JP2012159307A - Three-dimensional shape deriving apparatus and three-dimensional shape deriving method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To derive, by using a sensor network, shapes of even such regions as are difficult to measure by camera or laser.SOLUTION: A three-dimensional shape deriving apparatus 100 comprises a plurality of fluid position sensors 112 that transmit positional information indicating their own three-dimensional positions or physical quantities corresponding thereto, an enveloping body (hand 92) that envelops the plurality of position sensors, a sensor information integrating unit 114 that receives the positional information transmitted by the plurality of position sensors, and a three-dimensional shape deriving unit 130 that derives the three-dimensional shape of the enveloping body on the basis of the positional information from the plurality of position sensors.

Description

  The present invention relates to a three-dimensional shape deriving apparatus and a three-dimensional shape deriving method for deriving a three-dimensional shape through a sensor network.

  In recent years, a technique for specifying the position of an arbitrary object using a sensor that is lightweight and small and capable of wireless communication has become widespread. For example, by using communication means with different radio field intensity, electromagnetic induction mode, etc., and limiting the communication with the sensor to be established only within a specific range, the sensor is attached depending on whether or not the sensor communication is established. A technique for detecting an approximate position of a device has been disclosed (for example, Patent Document 1).

  Further, a technique for specifying the position of the monitoring target by attaching a sensor capable of transmitting an identification signal to the monitoring target and specifying the position of the sensor through a sensor network such as ad hoc is disclosed (for example, Patent Document 2). ). Furthermore, a technique is also disclosed in which two sensors whose positions are known are used in the sensor network, and the positions of the arbitrary sensors are sequentially derived based on the distance between the arbitrary sensor and the two sensors whose positions are known. (For example, patent document 3).

JP 2008-59087 A JP 2007-279895 A JP 2009-250627 A

  In recent years, there has been a demand for a technique for measuring positions and shapes of various objects in a three-dimensional space. The position and shape in the three-dimensional space obtained by such measurement are, for example, autonomously recognizing a plurality of regularly placed objects, and placing the recognized objects on the hand of an industrial robot. It can be used for various purposes such as picking processing for gripping one by one. However, in the techniques described in Patent Documents 1 to 3 described above, the position of the object can be specified, but the shape of the object cannot be specified because the sensor is fixed to the object.

  Hands used for the picking process have a wide variety of shapes, and not only a simple hand having only a joint angle but also a completely free-form universal hand has been devised. In such a free-form hand, the degree of freedom of the shape is very large, and even if the same object is gripped, the shape of the hand is not necessarily the same, and it is necessary to grasp the shape of the hand. In order to measure the shape in such a three-dimensional space, a large-scale apparatus such as a plurality of cameras and laser sensors is required. However, in the picking process, a blind spot may be formed depending on the object being gripped, or a shadowed part of the laser beam may be generated. Even with such a large device, the shape of the hand can be accurately measured. I couldn't. In addition, it is impossible to measure the inside (palm side) of the hand that is grasping in the first place.

  Further, in the conventional technique, since the shape and pressure of the surface of the hand in contact with the object cannot be grasped, there is no way of knowing how firmly the object is grasped.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide a three-dimensional shape deriving apparatus and a three-dimensional shape deriving method capable of deriving even the shape of a part difficult to measure with a camera or laser using a sensor network. .

  In order to solve the above-described problems, a three-dimensional shape deriving device according to the present invention includes a plurality of position sensors that transmit position information indicating its own three-dimensional position or a physical quantity corresponding thereto, and a container that includes the plurality of position sensors. And a sensor information aggregating unit that receives position information transmitted from the plurality of position sensors, and a three-dimensional shape deriving unit that derives the three-dimensional shape of the container based on the position information of the plurality of position sensors.

  A plurality of signal supply sources arranged at different three-dimensional positions are further provided, and the plurality of position sensors calculate a distance from the signal supply source or a direction of the signal supply source based on a signal supplied from the signal supply source. May derive its own three-dimensional position.

  The container may be provided with a deformation mechanism.

  The inclusion body may hold an arbitrary object.

  The position sensor may further measure a pressure applied to the position sensor.

  A plurality of position sensors may transmit position information only when a predetermined condition is satisfied.

  The predetermined condition is to be the only representative position sensor within a predetermined space range, and the position sensor may transmit the position information only when the position sensor itself becomes the representative position sensor.

  The predetermined condition is that the position sensor is located at the boundary of the inclusion body, and the position sensor determines whether or not the position sensor itself is located at the boundary of the inclusion body based on the three-dimensional positions of a plurality of other position sensors existing in the vicinity. The position information may be transmitted only when it is determined and located at the boundary of the container.

  The position sensor may determine that the position sensor itself is located at the boundary of the enclosure if the number of other position sensors existing within the predetermined range is equal to or less than a predetermined threshold value based on the density of the position sensors.

  A position sensor is located within a specified steradian and within a specified range with a connection between the center of gravity of the position of the center of gravity of a plurality of other position sensors and the position sensor as the center point. , It may be determined that itself is located at the boundary of the inclusion body.

  In order to solve the above-described problem, a container is provided using a three-dimensional shape derivation device including a plurality of position sensors, a container that contains the plurality of position sensors, and a sensor information aggregating unit that can communicate with the plurality of position sensors. In the three-dimensional shape deriving method of the present invention for deriving a three-dimensional shape of the present invention, the plurality of position sensors transmit position information indicating their own three-dimensional position or a physical quantity corresponding thereto, and the sensor information aggregating unit includes a plurality of positions. The position information transmitted from the sensor is received, and the three-dimensional shape of the inclusion body is derived based on the position information of the plurality of position sensors.

  According to the present invention, by using a sensor network, it is possible to derive a shape of a part that is difficult to measure with a camera or a laser.

It is explanatory drawing which showed the rough connection relation of the picking system. It is a functional block diagram showing a schematic configuration of a three-dimensional shape deriving device. It is the functional block diagram which showed the schematic structure of the sensor. It is explanatory drawing for demonstrating the sensor network constructed | assembled by several sensors. It is explanatory drawing for demonstrating the predetermined conditions which transmit a positional infomation. It is explanatory drawing for demonstrating the boundary determination by a sensor. It is explanatory drawing for demonstrating the boundary determination by a sensor. It is explanatory drawing for demonstrating the boundary determination by a sensor. It is the flowchart which showed the flow of the process of the three-dimensional shape derivation method.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. The dimensions, materials, and other specific numerical values shown in the embodiments are merely examples for facilitating the understanding of the invention, and do not limit the present invention unless otherwise specified. In the present specification and drawings, elements having substantially the same function and configuration are denoted by the same reference numerals, and redundant description is omitted, and elements not directly related to the present invention are not illustrated. To do.

(Picking system 10)
FIG. 1 is an explanatory diagram showing a schematic connection relationship of the picking system 10. The picking system 10 includes an object measuring device 60, an object recognizing device 70, a robot control device 80, a robot 90, and a three-dimensional shape deriving device 100.

  The object measuring device 60 is composed of, for example, a laser scanner type three-dimensional distance sensor, and a plurality of detection points on measurement surfaces (surfaces) of a plurality of objects (workpieces) 62 that are irregularly stacked and overlapped randomly. The three-dimensional coordinates of are detected. Specifically, the object measuring device 60 passes through a three-dimensional distance sensor, a Z-axis coordinate indicating the distance from the object measuring device 60 of a plurality of detection points on the measurement surface, and a surface (XY) perpendicular to the detection direction. The X and Y axis coordinates on the plane) are combined into a three-dimensional coordinate, and information indicating the three-dimensional coordinate is transmitted to the object recognition device 70.

  The object recognition device 70 selects one object 62 based on the three-dimensional coordinates indicated in the information received from the object measurement device 60, and the object 62 is prepared in advance by pattern matching or the like. It is determined whether the object 62 is a desired object indicated by the model, and the position and orientation of the object 62 are specified. In addition, the object recognition device 70 transmits information indicating the position and orientation of one recognized object 62 to the robot control device 80, and causes the robot control device 80 to pick the object 62.

  The robot control device 80 receives information indicating the position and orientation of the object 62 from the object recognition device 70 and issues a control command for causing the robot 90 to pick the object 62. In response to the control command, the robot 90 picks the instructed one object 62 through the hand (container) 92.

  The three-dimensional shape deriving device 100 determines how the hand 92 that holds the object 62 is holding the object 62, that is, the three-dimensional shape of the hand 92 in the state where the object 62 is held, The obtained pressure or the like is derived, and information indicating the three-dimensional shape or the like is transmitted to the robot controller 80. The robot control device 80 performs control so as to appropriately hold the object 62 using the three-dimensional shape of the hand 92 indicated by the information received from the three-dimensional shape deriving device 100. When approaching the hand 92 to one of the objects 62, measure the shape so that the hand 92 does not collide with surrounding structures or other objects 62, and calculate an appropriate approach path to avoid interference. There is a need to. With such a picking system 10, for example, an operation of aligning irregularly stacked objects 62 at a predetermined position can be automated. Hereinafter, the three-dimensional shape deriving device 100 characteristic of the present embodiment will be described in detail.

(Three-dimensional shape deriving device 100)
FIG. 2 is a functional block diagram illustrating a schematic configuration of the three-dimensional shape deriving device 100, and FIG. 3 is a functional block diagram illustrating a schematic configuration of the position sensor 112. The three-dimensional shape deriving device 100 includes a plurality of signal supply sources 110, a plurality of position sensors 112, a hand 92, a sensor information aggregating unit 114, a holding unit 116, and a central control unit 118. The three-dimensional shape of the hand 92 corresponding to the container of the three-dimensional shape deriving device 100 is derived.

  The signal supply sources 110 are arranged at different three-dimensional positions, and supply signals based on an electric field, a magnetic field, or electromagnetic waves to the plurality of position sensors 112. The signal supply source 110 plays two roles of a position calculation signal transmission source to the position sensor 112 and a power supply source. However, these functions may be separated and arranged as separate solids.

  The position sensor (microsensor) 112 is formed in a particulate (lightweight, small) and free-flowing manner by microfabrication technology such as MEMS (Micro Electro Mechanical Systems) or miniaturization of a semiconductor process, and is supplied from a signal supply source 110. By calculating the distance to the signal supply source 110 or the direction of the signal supply source 110 based on the received signal, the three-dimensional position of itself is derived in a self-contained manner, and position information indicating the three-dimensional position is transmitted. The position information here is not limited to only the three-dimensional position coordinates (x, y, z coordinates), but finally the position information from the measured physical quantity corresponding to the three-dimensional position, such as the intensity and phase of the signal. It may be information that can be converted to. Here, an example is given in which the position sensor 112 is formed to be freely flowable. However, the present invention is not limited to such a case. For example, the position sensor 112 is kneaded into a resin such as plastic, so that the fluidity is reduced or lost. Is included in the technical scope of the present embodiment. Here, the position sensor 112 is assumed to be spherical and coated in consideration of friction with other position sensors 112. However, the position sensor 112 is not limited to this shape, and may be any shape that is easy to handle. . As shown in FIG. 3, the position sensor 112 includes an antenna 112a, a power supply unit 112b, a communication unit 112c, and a microcomputer 112d.

  The antenna 112a receives electromagnetic waves in all directions, and transmits a signal generated in the position sensor 112 as electromagnetic waves. As a communication method in the position sensor 112, for example, near field communication (NFC) with low power consumption is adopted. The power supply unit 112b obtains an induced electromotive force from the received electromagnetic wave, and operates the position sensor 112 with the electric power (powerless sensor). Since the power supply part 112b consumes little electric power, it can also have a small battery in itself. The communication unit 112c performs data communication with other nearby position sensors 112 and the sensor information aggregation unit 114 through the antenna 112a. The microcomputer 112d serves as a three-dimensional position deriving unit 112e that derives its own three-dimensional position based on a signal from the signal supply source 110, and a boundary determination unit 112f that determines that the self is located at the boundary of the hand 92. Function. The position sensor 112 can also measure other environmental parameters such as the pressure applied to the position sensor 112, the temperature near the position sensor 112, the radio wave intensity and phase of the electromagnetic wave received by the antenna 112a.

  The hand 92 has a deformation mechanism (not shown), and the deformation mechanism deforms the hand 92 to a position and posture where the object 62 can be gripped in response to a control command from the robot controller 80. Such a deformation mechanism may be provided inside or outside the hand 92. The hand 92, also called a universal hand, encloses a plurality of spherical particles (in the present embodiment, a plurality of position sensors 112) in a flowable manner, and is itself formed of a member having a high degree of deformation freedom such as an elastic member. Has been. In addition, a resin having a low degree of deformation or no degree of deformation, such as a resin such as plastic, is included in the technical scope of the present embodiment. Each position sensor 112 contained in the hand 92 is weakly coupled with other nearby position sensors 112, can be freely rearranged within the hand 92, and has a high degree of freedom of movement. After the hand 92 is deformed to a position and posture at which the object 62 can be gripped, the air in the hand 92 is removed, whereby the position of the particles (position sensor 112) in the hand 92 is fixed, and the object 62 is moved. It can be gripped.

  At this time, the plurality of position sensors 112 in the hand 92 measure the pressure applied to the position sensor 112 together with the three-dimensional position, and transmit it together as position information. In this way, the three-dimensional shape deriving device 100 can grasp whether or not the hand 92 is correctly grasping the object 62. In addition, since the pressure applied to the object 62 by the hand 92 can be adjusted, the object 62 can be appropriately held even when the object 62 is fragile or the shape is easily deformed. Further, the temperature MAP in the hand 92 can also be generated by measuring the temperature in the vicinity of the position sensor 112.

  The sensor information aggregating unit 114 receives position information transmitted from the position sensor 112. In addition, an electric field or electromagnetic wave may be supplied from the sensor information aggregating unit 114 to the position sensor 112 to serve as the signal supply source 110. The holding unit 116 includes a ROM, a nonvolatile RAM, a flash memory, an HDD (Hard Disk Drive), and the like, and temporarily holds position information received by the sensor information aggregation unit 114. The central control unit 118 has a three-dimensional shape by a semiconductor integrated circuit including a central processing unit (CPU), a signal processing unit (DSP: Digital Signal Processor), a ROM and memory storing programs, a RAM as a work area, and the like. The entire derivation device 100 is managed and controlled. In the present embodiment, the central control unit 118 derives a three-dimensional shape of a portion of the hand 92 where the plurality of position sensors 112 exist based on position information of the plurality of position sensors 112. Also functions as 130. Hereinafter, specific processing of the three-dimensional shape deriving device 100 will be described.

(Derivation process of three-dimensional position in position sensor 112)
For example, when the signal supply source 110 supplies a signal based on an electromagnetic wave to the position sensor 112, the three-dimensional position deriving unit 112e of the position sensor 112 grasps the distance from the signal supply source 110 through the radio wave intensity of the electromagnetic wave. be able to. In addition, a relative distance difference can be calculated by using a phase difference with another signal supply source 110. Further, if the antenna 112a of the position sensor 112 is a directional microarray antenna or the like, the direction of the signal supply source 110 can be grasped.

  Therefore, in order for the three-dimensional position deriving unit 112e of the position sensor 112 to derive its own three-dimensional position, if the position sensor 112 can grasp the distance and direction of the signal supply source 110, there is only one signal supply source 110. It is sufficient if the position sensor 112 can grasp only the direction of the signal supply source 110, and if there are only two signal supply sources 110, and the position sensor 112 can grasp only the distance from the signal supply source 110, the signal supply source 110 is sufficient. Three is enough.

  Further, when the spatial degree of freedom of the position sensor 112 in the hand 92 is limited, the position sensor 112 can derive its own three-dimensional position with fewer signal supply sources 110. In addition, when the position sensor 112 can use an existing system such as GPS (Global Positioning System), the three-dimensional position deriving unit 112e can also derive a three-dimensional position using the GPS.

  Further, when the signal supply source 110 supplies a signal based on the magnetic field to the position sensor 112, the plurality of signal supply sources 110 exclusively turn on / off the magnetic field, and at that time, which signal supply source 110 In order to inform the position sensor 112 of whether or not the signal source is turned on, an identifier for identifying the signal source 110 is transmitted. The three-dimensional position deriving unit 112e of the position sensor 112 can grasp the distance from each signal supply source 110 through the identifier and the magnetic field strength of the magnetic field. Thus, the three-dimensional position deriving unit 112e of the position sensor 112 can derive its own three-dimensional position by the three-point positioning method based on the distances from the plurality of signal supply sources 110.

(Sensor network)
FIG. 4 is an explanatory diagram for explaining a sensor network constructed by a plurality of position sensors 112. As described above, the position sensor 112 that has received the signal from the signal supply source 110 and derived its own three-dimensional position transmits position information indicating the three-dimensional position to the sensor information aggregating unit 114. At this time, a plurality of position sensors 112 in the vicinity (adjacent) form an autonomous distributed sensor network that exchanges information by wirelessly communicating with each other. As shown in FIG. 112 can finally transmit position information to the sensor information aggregating unit 114 by passing (hopping) a plurality of other position sensors 112. Therefore, the sensor information aggregating unit 114 does not receive the position information individually from all the plurality of position sensors 112, but receives the position information collected locally in the position sensor 112 group 138 near the sensor information aggregating unit 114. As a result, it is possible to acquire position information of all the plurality of position sensors 112.

  In this embodiment, a sensor network represented by Zigbee (registered trademark) standardized as IEEE 802.15.4 is adopted, and the position information of each position sensor 112 is used by using an existing ad hoc function or routine function. Build an acquisition path. The ad hoc function and the routine function can apply various existing technologies, and thus detailed description thereof is omitted here. However, in this embodiment, since each position sensor 112 can grasp the distance from the signal supply source 110 and its own three-dimensional position, the sensor information aggregating unit 114 uses the three-dimensional position information and the spanning tree protocol. It is possible to construct a route through the information from the distant position sensor 112 to the closer position sensor 112 (see FIG. 4). Since the position sensor 112 is flowing in the hand 92, it is desirable to reconstruct the information transmission path in real time. However, after the hand 92 grips the object 62, that is, the air in the hand 92 is removed and the position sensor is removed. After the arrangement of 112 is fixed, the constructed information transmission path can be maintained (network fixed).

(Regarding predetermined conditions for sending location information)
Each of the plurality of position sensors 112 has a capability of transmitting position information. However, in the present embodiment, the position information is not transmitted to all the position sensors 112 and a predetermined condition is satisfied. Only position information will be sent.

  The position sensor 112 may be hundreds, thousands, or more depending on the size and shape of the hand 92. However, if all three-dimensional positions of the hundreds or thousands of position sensors 112 are strictly acquired, the information transmission load in the sensor network and the calculation load in the central control unit 118 become enormous, and the desired result You may not get. Here, only the position sensor 112 satisfying a predetermined condition transmits the position information, thereby thinning out the number of position sensors 112 serving as the position information source, and the information transmission load in the sensor network and the calculation load in the central control unit 118. Mitigating. However, by imposing a predetermined condition on the position sensor 112, the information transmission load to the sensor information aggregating unit 114 is reduced, but communication with other position sensors 112 in the vicinity is continuously executed. Therefore, communication with other position sensors 112 in the vicinity and communication for transmitting position information to the sensor information aggregating unit 114 are performed using different protocols, for example, using headers with different identifiers or headers indicating the number of hops. Yes.

  Hereinafter, with reference to FIG. 5, as the above-described predetermined conditions, (1) a process for deriving a representative position sensor appropriately, (2) a process for deriving a representative position sensor within a predetermined space range, and (3) a position at the boundary A process for deriving a position sensor will be exemplified. However, the predetermined condition is not limited to such a case, and various conditions can be set such as when other parameters such as pressure and temperature are equal to or higher than a predetermined threshold value or lower than a predetermined threshold value.

((1) Processing for deriving a representative position sensor appropriately)
As shown in FIG. 5A, an arbitrary position sensor 112g in the hand 92 has a plurality of (for example, several tens or several hundreds) other nearby (for example, within two hops or within a predetermined delay amount). The sensor 112 is regarded as one group 140 (clustered), and its own position information is transmitted to the sensor information aggregating unit 114 as a representative value of the group 140. At this time, the arbitrary position sensor 112 that has collected the three-dimensional position becomes the representative position sensor of the other position sensor 112 located in the group 140, and therefore, the other position sensor in the group 140 to which the three-dimensional position is referenced. 112 may not need to transmit the three-dimensional position to the sensor information aggregation unit 114. Here, the number of hops, the predetermined delay amount, and the number of sensors to be averaged for the three-dimensional positions are determined as appropriate according to the accuracy and resolution required for the three-dimensional shape deriving device 100. .

((2) Processing for deriving a representative position sensor within a predetermined space range)
As another example, it is conceivable to derive the representative position sensor 112 within a predetermined space range. This is similar to the above-described (1) process of deriving a representative position sensor appropriately in that the representative position sensor 112 is derived, but differs in that only one representative position sensor is derived within a predetermined range.

  Here, as shown in FIG. 5B, the three-dimensional space is divided into a plurality of predetermined space ranges 142 in a matrix (the predetermined space range 142 is represented by the length of each of the x, y, and z coordinates. be able to.). Then, the arbitrary position sensor 112g in the hand 92 regards a plurality of (for example, several tens and several hundreds) other sensors 112 in the predetermined space range 142 as one group, and the position information of itself is set in the predetermined space. It transmits to the sensor information aggregating unit 114 as a representative value of the range 142. At this time, the arbitrary position sensor 112 that has collected the three-dimensional position becomes the representative position sensor of the other position sensor 112 located in the predetermined space range 142, and therefore, the three-dimensional position in the predetermined space range 142 to which the three-dimensional position is referenced. The other position sensor 112 may not need to transmit the three-dimensional position to the sensor information aggregation unit 114.

  Further, in the three-dimensional shape deriving device 100, as means for specifying the predetermined space range 142, for example, means by broadcast or means by MAP are used. In the broadcast means, the sensor information aggregating unit 114 transmits information asking whether or not the position sensor 112 exists in the predetermined space range 142 to each predetermined space range 142, and the position sensor 112 is designated. If it exists in the predetermined space range 142, the fact is transmitted to the sensor information aggregating unit 114 to that effect. The three-dimensional shape deriving unit 130 determines that the hand 92 occupies the predetermined space range 142 if a predetermined number or more of the position sensors 112 are present in one predetermined space range 142. When the position measurement accuracy of each position sensor 112 is high, 1 can be included as the predetermined number. The three-dimensional shape deriving unit 130 determines that the outer peripheral portion (contour portion) of the predetermined space range 142 where the predetermined number or more of the position sensors 112 are present is the outer shape of the hand 92.

  In the MAP means, each position sensor 112 has a MAP capable of specifying a plurality of predetermined space ranges 142, and it is determined whether or not a flag is set in the predetermined space range 142 where the position sensor 112 is located. If is not set, a flag is set to indicate that the predetermined space range 142 exists. Then, the position sensor 112 transmits the MAP for which the flag determination process has been completed to the nearby position sensor 112. The transmitted position sensor 112 updates its MAP. Thus, in each predetermined space range 142 of the MAP that has passed through the plurality of position sensors 112, there are places where a flag is set and places where no flag is set. The three-dimensional shape deriving unit 130 refers to the MAP and determines that the hand 92 occupies the predetermined space range 142 if the flag is set. If the flag is not set, the three-dimensional shape deriving unit 130 determines that predetermined space. The range 142 is determined to be a space outside the hand 92. The three-dimensional shape deriving unit 130 determines that the outer peripheral portion (contour portion) of the predetermined space range 142 for which the flag is set is the outer shape of the hand 92.

((3) Process for deriving a position sensor located at the boundary)
Further, as shown in FIG. 5C, the plurality of position sensors 112g are positioned at the boundaries of the inclusion wall 144 of the hand 92 based on the three-dimensional positions of the other plurality of position sensors 112 existing in the vicinity. The position information can be transmitted only when it is determined whether or not it is positioned and located at the boundary of the inclusion wall 144 of the hand 92. Hereinafter, the boundary determination of the position sensor 112 will be described in detail.

(Judgment process for boundary)
6, 7, and 8 are explanatory diagrams for explaining boundary determination by the position sensor 112. As described above, the position sensor 112 is formed in a particle shape and moves in a freely flowing manner in a state of being packed in the hand 92. Although the size of the position sensor 112 is arbitrary, the smaller the position sensor 112, the higher the measurement accuracy. Here, a position sensor 112 having a size of about several millimeters is assumed. At this time, not only the position sensor 112 is packed in the hand 92 but also a fluid such as sand may be arranged (mixed) side by side. The position sensor 112 does not enter the containing wall 144 of the hand 92 shown in FIG. 6A, but the other gaps 154 are equally filled.

  Focusing only on the position sensor 112 in the hand 92, as shown in FIG. 6 (b), the position sensor 112 does not exist only in a portion corresponding to the containing wall 144, so that it touches the containing wall 144 in the hand 92. The position sensor 112 that is in contact cannot detect the presence of another position sensor 112 in the direction of the enclosure wall 144. In this embodiment, this event is used to determine whether or not the position sensor 112 is located at the boundary.

  As a boundary determination process, there is a method of calculating the number of position sensors 112 in the vicinity of a predetermined range (for example, within 2 hops) when the calculation load is small. For example, as shown in FIG. 7A, when the position sensors 112 are evenly arranged in the hand 92, the position sensor 112h located in the gap 154 is within the vicinity range 160, for example, 20 other position sensors. 112 can be detected. However, as shown in FIG. 7B, the position sensor 112i located at the boundary of the enclosure wall 144 can be detected by half, and only 12 other position sensors 112 can be detected in the vicinity range 160, for example. It will be. As described above, if the number of position sensors 112 is less than or equal to a predetermined threshold based on the density of the position sensors 112 depending on the number of other position sensors 112 that can be detected, the position sensor 112 itself is the boundary of the enclosure wall 144 of the hand 92. It is possible to easily grasp that it is located.

  As another determination process of the boundary, as illustrated in FIG. 8, the position sensor 112 i includes a point symmetry 174 of the centroid positions 172 of the plurality of other position sensors 112 within the predetermined range 160 with the position sensor 112 i as the symmetry point 170. , When there is no other position sensor 112 within a predetermined steradian with a connection line 176 connecting to the position sensor 112 as a central axis, for example, (2-√2) π steradian and within a predetermined range 160, the self-contained wall of the hand 92 It is determined that it is located at the boundary of 144.

  This is based on the fact that when the position sensor 112 is located at the boundary of the enclosure wall 144 of the hand 92, the location where the position sensor 112 is concentrated and the location where the position sensor 112 is not located are facing each other. The absence of the position sensor 112 at the point symmetry 174 of the center of gravity position 172 where the position sensor 112 is concentrated determines that the position sensor 112 is located at the boundary. Here, (2-√2) π steradian indicates a solid angle of a cone having an apex angle of 90 degrees. Further, the angle is not limited to 90 degrees, and the angle can be appropriately changed according to the shape of the obstacle 150 in the hand 92.

  In addition, as described above, when only the position sensor 112 that satisfies the predetermined condition transmits the position information, the predetermined condition can be positioned at the boundary of the enclosure wall 144 of the hand 92 here. The position sensor 112 determines whether or not the position sensor 112 itself is located at the boundary of the inclusion wall 144 of the hand 92 based on the three-dimensional positions of the plurality of other position sensors 112 existing in the vicinity. The position information is transmitted only when it is located at the boundary of 144.

  With such a configuration, as shown in FIG. 6C, only the position sensor 112, which is located at the boundary of the enclosure wall 144 of the hand 92 and represented by black filling in FIG. Therefore, the three-dimensional shape deriving device 100 can reduce the calculation load by not extracting the position information of all the position sensors 112, and only by the position information of the position sensors 112 located at the boundary. The shape of the containing wall 144 of the hand 92 can be accurately derived.

(Location information averaging process)
By the way, even if it is based on an electric field or electromagnetic wave, or based on a magnetic field, the three-dimensional position derived by each sensor 112 may not have sufficient accuracy. Therefore, the sensor 112 in the present embodiment can improve the accuracy of the three-dimensional position by using the three-dimensional positions of the other sensors 112 in the sensor network. For example, (1) when a suitable representative position sensor is derived as described with reference to FIG. 5A, three-dimensionally from a plurality of other sensors 112 in the vicinity of which an arbitrary sensor 112 is regarded as a group 140 The position is acquired, and the average value of the acquired three-dimensional positions of all the sensors 112 and the three-dimensional position derived by itself is obtained and set as the own three-dimensional position. It goes without saying that such an averaging process can also be applied to (2) deriving a representative position sensor within a predetermined spatial range, and (3) deriving a position sensor located at a boundary.

  For example, it is assumed that the position sensor 112 has a diameter of 0.1 mm and the position measurement accuracy of each position sensor 112 is 1 cm. Then, 1000 position sensors 112 are contained in a cubic space of 1 mm × 1 mm × 1 mm in the hand 92. When the average value of the position measurement values of the position sensor 112 is taken, the measurement accuracy is close to 1 mm. (Up to 2 mm or less).

  The method for improving the accuracy of the three-dimensional position is not limited to such a case. For example, the derivation of the three-dimensional position is repeatedly executed for each position sensor 112, the derivation results for a plurality of times are averaged, and the position sensor 112 is used. It is also possible to improve the accuracy of the three-dimensional position by setting the three-dimensional position.

(Three-dimensional shape derivation method)
Next, a method for deriving a three-dimensional shape will be described based on the flowchart of FIG. Here, a plurality of position sensors 112 are packed in the hand 92 so as to flow. First, all the position sensors 112 packed in the hand 92 derive their own three-dimensional positions based on the signal supplied from the signal supply source 110 (S200). Then, the three-dimensional position deriving unit 112e of each position sensor 112 acquires the three-dimensional positions of other neighboring position sensors 112, averages the three-dimensional positions, and obtains the barycentric position 172 of the other position sensors 112 ( S202).

  Then, the three-dimensional position deriving unit 112e obtains a connection line 176 that connects the point sensor 174 with the position sensor 112 as the symmetry point 170 and the position sensor 112 at the center of gravity position 172, and uses the connection line 176 as the central axis (2). −√2) It is determined whether or not another position sensor 112 is included in π steradians (S204). When another position sensor 112 is included (YES in S204), the three-dimensional position deriving unit 112e determines that itself (position sensor 112) is located in the gap 154 in the hand 92, and performs processing in the position sensor 112. finish. If no other position sensor 112 is included (NO in S <b> 204), it is determined that the position sensor 112 itself is located at the boundary of the enclosure wall 144 of the hand 92.

  Further, the three-dimensional position deriving unit 112e determined to be located at the boundary of the inclusion wall 144 of the hand 92 is a position determined to be located at the boundary of the inclusion wall 144 of the hand 92 in the vicinity (for example, within two hops). A three-dimensional position is acquired from the sensor 112 (S206), and an average value of the three-dimensional positions of all the position sensors 112 in the vicinity is obtained and set as its own three-dimensional position (S208). At this time, the position sensor 112 becomes a representative position sensor of the group of position sensors 112 positioned at the boundary of the nearby hand 92, and the position sensor 112 in two hops to which the three-dimensional position is referenced refers to the three-dimensional position as sensor information. It is not necessary to transmit to the consolidating unit 114. By doing so, it is possible to reduce the information transmission load in the sensor network and the calculation load in the central control unit 118.

  Thus, the communication unit 112c of the position sensor 112 located at the boundary of the hand 92 that has become the representative position sensor transmits position information indicating its own three-dimensional position (S210), and the sensor information aggregation unit 114 The position information transmitted from the sensor 112 is received (S212), and the three-dimensional shape deriving unit 130 obtains the three-dimensional shape of the portion of the hand 92 where the plurality of position sensors 112 exist based on the position information of the plurality of position sensors 112. Is derived (S214). Since various existing techniques can be used for the algorithm for deriving the three-dimensional shape from the three-dimensional positions of the plurality of position sensors 112, detailed description thereof is omitted here.

  As described above, the complicated shape of the hand 92 that cannot be confirmed from the outside, which is difficult to measure even with a large-scale apparatus such as a plurality of cameras or laser sensors, using the three-dimensional shape deriving device 100 described above, can be obtained using the sensor network. It can be derived. Further, since it is not necessary to use a visual sensor in the three-dimensional shape deriving device 100, even if the hand 92 is not projected, or under a situation where laser measurement does not function with another light source, it is possible to grasp the hand. A three-dimensional shape can be reliably derived, for example, a cavity is generated in the hand 92. Therefore, it is possible to accurately grasp the three-dimensional shape in various applications such as picking processing by a robot.

  Further, only when the position sensor 112 satisfies a predetermined condition, for example, the position sensor 112 itself is positioned at the boundary of the enclosure wall 144 of the hand 92 based on the three-dimensional positions of the plurality of other position sensors 112 present in the vicinity. It is possible to reduce the information transmission load in the sensor network and the calculation load in the central control unit 118 by transmitting the position information only when it is located at the boundary of the inclusion wall 144 of the hand 92. It becomes possible.

  Furthermore, since the three-dimensional shape deriving device 100 uses a large number of position sensors 112, the number of measurement points is large, and even if some of the position sensors 112 do not function, the measurement results are not hindered and robust. Sexuality can be secured. Also, high accuracy can be achieved by averaging a large number of position sensors 112.

  In addition, a program that functions as the three-dimensional shape deriving device 100 by a computer and a storage medium that stores the program are also provided. Further, the program may be read from a storage medium and taken into a computer, or may be transmitted via a communication network and taken into a computer.

  As mentioned above, although preferred embodiment of this invention was described referring an accompanying drawing, it cannot be overemphasized that this invention is not limited to this embodiment. It will be apparent to those skilled in the art that various changes and modifications can be made within the scope of the claims, and these are naturally within the technical scope of the present invention. Is done.

  For example, in the above-described embodiment, an example is given in which the position sensor 112 is packed in the hand 92 as it is, but the position sensor 112 does not need to be in the air and can be operated in a liquid or a solid. Therefore, there may be a method in which the liquid mixed with the position sensor 112 is poured into the hand 92 and measured. In addition, by mixing the position sensor 112 when forming plastic or the like, a material capable of measuring its own shape can be made.

  In the above-described embodiment, the configuration in which the three-dimensional position is derived in the position sensor 112 has been described. However, the signal supplied from the signal supply source 110 acquired by the position sensor 112 is transmitted to the sensor information aggregation unit 114 as it is. Thus, the three-dimensional position can be derived by another computer such as the central control unit 118.

  Further, by forming artificial skin as the inclusion wall 144 of the hand 92 and attaching the position sensor 112 to the artificial skin, the present embodiment can be performed with fewer position sensors 112 (attached position sensors 112). It becomes possible.

  If high measurement accuracy is not required, the signal supply source 110 can be omitted and the three-dimensional position can be derived from the relative relationship between the position sensors 112.

  Note that each step in the three-dimensional shape derivation method of the present specification does not necessarily have to be processed in time series in the order described in the flowchart, and may include parallel or subroutine processing.

  The present invention can be used in a three-dimensional shape deriving apparatus and a three-dimensional shape deriving method for deriving a three-dimensional shape through a sensor network.

92 ... hand (container)
DESCRIPTION OF SYMBOLS 100 ... Three-dimensional shape derivation | leading-out apparatus 110 ... Signal supply source 112 ... Position sensor 114 ... Sensor information aggregation part 130 ... Three-dimensional shape derivation | leading-out part 142 ... Predetermined spatial range 170 ... Symmetry point 172 ... Center of gravity position 174 ... Point symmetry 176 ... Connection

Claims (11)

A plurality of position sensors for transmitting position information indicating its own three-dimensional position or a physical quantity corresponding thereto;
A container for containing the plurality of position sensors;
A sensor information aggregating unit for receiving position information transmitted by the plurality of position sensors;
A three-dimensional shape deriving unit for deriving a three-dimensional shape of the inclusion body based on position information of the plurality of position sensors;
A three-dimensional shape deriving device comprising: And further comprising a plurality of signal supply sources arranged at different three-dimensional positions,
The plurality of position sensors derive their own three-dimensional positions by calculating a distance to the signal supply source or a direction of the signal supply source based on a signal supplied from the signal supply source. The three-dimensional shape deriving device according to claim 1.   The three-dimensional shape deriving device according to claim 1, wherein the enclosure is provided with a deformation mechanism.   The three-dimensional shape deriving device according to claim 3, wherein the inclusion body holds an arbitrary object.   The three-dimensional shape deriving device according to claim 4, wherein the position sensor further measures a pressure applied to the position sensor.   The three-dimensional shape derivation device according to claim 1, wherein the plurality of position sensors transmit position information only when a predetermined condition is satisfied. The predetermined condition is to be the only representative position sensor within a predetermined space range;
The three-dimensional shape deriving device according to claim 6, wherein the position sensor transmits position information only when the position sensor itself becomes the representative position sensor. The predetermined condition is to be located at a boundary of the inclusion;
The position sensor determines whether or not the position sensor itself is located at the boundary of the inclusion body based on the three-dimensional positions of a plurality of other position sensors existing in the vicinity. The three-dimensional shape deriving device according to claim 6, wherein only the position information is transmitted.   The position sensor determines that the position sensor itself is located at the boundary of the inclusion body if the number of other position sensors existing within the predetermined range is equal to or less than a predetermined threshold value based on the density of the position sensors. The three-dimensional shape deriving device according to claim 8.   The position sensor is located within a predetermined steradian and within a predetermined range with a center line as a connection line connecting the point sensors of the center of gravity of the plurality of other position sensors and the position sensor. The three-dimensional shape deriving device according to claim 8, wherein when there is not, it is determined that itself is located at a boundary of the inclusion body. A three-dimensional shape of the inclusion body is obtained using a three-dimensional shape deriving device including a plurality of position sensors, a inclusion body containing the plurality of position sensors, and a sensor information aggregating unit capable of communicating with the plurality of position sensors. A method for deriving a three-dimensional shape to be derived,
The plurality of position sensors transmit position information indicating their own three-dimensional position or a physical quantity corresponding thereto,
The sensor information aggregating unit receives position information transmitted by the plurality of position sensors,
A three-dimensional shape deriving method, wherein the three-dimensional shape of the inclusion is derived based on position information of the plurality of position sensors.

Images