avoidance unit 38 determines the distance variance value in the TOA if the variance value of the distance in the TOA is larger than the minimum value of the variance value of the distance in the TOA from the second factor node processing unit 18 as follows. Then, the distance is changed to the minimum value of the dispersion value in the TOA from the second factor node processing unit 18.
This is a process for preventing infinity from being generated by the process in the first variable node processing unit 20.
The r 1 node 86 is connected to the C 1 node 110 via the infinity avoidance unit 38, the r 2 node 88 is connected to the C 2 node 112 via the infinity avoidance unit 38, and the r 3 node 90 Is connected to the C 3 node 114 via the infinity avoidance unit 38, and the r 4 node 92 is connected to the C 4 node 116 via the infinity avoidance unit 38. Each node executes the sum product algorithm on the average value and the variance value of the distance in TOA as follows.
As described above, the execution result of the sum product algorithm may also be referred to as a distance average value and a variance value in TOA.
The D 1 and 2 nodes 74 in the second factor node processing unit 18 are connected to the r 1 node 86 and the r 2 node 88, and the D 1 and 3 nodes 76 are connected to the r 1 node 86 and the r 3 node 90, The D 1,4 node 78 is connected to the r 1 node 86 and the r 4 node 92, the D 2,3 node 80 is connected to the r 2 node 88 and the r 3 node 90, and the D 2,4 node 82 is The r 3 node 88 and the r 4 node 92 are connected, and the D 3 and 4 node 84 is connected to the r 3 node 90 and the r 4 node 92. Each node of the second factor node processing unit 18 updates the average value and the variance value of the distance in the TOA by executing the above-described equations (2) to (5).
  As described above, the control unit 30 controls the repetition process. The control unit 30 terminates the iterative process and outputs the average value to the output unit when the average value and the variance value derived by the equations (24), (25), (28), and (29) converge. 32. Here, convergence is determined when the amount of change between the derived average value and variance value and the already derived average value and variance value is smaller than the threshold value. The determination of convergence may be made based on either the average value or the variance value.
  The output unit 32 receives the average value from the second variable node processing unit 28 after the processing repeatedly executed by the control unit 30 and outputs this as the position coordinates of the wireless device.
  This configuration can be realized in terms of hardware by a CPU, memory, or other LSI of any computer, and in terms of software, it can be realized by a program loaded in the memory, but here it is realized by their cooperation. Draw functional blocks. Accordingly, those skilled in the art will understand that these functional blocks can be realized in various forms by hardware only, software only, or a combination thereof.
The operation of the estimation apparatus 100 having the above configuration will be described. FIG. 2 is a flowchart illustrating the procedure of the estimation process performed by the estimation apparatus 100. The external variable node processing unit 16 processes the expression (1) (S10). The second factor node processing unit 18 processes Expressions (2) to (5) (S12). From the second factor node processing unit 18 to the second variable node processing unit 28 execute the downlink processing (S14). If it has not converged (N of S16), the second variable node processing unit 28 to the second factor node processing unit 18 execute the upward processing (S18). Return to step 14. On the other hand, if the convergence (S16 of Y), the output unit 32 outputs the m x and m y as the final position coordinates (x, y).
  FIG. 3 is a flowchart showing the procedure of the downstream process. The first variable node processing unit 20 executes Expression (6) and Expression (7) (S40). When Expression (8) is satisfied (Y in S42), the zero value avoidance unit 34 executes Expression (9) (S44). If the expression (8) is not satisfied (N in S42), the step 44 is skipped. The third factor node processing unit 22 executes equations (10) to (13) (S46). The internal variable node processing unit 24 executes expressions (14) to (17) (S48). The fourth factor node processing unit 26 executes Expressions (18) to (21) (S50). The second variable node processing unit 28 executes Expressions (22) to (29) (S52).
FIG. 4 is a flowchart illustrating the procedure of the upward processing. The A 1 node 134, the A 2 node 136, the A 3 node 138, and the A 4 node 140 execute Expressions (30) and (31) (S70). The Δx 1 node 118, the Δx 2 node 120, the Δx 3 node 122, and the Δx 4 node 124 execute Expression (34) (S72) and execute Expression (35) (S74). When Expression (38) is satisfied (Y in S76), the imaginary value avoiding unit 36 executes Expression (39) (S78). The B 1 node 142, the B 2 node 144, the B 3 node 146, and the B 4 node 148 execute Expressions (32) and (33) (S80). The Δy 1 node 126, the Δy 2 node 128, the Δy 3 node 130, and the Δy 4 node 132 execute Expression (36) (S82) and execute Expression (37) (S84). When Expression (40) is satisfied (Y in S86), the imaginary value avoiding unit 36 executes Expression (41) (S88).
  When Expression (38) is not satisfied (N in S76) or when Step 78 is completed, the third factor node processing unit 22 executes Expression (42) (S90). Further, when the expression (40) is not satisfied (N of S86) or when the step 88 is completed, the third factor node processing unit 22 executes the expression (42) (S90). When step 74 ends or when step 78 ends, the third factor node processing unit 22 executes Expression (43) (S92). When Step 84 ends or Step 88 ends, the third factor node processing unit 22 executes Expression (43) (S92). When Expression (44) is satisfied (Y in S94), the infinity avoidance unit 38 executes Expression (45) (S96). If Expression (44) is not satisfied (N in S94), Step 96 is skipped. The first variable node processing unit 20 executes Expression (46) and Expression (47) (S98). Following this, the second factor node processing unit 18 updates the average value and the variance value.
  According to the embodiment of the present invention, the conversion from the average value and the dispersion value of the distance in TDOA to the average value and the dispersion value of the distance in TOA is performed. Therefore, even if the measurement is performed by TDOA, the TOA The factor graph processing can be executed. Moreover, since the factor graph process by TOA is performed, the position estimation accuracy can be improved. In addition, since the measurement is performed by TDOA, timing synchronization with the target wireless device can be eliminated. Further, if the average value of distances in TOA is smaller than a predetermined value, the average value of distances in TOA is changed to a predetermined value, so that the processing result can be prevented from becoming zero. Further, if the square value of the average value of the distance in the TOA from the first variable node processing unit is smaller than the square value of the average value of the position coordinates from the sensor, the distance in the TOA from the first variable node processing unit. Since the average value of the position coordinates from the sensor is changed based on the average value, it is possible to avoid the processing result from becoming an imaginary value. Further, if the variance value of the distance in the TOA is larger than the minimum value of the variance value of the distance in the TOA from the second factor node processing unit, the variance value of the distance in the TOA is calculated in the TOA from the second factor node processing unit. Since the distance variance value is changed to the minimum value, the processing result can be prevented from becoming infinite.
  In the above, this invention was demonstrated based on the Example. This embodiment is an exemplification, and it will be understood by those skilled in the art that various modifications can be made to combinations of the respective constituent elements, and such modifications are also within the scope of the present invention.
  10 sensors, 12 acquisition units, 14 first factor node processing units, 16 external variable node processing units, 18 second factor node processing units, 20 first variable node processing units, 22 third factor node processing units, 24 internal variable nodes Processing unit, 26 fourth factor node processing unit, 28 second variable node processing unit, 30 control unit, 32 output unit, 34 zero value avoidance unit, 36 imaginary value avoidance unit, 38 infinity avoidance unit, 100 estimation device.

Claims (5)

  1. A signal from a target wireless device is received by each of the plurality of sensors, and an acquisition unit that acquires a reception time difference between different sensors;
    A first factor node processing unit for deriving an average value and a variance value of a distance in TDOA (Time Difference of Arrival) on the basis of the reception time difference acquired in the acquisition unit and assuming that a measurement value error follows a Gaussian distribution When,
    A second factor node processing unit that performs conversion from an average value and a variance value of distances in TDOA to an average value and a variance value of distances in TOA (Time Of Arrival);
    A first variable node processing unit that executes a sum product algorithm on the average value and variance value of the distance in TOA;
    A third factor node processing unit that performs conversion between the average value and the variance value of the distance in the TOA, and the average value and the variance value of the position coordinates from the sensor;
    A fourth factor node processing unit that performs conversion between the average value and variance value of the position coordinates from the sensor and the average value and variance value of the position coordinates of the wireless device;
    A second variable node processing unit that executes a sum product algorithm on the average value and the variance value of the position coordinates of the wireless device;
    Processing performed in the order of the second factor node processing unit, the first variable node processing unit, the third factor node processing unit, the fourth factor node processing unit, and the second variable node processing unit; Is a control unit that repeatedly executes processes performed in the reverse order;
    An output unit that outputs an average value of the position coordinates of the wireless device after completion of the process repeatedly executed by the control unit;
    An estimation apparatus comprising:
  2.   When the average value of the distance in TOA and the variance value are passed from the first variable node processing unit to the third factor node processing unit, if the average value of the distance in TOA is smaller than a predetermined value, the distance in TOA The estimation apparatus according to claim 1, further comprising a zero value avoiding unit that changes the average value to a predetermined value.
  3.   When the average value of the position coordinates from the sensor is passed from the fourth factor node processing section to the third factor node processing section, the first variable node is more than the square value of the average position coordinates from the sensor. If the square value of the average distance in the TOA from the processing unit is small, the average value of the position coordinates from the sensor is changed based on the average value in the TOA from the first variable node processing unit. The estimation apparatus according to claim 1, further comprising a numerical value avoiding unit.
  4.   When the variance value of the distance in TOA is passed from the third factor node processing unit to the first variable node processing unit, the variance value of the distance in TOA is the variance of the distance in TOA from the second factor node processing unit. If it is larger than the minimum value, it further comprises an infinity avoiding unit that changes the dispersion value of the distance in the TOA to the minimum value of the dispersion value of the distance in the TOA from the second factor node processing unit. The estimation apparatus in any one of Claim 1 to 3.
  5. An acquisition step in which a signal from a target wireless device is received by each of a plurality of sensors, and a reception time difference between different sensors is acquired;
    A first factor node processing step for deriving an average value and a variance value of distances in TDOA (Time Difference of Arrival) while assuming that the measurement value error follows a Gaussian distribution based on the acquired reception time difference;
    A second factor node processing step for performing conversion from the average value and the variance value of the distance in TDOA to the average value and the variance value of the distance in TOA (Time Of Arrival);
    A first variable node processing step for executing a sum product algorithm on the average value and the variance value of the distance in TOA;
    A third factor node processing step for performing conversion between the average value and variance value of the distance in the TOA and the average value and variance value of the position coordinates from the sensor;
    A fourth factor node processing step for performing conversion between the average value and variance value of the position coordinates from the sensor and the average value and variance value of the position coordinates of the wireless device;
    A second variable node processing step for executing a sum product algorithm on the average value and the variance value of the position coordinates of the wireless device;
    Processing performed in the order of the second factor node processing step, the first variable node processing step, the third factor node processing step, the fourth factor node processing step, the second variable node processing step; Is a step of outputting the average value of the position coordinates of the wireless device after repeatedly performing the processing performed in the reverse order;
    An estimation method comprising:
JP2014227152A 2014-11-07 2014-11-07 Estimation method and estimation apparatus using the same Active JP6396765B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2014227152A JP6396765B2 (en) 2014-11-07 2014-11-07 Estimation method and estimation apparatus using the same

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP2014227152A JP6396765B2 (en) 2014-11-07 2014-11-07 Estimation method and estimation apparatus using the same

Publications (2)

Publication Number Publication Date
JP2016090471A JP2016090471A (en) 2016-05-23
JP6396765B2 true JP6396765B2 (en) 2018-09-26

Family

ID=56018372

Family Applications (1)

Application Number Title Priority Date Filing Date
JP2014227152A Active JP6396765B2 (en) 2014-11-07 2014-11-07 Estimation method and estimation apparatus using the same

Country Status (1)

Country Link
JP (1) JP6396765B2 (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111459009B (en) * 2020-04-21 2021-08-17 哈尔滨工业大学 Random error estimation system and estimation method for synchronization of multiple digital electronic devices

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2008115209A2 (en) * 2006-10-25 2008-09-25 Massachusetts Institute Of Technology Cooperative localization for wireless networks
CN102905363B (en) * 2012-07-20 2014-12-03 北京邮电大学 Factor graph based positioning method

Also Published As

Publication number Publication date
JP2016090471A (en) 2016-05-23

Similar Documents

Publication Publication Date Title
JP5075396B2 (en) Position estimation method and position estimation system
US9660740B2 (en) Signal strength distribution establishing method and wireless positioning system
JP6396765B2 (en) Estimation method and estimation apparatus using the same
KR101627419B1 (en) Method for estmating location of mobile node and apparatus thereof
CN105190346A (en) Techniques to improve the performance of a fixed, timing-based radio positioning network using external assistance information
KR101219913B1 (en) Location estimation method for mobile node
KR100977246B1 (en) Apparatus and method for estmating positon using forward link angle of arrival
JP4488177B2 (en) Angle measuring method and apparatus
KR101021674B1 (en) Apparatus and method for tracking direction of signal source
JP6612113B2 (en) Method and apparatus for estimating direction of arrival and intensity of radio wave
US9930462B2 (en) System and method for on-site microphone calibration
JP6286182B2 (en) Estimation method and estimation apparatus using the same
JP2009210408A (en) Wireless system and its location identification method
JP2016125854A (en) Node position estimation method, node position estimation device, physical quantity distribution detection method, and physical quantity distribution detector
JPWO2018062435A1 (en) Positioning system, positioning device and computer program
US8660815B2 (en) Methods and apparatus for filtering noise in a three-dimensional position measurement system
JP2009175096A (en) Signal source position estimation device
JP6433386B2 (en) Estimation method and estimation apparatus using the same
KR101762510B1 (en) Method for determining mobile reference node to improve the detection accuracy of the location
US9710922B2 (en) Image processing apparatus, method and medium storing a program for detecting a motion vector
US10330772B2 (en) Determining a location of a device
WO2017049914A1 (en) Terminal positioning method, apparatus, and system
JP2009281994A (en) Mobile object position estimation method and mobile object position estimation system
JP2015162188A (en) Data analysis device and method
JP2010139486A (en) Angle measurement processing device

Legal Events

Date Code Title Description
A621 Written request for application examination

Free format text: JAPANESE INTERMEDIATE CODE: A621

Effective date: 20171005

A977 Report on retrieval

Free format text: JAPANESE INTERMEDIATE CODE: A971007

Effective date: 20180724

TRDD Decision of grant or rejection written
A01 Written decision to grant a patent or to grant a registration (utility model)

Free format text: JAPANESE INTERMEDIATE CODE: A01

Effective date: 20180814

A61 First payment of annual fees (during grant procedure)

Free format text: JAPANESE INTERMEDIATE CODE: A61

Effective date: 20180830

R150 Certificate of patent or registration of utility model

Ref document number: 6396765

Country of ref document: JP

Free format text: JAPANESE INTERMEDIATE CODE: R150