JP5300797B2 - Shape estimation system, angle estimation method, and object number estimation method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To estimate the internal angle formed by line segments configuring the outer shape periphery of an object on a two-dimensional plane. <P>SOLUTION: A composite sensor node is equipped with: a plurality of sensors for detecting presence of a shape estimation object within a sensing area; and a transmission means which transmits to a server the detection result information indicative of presence of the shape estimation object detected by each of the sensors. The server is equipped with: a measurement calculating means which calculates, where the internal angle of the vertex of the shape estimation object is &alpha;, the measure of a set of the composite sensor nodes as sensing results of detecting the shape estimation object by the specific composite sensor node and not detecting by the others among the composite sensor nodes from the parameters of the composite sensor nodes and &alpha;; an expectation value calculation means which calculates the expectation value of the sensing results for internal angles &alpha;<SB>1</SB>,..., &alpha;<SB>n</SB>(n is the number of internal angles) from the measure; and a search means which searches the internal angles &alpha;<SB>1</SB>,..., &alpha;<SB>n</SB>which decrease the differences between the sensing results and the expectation value. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

  The present invention relates to a shape estimation system, an angle estimation method, and an object number estimation method for estimating an inner angle formed by a line segment that forms an outer periphery of a shape of an object on a two-dimensional plane and the number of objects.

  In recent years, attempts have been made to distribute small wireless sensors, network them, and use them for various applications (for example, see Non-Patent Documents 1 and 2). Since it is considered that a large number of sensors are used, it is desirable that the position of each sensor can be used without designing or registering. In such a situation, consider detecting an object having a two-dimensional or three-dimensional spread. Specifically, when the sensor can detect the presence or absence of the target object in the sensing area of the sensor, the sensor estimates the shape of the target object.

  In order to deal with such a problem, in Patent Document 1, the shape of the object is determined within a range in which the class of the object shape (circle or rectangle) from each sensor is known in advance without the position information of each sensor. Shape estimation is performed by estimating parameters to be characterized.

JP 2009-063554 A

I.F.Akyildiz, W.Su, Y. Sankarasubramaniam, and E. Cayirci, ASurveyon SensorNetworks, IEEE Communications Magazine, 40, 8, pp. 102-1142002. BenW.Cook, StevenLanzisera, andKristoferS.J.Pister, SoCIssuesforRFSmartDust, ProceedingsoftheIEEE, 94,6, pp.1177-1196, June2006.

  By the way, each sensor is a binary sensor and can report only whether or not there is any object in the sensing area, without designing and deploying the position of each sensor, Further, when estimating the shape of an object without specifying a sensor position using a position sensor such as GPS, the shape of the object on a two-dimensional plane, in particular, an inner angle 1 formed by a line segment constituting the outer periphery thereof. One needs to be estimated.

  The present invention has been made in view of such circumstances. When estimating the shape of an object, the shape of the object on a two-dimensional plane, in particular, one inner angle formed by a line segment constituting the outer periphery thereof. An object is to provide a shape estimation system, an angle estimation method, and an object number estimation method capable of estimating one and estimating the number of objects.

The present invention provides a shape estimation comprising: a plurality of composite sensor nodes that detect a shape estimation object; and a server that performs processing for estimating the shape of the shape estimation object based on information on a detection result of the composite sensor node. In the system, the composite sensor node is arranged to have a predetermined shape, a plurality of sensors for detecting the presence or absence of the shape estimation object in a sensing area, and a shape estimation object detected by each of the sensors Transmitting means for transmitting detection result information indicating the presence / absence of an object to the server, wherein the server receives the detection result information indicating the presence / absence of the shape estimation object transmitted from the transmitting means, and stores the detection result information therein. Detection information acquisition means for performing detection, sensor information holding means for holding parameters of the composite sensor node, and an inner angle of the vertex of the shape estimation object is α In particular, among the composite sensor nodes, a specific composite sensor node detects the shape estimation target object, and the others measure a set measure of the composite sensor nodes that is a sensing result as a parameter of the composite sensor node. Measure calculation means for calculating from α, expected value calculation means for calculating an expected value of a sensing result when the angle is an internal angle α ₁ ,..., Α _n (n is the number of internal angles), and the sensing And a search means for searching for internal angles α ₁ ,..., Α _n for reducing the difference between the result and the expected value.

  The present invention is characterized in that in the composite sensor node, the sensors are arranged at equal intervals on two parallel line segments having the same length and separated by a predetermined distance.

The present invention is characterized in that the search means searches for internal angles α ₁ ,..., Α _n that minimize a sum related to the squared sensing pattern of the difference.

  The present invention is characterized in that the measure calculation means includes operations of Expressions (3) to (32).

  The present invention is characterized in that the expected value calculation means includes operations of Expressions (1) and (2).

The present invention obtains the sum of the interior angles α ₁ ,..., Α _n obtained by the search means, and uses the formula for obtaining the sum of the interior angles of the two-dimensional figure, and the sum of the interior angles and the interior angle An object number calculating means for obtaining the number of objects from the number is further provided.

The present invention provides a shape estimation comprising: a plurality of composite sensor nodes that detect a shape estimation object; and a server that performs processing for estimating the shape of the shape estimation object based on information on a detection result of the composite sensor node. In the system, an angle estimation method for estimating an internal angle formed by a line segment constituting the outer periphery of the shape of the shape estimation object on a two-dimensional plane, wherein the composite sensor node has a predetermined shape. A plurality of sensors arranged to detect the presence / absence of the shape estimation object in a sensing area, and a transmission means for transmitting detection result information indicating the presence / absence of the shape estimation object detected by each of the sensors to the server. A detection information acquisition step in which the server receives detection result information indicating the presence / absence of the shape estimation object transmitted from the transmission unit and stores the detection result information therein A sensor information holding step for holding parameters of the composite sensor node, and a specific composite sensor node among the composite sensor nodes when the interior angle of the vertex of the shape estimation target object is α. A measure calculation step for calculating a measure of the set of composite sensor nodes, which is a sensing result that is detected by an object and not detected by others, from the parameter of the composite sensor node and α, and an internal angle α ₁ ,. , Α _n (where n is the number of interior angles), an expected value calculation step for calculating the expected value of the sensing result, and interior angles α ₁ ,..., Α _n for reducing the difference between the sensing result and the expected value. And a search step for searching for.

  The present invention is characterized in that in the composite sensor node, the sensors are arranged at equal intervals on two parallel line segments having the same length and separated by a predetermined distance.

The present invention is characterized in that the searching step searches for internal angles α ₁ ,..., Α _n that minimize a sum related to the squared sensing pattern of the difference.

The present invention obtains the sum of the interior angles α ₁ ,..., Α _n obtained by the search step, and uses the formula for obtaining the sum of the interior angles of the two-dimensional figure, and calculates the sum of the interior angles and the number of interior angles. And a target number calculation step for obtaining the number of the target objects from the above.

  According to the present invention, the line segment constituting the outer periphery of the shape of the object on the two-dimensional plane is formed only from the information on whether or not each sensor has detected the shape estimation object without the position information of each sensor. The effect that each inner angle can be estimated is obtained. Moreover, the effect that the number of shape estimation objects can be estimated from the total sum of the calculated inner angles is also obtained.

It is a block diagram which shows the structure of the shape estimation system which estimates the internal angle which the line segment which comprises the shape of the target object on a two-dimensional plane makes. It is a block diagram which shows the structure of the composite sensor node 2 shown in FIG. It is explanatory drawing which shows a calculation result. It is explanatory drawing which shows a calculation result. It is explanatory drawing which shows a calculation result. It is explanatory drawing which shows a calculation result.

  Hereinafter, a shape estimation system according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the embodiment. In this figure, reference numeral 1 denotes a center server that performs arithmetic processing for estimating each inner angle formed by a line segment that forms the outer periphery of the shape of the object on the two-dimensional plane when estimating the shape of the object. It is. The center server 1 is a computer terminal that includes a communication unit 11, a data processing unit 12, a storage unit 13, an input / output unit 14, and a power supply unit 15. The communication unit 11 performs communication processing for receiving detection result information transmitted from the composite sensor nodes 2-1, 2-2, 2-3. The data processing unit 12 performs a statistical processing calculation based on the detection result information received by the communication unit 11. The storage unit 13 stores information (parameters) used for calculation processing in the data processing unit 12 and information on detection results received by the communication unit 11. The input / output unit 14 includes, for example, an input unit that receives input of information from the user by a keyboard, mouse, touch panel, buttons, keys, and the like, and an output unit that displays on a display such as an LCD (Liquid Crystal Display). I / O device. The power supply unit 15 supplies power for driving each unit of the center server 1. Reference numerals 2-1 to 3 are composite sensor nodes scattered around the object 3 for estimating an inner angle formed by a line segment constituting the shape of the object on a two-dimensional plane. Here, only three composite sensor nodes 2 are illustrated, but actually, a large number of composite sensor nodes 2 are scattered around the object 3.

Next, the configuration of the composite sensor node 2 shown in FIG. 1 will be described with reference to FIG. Here, a composite sensor node 2 of J _c pieces of type. The compound sensor node 2 is a sensor node in which a plurality of sensors are arranged in a predetermined shape, and the type j compound sensor node 2 has two lengths L _j arranged in parallel with a distance W _j apart. On the line segment, n _j binary sensors 23-1 to n, 24-1 to n are attached at equal intervals (j = 1,..., Jc). The distance between the sensors is l _j = L _j / (n _j −1). The sensors 23-1 to 23-n and 24-1 to 24-n emit a signal 1 indicating that the object has been detected only when the object is at its own position, and a signal 0 in other cases. A large number of these types of composite sensor nodes 2 are scattered around the object 2 so as to have an average density λ _j (j = 1,..., Jc) at random. Each composite sensor node 2 includes a CPU 21 that controls the overall processing operation of the composite sensor node 2 and a wireless communication device 22. The CPU 21 aggregates the sensing results (signal 1 or signal 0) from the sensors 23-1 to n and 24-1 to n, establishes a communication line with the center server 1 via the communication device 22, Information on the sensing result of each sensor is transmitted. The composite sensor node 2 includes a power supply unit that supplies power for operating the sensors 23-1 to n, 24-1 to n, the CPU 21, and the communication device 22, but is illustrated in FIG. Omitted. The communication device 22 may be provided in each of the sensors 23-1 to n and 24-1 to n, and each sensor may transmit a sensing result to the center server 1.

Next, the operation of the shape estimation system shown in FIG. 1 will be described with reference to FIGS. First, the i-th composite sensor node 2 of type j (for example, composite sensor node 2-1) has 2n _j pieces of 0, 1 information (when each sensor in the composite sensor node detects an object 1 If not detected, information consisting of 0) I _{i, j} is transmitted to the center server 1 through the communication device 22 together with its own type number j. The communication unit 11 receives information transmitted from the composite sensor node 2 and stores it in the storage unit 13. In addition, the storage unit 13 stores parameters (λ _j , L _j , w _j , n _j , j = 1,..., Jc) of each type of composite sensor node 2 in advance. The data processing unit 12 executes the following statistical processing when the predetermined data I _{i, j} can be collected. Here, in order to simplify subscripts, subscripts belonging to types are omitted.

First, for the various i ₁ , i ₂ , j ₁ , j ₂ satisfying 1 <i ₁ ≦ j ₁ <n _j , 1 <i ₂ ≦ j ₂ <n _j , ₁ -1) th 0, the _{j 1} th from the _{first i} th 1, the _(j 1 +1) the _{n j} th sensor from No. 0, the _(i 2 -1) th from the first sensor 0, i _2nd to j _2nd is 1 and (j ₂ +1) th to n _jth sensors are 0. The number of composite sensor nodes 2 is counted from 0,1 information I _{i, j} . .

観測値を

とする。 Its value (random variable)

Observations

And

Similarly, the (i ₁ -1) th from the first sensor is 1, the i _1st to jth _1st is 0, the (j ₁ +1) th to the _njth sensor is 1, 1st The number of composite sensor nodes in which the (i ₂ −1) th sensor is 1, the i _2nd to j _2nd sensors are 0, and the (j ₂ +1) th to the n _jth sensor is 1 is 0. 1 information I _{i, j is} counted.

とする。ここで

を、角度αをもつ対象物３の頂点に対して、第１番目センサから第（ｉ_１−１）番目が０、第ｉ_１番目から第ｊ_１番目が１、第（ｊ_１＋１）番から第ｎ_ｊ番目センサが０、第１番目センサから第（ｉ_２−１）番目が０、第ｉ_２番目から第ｊ_２番目が１、第（ｊ_２＋１）番から第ｎ_ｊ番目センサが０となるタイプｊの複合センサノード２の集合の測度と定義すると、

であることが積分幾何学の計算により分かる。ただし、ｎ_ｖ，ｎ’_ｖは、対象物の鋭角、鈍角の頂点数である。さらに、同計算により、以下が分かる。添え字簡略化のため、ｌ_ｊ，ｎ_ｊ，ｗ_ｊをｌ，ｎ，ｗと書く。 Its value (random variable)

And here

To the vertex of the object 3 having the angle α, the (i ₁ −1) th from the first sensor is 0, the i _1st to j _1st is 1, and the (j ₁ +1) th the _{n j} th sensor from 0, from the first sensor first _(i 2 -1) th 0, the _{j 2} th from the _{i 2} th 1, the _{n j} th sensor from the _(j 2 +1) th Is defined as a measure of a set of composite sensor nodes 2 of type j for which

It can be seen from the calculation of integral geometry. Here, n _v and n ′ _v are the number of vertices of acute and obtuse angles of the object. Furthermore, the following is understood from the calculation. In order to simplify subscripts, l _j , n _j , and w _j are written as l, n, and w.

とを用いて

と書ける。 Therefore, the error term

And with

Can be written.

Ε (i ₁ , j ₁ , i ₂ , j ₂ | l _j , n _j , w _j ) for various (i ₁ , j ₁ , i ₂ , j ₂ , l _j , n _j , w _j ) An error vector v _{N, m} is defined.

同様にして、鈍角と鈍角数の推定値も得られる。 Then, by solving the following least square error problem, an estimated value of each acute angle and the number of acute angles of the object can be obtained.

Similarly, the obtuse angle and the estimated number of obtuse angles are also obtained.

は、α_＊（ｉ_１，ｊ_１，ｉ_２，ｊ_２｜ｌ，ｎ，ｗ）で最大値をとり、その値は、ｍ_＊（α_＊，α，ｉ_１，ｊ_１，ｉ_２，ｊ_２｜ｌ，ｎ，ｗ）であるとする。 There are many algorithms for solving the least square error problem, but a simple example is shown here for reference.
<Initial value estimation processing>

Takes a maximum value at α _* (i ₁ , j ₁ , i ₂ , j ₂ | l, n, w), which is m _* (α _* , α, i ₁ , j ₁ , i ₂ , j ₂ | l, n, w).

となる。また、Ｓを利用可能なすべての（ｉ_１，ｊ_１，ｉ_２，ｊ_２，ｌ，ｎ，ｗ）の組み合わせとする。 (1) Initialization i = 1, n _e = 1, for each (i ₁ , j ₁ , i ₂ , j ₂ | l, n, w),

It becomes. Also, let S be a combination of all available (i ₁ , j ₁ , i ₂ , j ₂ , l, n, w).

が最大になる（ｉ_１，ｊ_１，ｉ_２，ｊ_２，ｌ，ｎ，ｗ）を見つけて、その（ｉ_１，ｊ_１，ｉ_２，ｊ_２）を〜θ_１（ｉ）（θの前の「〜」は、θの頭に付く）、（ｌ，ｎ，ｗ）を〜θ_２（ｉ）（θの前の「〜」は、θの頭に付く）と表記する。 (2) Next, for (i ₁ , j ₁ , i ₂ , j ₂ , l, n, w) εS, its normalized peak

(I ₁ , j ₁ , i ₂ , j ₂ , l, n, w) is found, and (i ₁ , j ₁ , i ₂ , j ₂ ) is changed to ˜θ ₁ (i) (θ "~" In front of is attached to the head of [theta], and (l, n, w) is expressed as ~ [theta] ₂ (i) ("~" in front of [theta] is attached to the head of [theta]).

を計算し、予め定めたγ＜１と比較し、γ以下であれば（４）の処理へ進む。一方、γより大きく、かつ、ｋ−１＋γ≦β＜ｋ＋γであれば、初期推定値として、

を得る。そして、パラメータの更新を以下のようにする。ｎ_ｅ←ｎ_ｅ＋ｋ，各（ｉ_１，ｊ_１，ｉ_２，ｊ_２｜ｌ，ｎ，ｗ）（←は値の代入を表す）について、

とする。 (3) Next,

Is compared with a predetermined γ <1, and if it is equal to or less than γ, the process proceeds to (4). On the other hand, if it is larger than γ and k−1 + γ ≦ β <k + γ, as an initial estimated value,

Get. Then, the parameter is updated as follows. For n _e ← _ne + k, each (i ₁ , j ₁ , i ₂ , j ₂ | l, n, w) (← represents substitution of values),

And

により限定し、ｉ←ｉ＋１とする。仮にｎ_ｖが既知でｎ_ｅ＞ｎ_ｖなら処理を終了させ、ｎ_ｖが未知ならＳ＝０の場合終了し、ｎ_ｖの推定値はｎ_ｅ−１となる。終了でなければ（２）の処理へ戻って処理を繰り返す。この初期値推定は、特に、ｎ_ｖが未知のときに重要である。既知の場合は、適当な初期値を用いて、以下に説明する最小二乗推定値探索を行ってもよい。 (4) Search range

And i ← i + 1. If n _v is known and n _e > n _{v, the} process is terminated. If n _v is unknown, the process is terminated when S = 0, and the estimated value of n _v is n _e −1. If not completed, the process returns to the process (2) and the process is repeated. This initial value estimation is particularly important when _nv is unknown. If known, a least square estimation value search described below may be performed using an appropriate initial value.

を初期値設定する。 <Least-square estimated value search processing>
(5) First, for each k,

Is set to the initial value.

とする。 (6) Next, for each k,

And

として、ｘ_１，・・・，ｘ_ｎｖの全組み合わせに対して、

を計算する。ここでε（ｘ_１，・・・，ｘ_ｎｖ，ｉ_１，ｊ_１，ｉ_２，ｊ_２，ｌ，ｎ，ｗ）は、

となる。 (7) Next,

For all combinations of x ₁ ,..., X _nv ,

Calculate Where ε (x ₁ ,..., X _nv , i ₁ , j ₁ , i ₂ , j ₂ , l, n, w) is

It becomes.

を最小化する組み合わせ

を選ぶ。 (8) Next,

Combinations to minimize

Select.

を得る。さらに推定値を更新するため、δ＞０を小さくして、（６）へ戻り処理を繰り返す。 (9) Next, for each k, as an estimated value

Get. Further, in order to update the estimated value, δ> 0 is reduced, and the process returns to (6) and is repeated.

を図３に示す。実際の計算では、

を用いて計算した。図３は、ｎ＝５にして、それ以外のパラメータはいろいろな値に変えて計算した結果を示している。図３から分かるように、パラメータを適当に選ぶことにより、特定の角度に対して、ｍ（α，ｉ_１，ｊ_１，ｉ_２，ｊ_２，｜ｌ，ｎ，ｗ）が高い値をとるようになる。 Next, the effects of the angle estimation method described above will be described with reference to FIGS. Normalized what kind of characteristic m (α, i ₁ , j ₁ , i ₂ , j ₂ , | l, n, w) described above shows

Is shown in FIG. In the actual calculation,

Calculated using FIG. 3 shows the calculation results when n = 5 and other parameters are changed to various values. As can be seen from FIG. 3, m (α, i ₁ , j ₁ , i ₂ , j ₂ , | l, n, w) takes a high value for a specific angle by appropriately selecting parameters. It becomes like this.

は、ｍ（α，ｉ_１，ｊ_１，ｉ_２，ｊ_２，｜ｌ，ｎ，ｗ）と、式（１）、（３）と関連するので、パラメータを適当に選ぶことにより、特定の角度に対して、観測値

が高い値をとることができるようになる。すなわち、複合センサノード２は、特定の角度に対するフィルタのような動作をすると期待される。 Observed value

Is related to m (α, i ₁ , j ₁ , i ₂ , j ₂ , | l, n, w) and Equations (1) and (3). Observed value against angle

Can take a high value. That is, the composite sensor node 2 is expected to operate like a filter for a specific angle.

  Further, the composite sensor node 2 of only λ = 1, 1 type is used, l = 2.5, n = 5, w = 10, and the object 3 has an internal angle of π / 2, π / 3, and π / 6. When the longest side is a triangle of 80, the simulation is performed 10,000 times and the result compared with the expected value given by the equation (1) is shown in FIGS. As shown in FIG. 4 and FIG. 5, it can be seen that both the calculation result by simulation and the expected value are in good agreement.

FIG. 6 shows the result of estimating the interior angles of this triangle, with the object having an interior angle of π / 2, π / 3, and π / 6 and the longest side being a triangle with 80. FIG. 6 shows how the estimated value is improved in each step of the least square search process using the initial estimation process and the least square search process described above. The set of parameters used here is (i ₁ , j ₁ , i ₂ , j ₂ , l, n, w) = (2, 4, 2, 2, 2.5, 5, 10), (2, 2,2,2,2.5,5,10), (2,4,2,2,2,5,5,2), (2,4,3,3,2.5,5,2) It is. As shown in FIG. 6, it can be seen that the estimated value rapidly converges to the true value.

とし、観測値を

とする。 Thus, based on the output of the composite sensor node 2 of J _c-number of types, and to estimate the respective internal angles of the object. A compound sensor is a sensor node in which a plurality of sensors are arranged in a predetermined shape, and a type j compound sensor node is a line segment of two lengths L _j arranged in parallel at an interval W _j. In addition, n _j binary sensors are attached at equal intervals (j = 1,..., J _c ). Here, l _j = L _j / (n _j −1). The (i ₁ -1) th from the first sensor is 0, the i _1st to jth _1st is 1, the (j ₁ +1) th to the _njth sensor is 0, the 1st sensor to the 1st Let I _{i, j be the} number of combined sensors in which (i ₂ −1) -th is 0, i _2nd to j _2nd is 1, and (j ₂ +1) th to n-th _j- th sensor is 0. Its value (random variable)

And the observed value is

And

を、角度αをもつ対象物の頂点に対して、第１番目センサから第（ｉ_１−１）番目が０、第ｉ_１番目から第ｊ_１番目が１、第（ｊ_１＋１）番から第ｎ_ｊ番目センサが０、第１番目センサから第（ｉ_２−１）番目が０、第ｉ_２番目から第ｊ_２番目が１、第（ｊ_２＋１）番から第ｎ_ｊ番目センサが０となる複合センサの集合の測度と定義すると、前述した式（１）となる。ここで、λ_ｊは、タイプｊの複合センサノードの密度、ｎ_ｖは、鋭角の頂点数である。鈍角についても同様の式を得ることができる。左辺側を測定値で置き換える。右辺は、諸々のパラメータを含む陽な式であるので、誤差項

を用いて、最小二乗誤差ｍｉｎΣ_ｊε（ｉ_１，ｊ_１，ｉ_２，ｊ_２｜ｌ_ｊ，ｎ_ｊ，ｗ_ｊ）^２を達成するα_１，・・・，α_ｎｖを得ることができる。 here

With respect to the vertex of the object having an angle α, the (i ₁ −1) th from the first sensor is 0, the i _1st to j _1st is 1, and the (j ₁ +1) th the _{n j} th sensor is 0, from the first sensor first _(i 2 -1) th 0, the _{j 2} th from the _{i 2} th 1, the second _{n j} th sensor from the _(j 2 +1) th When defined as a measure of a set of composite sensors that becomes 0, the above-described equation (1) is obtained. Here, λ _j is the density of type j composite sensor nodes, and n _v is the number of acute vertices. Similar equations can be obtained for obtuse angles. Replace the left side with the measured value. Since the right side is an explicit expression that includes various parameters, the error term

Can be used to obtain α ₁ ,..., Α _nv that achieves the least square error minΣ _j ε (i ₁ , j ₁ , i ₂ , j ₂ | l _j , n _j , w _j ) ^2. .

は、加法的である。すなわち、ｎ_ｖを複数の対象物全体にわたる鋭角の頂点数として、式（１）が成立する。従って、複数の対象物全体の鋭角の内角α_１，・・・，α_ｎｖについての最小二乗誤差推定の問題となる。複合センサノード２のタイプ数が十分大きければ、これらの推定値を得ることは、１つの対象物の場合と同じである。鈍角についても同様に得ることができる。結果として、全鋭角の内角の総和の推定値Ａ、全鈍角の内角の総和の推定値Ｂを得ることができる。 Next, a processing operation for estimating the number of objects from the estimated interior angle will be described. Unless one compound sensor node 2 detects a plurality of objects at a time,

Is additive. That is, equation (1) is established, where n _v is the number of acute vertices over a plurality of objects. Therefore, there is a problem of least square error estimation for acute inner angles α ₁ ,..., Α _nv of a plurality of objects. If the number of types of the composite sensor node 2 is sufficiently large, obtaining these estimated values is the same as in the case of one object. The obtuse angle can be obtained similarly. As a result, it is possible to obtain an estimated value A of the sum total of all acute angles and an estimated value B of the sum of all obtuse angles.

である。ここで、Ｊは、全対象物の鋭角、鈍角を合わせた頂点の数である。一般にｋ_ｉ個の頂点をもつ２次元図形の内角の和は（ｋ_ｉ−２）πである。よって、Ａ，Ｂが正しいとすると、

となる。したがって、Ｋ＝（Ｊπ−（Ａ＋Ｂ））／２πにより対象物数を求めることができ、推定した各内角の値から対象物の数を推定することが可能となる。 The object number K, the number of vertices of the i-th object (the acute angle, the combined obtuse) and k _i. From the definition

It is. Here, J is the number of vertices obtained by combining the acute and obtuse angles of all objects. In general, the sum of the inner angles of a two-dimensional figure having k _i vertices is (k _i −2) π. Therefore, if A and B are correct,

It becomes. Therefore, the number of objects can be obtained by K = (Jπ− (A + B)) / 2π, and the number of objects can be estimated from the estimated values of the respective internal angles.

  As described above, the line segment constituting the outer periphery of the shape of the object on the two-dimensional plane is formed only from the information on whether or not each sensor has detected the shape estimation object without the position information of each sensor. It becomes possible to estimate the interior angles one by one and also the number of objects.

  It should be noted that the program for realizing the functions of the center server 1 shown in FIG. 1 is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read into the computer system and executed to estimate the shape. Processing and object quantity estimation processing may be performed. Here, the “computer system” includes an OS and hardware such as peripheral devices. The “computer system” includes a WWW system having a homepage providing environment (or display environment). The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Further, the “computer-readable recording medium” refers to a volatile memory (RAM) in a computer system that becomes a server or a client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line. In addition, those holding programs for a certain period of time are also included.

  The program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, what is called a difference file (difference program) may be sufficient.

  Applications where it is indispensable to disperse sensors with small wireless communication devices, network them, and detect detection objects with a two-dimensional or three-dimensional extent without designing or registering the position of each sensor Applicable to.

  DESCRIPTION OF SYMBOLS 1 ... Center server, 11 ... Communication part, 12 ... Data processing part, 13 ... Memory | storage part, 14 ... Input-output part, 15 ... Power supply part, 2 ... Composite sensor Node, 21 ... CPU, 22 ... Communicator, 23, 24 ... Sensor, 3 ... Object

Claims (10)

In a shape estimation system comprising: a plurality of composite sensor nodes for detecting a shape estimation object; and a server that performs processing for estimating the shape of the shape estimation object based on information on a detection result of the composite sensor node;
The composite sensor node is:
A plurality of sensors that are arranged to form a predetermined shape and detect the presence or absence of the shape estimation object in a sensing area;
Transmission means for transmitting detection result information indicating the presence or absence of a shape estimation target detected by each of the sensors to the server,
The server
Detection information acquisition means for receiving detection result information indicating the presence or absence of the shape estimation object transmitted from the transmission means, and storing the detection result information therein;
Sensor information holding means for holding parameters of the composite sensor node;
When the inner angle of the apex of the shape estimation target object is α, among the composite sensor nodes, a specific composite sensor node detects the shape estimation target object, and the other sensor sensor results in a sensing result not detected by others. A measure calculation means for calculating a measure of the set from the parameter of the composite sensor node and α;
Expected value calculation means for calculating an expected value of the sensing result when the angle is an internal angle α ₁ ,..., Α _n (n is the number of internal angles) from the measure;
A shape estimation system comprising: search means for searching for internal angles α ₁ ,..., Α _n that reduce a difference between the sensing result and the expected value.   The shape estimation system according to claim 1, wherein the composite sensor node is configured by disposing sensors at equal intervals on two parallel line segments having the same length and spaced apart by a predetermined distance. 3. The shape estimation system according to claim 1, wherein the search unit searches for an internal angle α ₁ ,..., Α _n that minimizes a sum related to a square sensing pattern of the difference. The measure calculation means is a formula

The shape estimation system according to claim 1, further comprising: The expected value calculation means is an expression

The shape estimation system according to claim 1, further comprising: The sum of the interior angles α ₁ ,..., Α _n obtained by the search means is obtained, and the sum of the interior angles and the number of interior angles are calculated using an equation for obtaining the sum of interior angles of a two-dimensional figure. The shape estimation system according to claim 1, further comprising an object number calculating means for calculating the number of objects. 2. A shape estimation system comprising: a plurality of composite sensor nodes that detect a shape estimation object; and a server that performs processing for estimating the shape of the shape estimation object based on information on a detection result of the composite sensor node. An angle estimation method for estimating an internal angle formed by a line segment constituting the outer periphery of the shape of the shape estimation object on a three-dimensional plane,
The composite sensor node is:
A plurality of sensors that are arranged to form a predetermined shape and detect the presence or absence of the shape estimation object in a sensing area;
Transmission means for transmitting detection result information indicating the presence or absence of a shape estimation target detected by each of the sensors to the server,
The server is
A detection information acquisition step of receiving detection result information indicating the presence or absence of the shape estimation object transmitted from the transmission means, and storing the detection result information inside;
A sensor information holding step for holding parameters of the composite sensor node;
When the inner angle of the apex of the shape estimation target object is α, among the composite sensor nodes, a specific composite sensor node detects the shape estimation target object, and the other sensor sensor results in a sensing result not detected by others. A measure calculation step for calculating a measure of the set from the parameter of the composite sensor node and the α;
An expected value calculation step of calculating an expected value of the sensing result when the angle is an internal angle α ₁ ,..., Α _n (where n is the number of internal angles);
An angle estimation method comprising: a search step of searching for internal angles α ₁ ,..., Α _n that reduce a difference between the sensing result and the expected value.   The angle estimation method according to claim 7, wherein the composite sensor node has sensors arranged at equal intervals on two parallel line segments having the same length and spaced apart by a predetermined distance. 9. The angle estimation method according to claim 7, wherein the searching step searches for an internal angle α ₁ ,..., Α _n that minimizes a sum related to the squared sensing pattern of the difference. The sum of the interior angles α ₁ ,..., Α _n obtained by the angle estimation method according to claim 7 is obtained, and the sum of the interior angles and the interior angle are calculated using an equation for obtaining the sum of interior angles of a two-dimensional figure. And a target number calculating step for obtaining the number of the target objects from the number of the target objects.

Images