JP2008026178A - Tactile sensor device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a tactile sensor device capable of containing a multitude of tactile sensors having a novel structure. <P>SOLUTION: The tactile sensor device includes a plurality of tactile sensor units. The tactile sensor unit includes the tactile sensor 100 and a processor element processing an output from the tactile sensor 100. The processor element is connected to the processor element of an adjacent tactile sensor unit so that data acquired from the tactile sensor can be transmitted. The tactile sensor 100 includes a substrate 101 and a coating film 104 arranged on the substrate 101, and detects force applied to the coating film 104. The tactile sensor 100 includes a laminate part 102 including a SiGe layer 122 and a Si layer 123 sequentially arranged from the side of the substrate 101, an erected part 103 formed by bending the SiGe layer 122 and the Si layer 123 to the side of the Si layer 123, and a detection means for detecting deformation of the erected part 103. The erected part 103 is deformed by application of force to the coating film 104. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a tactile sensor device.

A tactile sensor is used as a sensor for detecting the hand shape of a robot or the surface shape of an object. In recent years, a tactile sensor that detects the direction of a force applied to the sensor has also been proposed (for example, Patent Document 1). This sensor includes a substrate, a support member disposed above the substrate so as to be parallel to the substrate surface, and two electrodes disposed on the substrate and the support member. Then, the force applied to the sensor is detected based on the capacitance between the two electrodes.
JP 2004-264172 A

  However, the conventional sensor is not sufficiently sensitive to a force in a direction parallel to the surface of the substrate on which the sensor is formed. Further, when an elastic body such as an elastomer is formed so as to cover the tactile sensor, it is not easy to detect the force applied to the elastic body with high sensitivity.

  In view of such circumstances, an object of the present invention is to provide a tactile sensor device using a tactile sensor having a novel structure that is completely different from the conventional tactile sensor. Another object of the present invention is to provide a tactile sensor device that can include a large number of tactile sensors.

  In order to achieve the above object, a tactile sensor device of the present invention is a tactile sensor device including a plurality of tactile sensor units, and the tactile sensor unit includes at least one tactile sensor and the at least one tactile sensor. A processor element for processing the output, wherein the processor element of the tactile sensor unit is connected to the processor element of the adjacent tactile sensor unit so as to be able to transmit data obtained from the tactile sensor Yes. The tactile sensor includes a substrate and a film disposed on the substrate, and detects a force applied to the film, and the tactile sensor is disposed on the substrate in order from the substrate side. A laminated portion including the first and second layers formed, a standing portion formed by bending the first and second layers to the second layer side, and a deformation of the standing portion Each of the stacked portion and the upright portion includes a plurality of layers having different lattice constants, and the first and second layers are formed by a difference in lattice constants in the plurality of layers. It is bent by the generated force, the upright portion and the coating are in contact with each other, and the upright portion is deformed by applying a force to the coating.

  Unless otherwise specified, in this specification, “stacked on” or “arranged on” a predetermined object means that the object is directly stacked (or disposed) on the object, It means that it includes the case where it is laminated (or arranged) via another layer or space.

  According to the present invention, a tactile sensor device capable of detecting a force applied to a film is obtained. According to this tactile sensor device, it is possible to obtain information related to the magnitude and direction of the force applied to the film. In addition, according to the tactile sensor device, it is possible to obtain information on the force applied in the direction parallel to the surface of the substrate and the force applied in the direction perpendicular to the surface of the substrate. Unlike the conventional tactile sensor device, this tactile sensor device detects the force applied to the film by detecting the deformation of the standing part rising from the substrate. Can be detected with high resolution. In addition, by using this tactile sensor at the tip of the robot hand, it is possible to hold an object dexterously or perform a handshake operation.

  In the tactile sensor device of the present invention, the direction of the force applied to the film can be detected in two or three dimensions by combining a plurality of tactile sensors. The sensor device of the present invention including a plurality of such tactile sensors can detect an external pressure of an arbitrary area. In particular, the tactile sensor device of the present invention is characterized in that a tactile sensor can be formed finely, a plurality of sensors can be arranged at high density, and a force in a three-dimensional direction can be detected.

  Embodiments of the present invention will be described below. In addition, the example demonstrated below is an example of this invention and this invention is not limited to the following description.

[Tactile sensor device]
The tactile sensor device of the present invention includes a plurality of tactile sensor units. One tactile sensor unit includes at least one tactile sensor and a processor element that processes output from the at least one tactile sensor. The processor element of the tactile sensor unit is connected to the processor element of the adjacent tactile sensor unit so that data obtained from the tactile sensor can be transmitted.

  The tactile sensor is a sensor that includes a substrate and a film disposed on the substrate, and detects a force applied to the film. The tactile sensor includes a laminated portion including first and second layers arranged in order from the substrate side on the substrate, and an upright portion formed by bending the first and second layers to the second layer side. And detecting means for detecting the deformation of the upright portion.

  Each of the stacked portion and the standing portion includes a plurality of layers having different lattice constants. The first and second layers are bent by a force generated by a difference in lattice constant between the plurality of layers. The standing portion and the coating are in contact with each other. And a standing part deform | transforms when force is applied to the said film.

  The processor element includes an arithmetic unit and a memory. A simple arithmetic unit called an arithmetic logic unit (ALU) may be applied to the arithmetic unit of the processor element.

  The processor element processes tactile sensor data in the tactile sensor unit in which it is included. In addition, the processor element performs data processing using data in the processor element of the tactile sensor unit adjacent to the tactile sensor unit in which the processor element is included, as necessary.

  In the tactile sensor device of the present invention, the feature value of the output data of the tactile sensor included in the plurality of tactile sensor units existing in one area is extracted and output by at least one processor element in the area. It is preferable. In this case, the feature amount of the output data may be extracted by forming a network with a plurality of processor elements existing in the area. For example, if the data of one tactile sensor unit is “1” (the force detected by the tactile sensor is equal to or greater than the threshold value), the processor element of that tactile sensor unit includes the tactile sensor unit to which it belongs and the data is “ The range of the tactile sensor unit that is “1” may be determined, and the feature amount of the data in the range may be calculated and output. By extracting and outputting feature quantities of data from a plurality of tactile sensors in advance, the data processing load can be distributed, and the amount of output data to be transferred can be reduced and high-speed post-processing can be performed. .

  In the tactile sensor device of the present invention, a plurality of tactile sensor units may be arranged in a matrix. In this case, the processor elements of the tactile sensor units arranged at positions other than the outer edge are connected to the processor elements of four or more adjacent tactile sensor units so that data obtained from the tactile sensors can be transmitted. It may be. For example, the tactile sensor processor elements arranged at positions other than the outer edge may be connected to the processor elements of four tactile sensor units adjacent in the row and column directions. Further, the processor elements of the tactile sensor unit arranged at a position other than the outer edge may be connected to the processor elements of the four tactile sensor units adjacent in the oblique direction. The tactile sensor processor elements arranged at positions other than the outer edge include a total of eight tactile sensor unit processor elements including four tactile sensor units adjacent in the matrix direction and four tactile sensor units adjacent in the diagonal direction. It may be connected to. As described above, when the processor elements constitute a mesh network to exchange data, the feature amount of the data can be extracted.

  Note that the arrangement of the plurality of tactile sensor units is not limited to a matrix, and may be other arrangements. For example, the tactile sensor unit may be arranged at the position of the honeycomb-shaped lattice point. In this case, the processor element of the tactile sensor unit arranged at a position other than the outer edge may be connected to the processor elements of the three adjacent tactile sensor units. Moreover, you may arrange | position a tactile sensor unit in the center position of each hexagon of honeycomb shape.

  The tactile sensor device of the present invention may further include a computing unit connected to a plurality of processor elements. This computing unit is connected to the processor element by a general-purpose bus or a clock bus, and controls the processor element. This computing unit is connected to the processor element by a data collection network (including a memory, gate switch, and other computing units in the middle if necessary), and collects and processes data output from the processor element. I do. Hereinafter, this calculator may be referred to as a “control calculator”. The control calculator may be a part of the tactile sensor device or may be outside the tactile sensor device.

  The tactile sensor device of the present invention may further include another arithmetic unit and a memory arranged on a network connecting the plurality of processor elements and the control arithmetic unit. By causing the other arithmetic units to perform some functions of the control arithmetic unit, the burden on the control arithmetic unit can be reduced.

[Tactile sensor]
Hereinafter, the tactile sensor included in the tactile sensor unit will be described. The tactile sensor unit includes at least one tactile sensor. The number of tactile sensors included in one tactile sensor unit may be one, two, three, four, or four. It may be the above. By arranging the four tactile sensors so as to be shifted by 90 °, the direction of the shearing force can be detected.

  Hereinafter, the structure of one tactile sensor will be described. The tactile sensor is a tactile sensor that includes a substrate and a coating disposed on the substrate (and the laminated portion), and detects a force applied to the coating, and is arranged on the substrate in order from the substrate side. A laminated portion including the first and second layers, an upright portion formed by bending the first and second layers to the second layer side, and a detecting means for detecting deformation of the upright portion. . Each of the stacked portion and the upright portion includes a plurality of layers having different lattice constants. The first and second layers are bent by a force generated by a difference in lattice constant between the plurality of layers. The standing portion and the coating are in contact with each other, and the rising portion is deformed by applying a force to the coating.

  In the tactile sensor of the present invention, the force applied to the coating film can be detected by detecting the deformation of the standing part by the detecting means. The detection means may detect the deformation of the standing part based on the capacitance between the two electrodes formed on the substrate (on the stacked part) and the standing part. In this case, each electrode may be an electrode group including a plurality of electrodes. By using the electrode group, it becomes easy to detect the direction of the force. Further, the detecting means may detect the deformation of the rising portion based on the resistance value of the piezoresistive element formed at the rising portion or the boundary between the rising portion and the stacked portion. In this case, the influence of the temperature change may be compensated by using a known circuit such as a Wheatstone bridge. Further, the detecting means may detect the deformation of the upright portion by other known methods such as a method using magnetism. The magnitude of the force applied to the film can be obtained, for example, by preparing a calibration curve in advance for the relationship between the magnitude of the force applied to the film and the measured value measured by the detection means.

  A board | substrate is selected according to the laminated part formed on it. For example, when the stacked portion is formed of a semiconductor layer containing Si, an example of a preferable substrate is a Si substrate. Moreover, when a laminated part contains a III-V group compound semiconductor as a main layer, an example of a preferable substrate is a III-V group compound semiconductor substrate such as a GaAs substrate.

  The stacked portion includes a plurality of layers made of materials having different lattice constants. Due to the strain (internal stress) generated in the layers due to the difference in the lattice constants of these layers, at least a part of the layers constituting the laminated portion is bent to form an upright portion. The plurality of layers having different lattice constants are composed of, for example, semiconductor layers. Examples of such a combination of layers include a combination of a Si layer and a SiGe layer, and a III-V group compound semiconductor having a different composition and / or composition ratio. Examples of combinations of III-V compound semiconductors include GaAs / InGaAs, GaN / AlGaN, and GaN / InGaN.

  When the layer A having a large lattice constant and the layer B having a small lattice constant are adjacent to each other, an internal stress that tends to bend is generated in the layer B. Further, when the layer A having a large lattice constant and the layer B having a small lattice constant are laminated as thick layer A / layer B / thin layer A, the stress applied to the layer B by the thick layer A is thinner. Is larger than the stress exerted on the layer B, so that a total stress that tends to bend on the thin layer A side is generated. Similarly, in the laminated film of thick layer B / layer A / thin layer B, a stress that tends to bend toward the thick layer B is generated in total. By utilizing such a relationship, a plurality of stacked layers can be bent in a predetermined direction.

In order to separate the part of the stacked portion from the substrate to form the standing portion, a layer (sacrificial layer) that is selectively removed is formed between the substrate and the stacked portion. When the stacked portion is formed of a semiconductor containing Si, the sacrificial layer can be formed of, for example, SiO ₂ . Further, when the stacked portion is formed of a III-V group compound semiconductor, the sacrificial layer can be formed of a film in which, for example, a plurality of Al _0.5 Ga _0.5 As and AlAs layers are alternately stacked.

  The film is usually formed so as to cover the laminated portion and the standing portion. The space between the standing part and the substrate may be filled with a film or may be a space. The coating is formed of a material that is deformed by a force applied to the coating, and is usually an elastic body. The material of the coating, the thickness of the coating, and the shape and size of the element are selected according to the application required for the sensor.

  In the tactile sensor of the present invention, the detection means may include a piezoresistive element formed on the standing part. In this configuration, the deformation of the upright portion can be detected by monitoring the resistance value of the piezoresistive element.

  In the tactile sensor of the present invention, the first and second layers are bent to the second layer side in the first bent portion located at the boundary between the stacked portion and the standing portion, and the detection means includes the first A piezoresistive element formed in the bent portion of the above may be included. In this configuration, the deformation (tilt) of the upright portion can be detected by monitoring the resistance value of the piezoresistive element.

  In the tactile sensor of the present invention, the first and second layers are on the second layer side in the first bent portion located at the boundary between the stacked portion and the standing portion and the second bent portion in the standing portion. The standing portion includes a facing portion disposed so as to face the surface of the substrate by the second bending portion, and the detection means faces the substrate (on the stacked portion) so as to face each other. The first and second electrodes arranged on the part may be included, and deformation of the standing part may be detected based on a change in capacitance between the first electrode and the second electrode. In this case, the standing part may include a tip part formed by folding the first and second layers toward the first layer in the third bent part, and the tip part may be in contact with the coating. In this configuration, the force applied to the coating can be detected with high sensitivity by the tip.

  In the tactile sensor of the present invention, the upright portion includes a slide portion that slides on the substrate (on the laminated portion) in accordance with the movement of the film that contacts the upright portion, and the detection means is on the substrate (laminated) so as to face each other. Part) and the first and second electrodes arranged on the slide part, and deformation of the standing part is detected based on a change in capacitance between the first electrode and the second electrode. May be. The slide portion is formed, for example, by bending the first and second layers at the first bent portion located at the boundary between the stacked portion and the upright portion, and the second and third bent portions in the upright portion. it can.

  In the tactile sensor of the present invention, the coating may be made of an elastomer. A preferred example of the elastomer is silicone rubber. The hardness (JIS-A hardness) of the silicone rubber is, for example, in the range of 20-80. The viscosity of the silicone rubber before curing is, for example, in the range of 1 to 500 Pa · s. In addition, a film should just deform | transform according to the force added to the film and transmit the force to an upright part. Therefore, the film may be formed of a resin or a polymer.

  Further, the coating may have a multilayer structure. For example, it may be formed of a softer material (for example, low density silicone rubber) closer to the substrate and a harder material (for example, high density silicone rubber) closer to the surface.

  The tactile sensor unit of the present invention includes a plurality of tactile sensors of the present invention. The plurality of sensors are arranged in a predetermined position and a predetermined direction as necessary. For example, a plurality of tactile sensors may be arranged linearly or two-dimensionally. A plurality of tactile sensors may be arranged in the same direction or may be arranged in different directions. Further, a tactile sensor suitable for detecting a force applied in a direction horizontal to the substrate surface may be combined with a tactile sensor suitable for detecting a force applied in a direction perpendicular to the substrate surface.

  The tactile sensor device of the present invention can be applied to various fields as a device for recognizing external pressure. For example, the sensor device may be used as a sensor device that is embedded in a passage or stairs and recognizes an object moving on the passage. Moreover, you may use as a sensor apparatus which recognizes the external shape and unevenness | corrugation of an object. In the tactile sensor device of the present invention, minute sensors can be arranged at high density using a semiconductor process, so that the tactile sensor device can also be used as a device for recognizing minute irregularities such as fingerprints and palms. Further, it may be used as a sensor device for determining the state of the operator by being embedded in a machine operation unit (for example, a steering wheel of an automobile) or the like. Moreover, you may use as a sensor apparatus used for the artificial skin of the robot hand of an industrial robot, or a humanoid robot. Further, it may be used as a tactile sensor device for an artificial hand or an artificial leg.

  Hereinafter, an example of the tactile sensor device of the present invention will be described. In the following description, the same parts may be denoted by the same reference numerals and redundant description may be omitted. Moreover, in the following drawings, illustration of a film etc. may be omitted for easy understanding.

[Example of tactile sensor]
FIG. 1A is a perspective view schematically showing the shape of an example of a tactile sensor used in the apparatus of the present invention, and FIG. The tactile sensor 100 of FIG. 1 includes a substrate 101, a SiO ₂ layer 121 and a laminated portion 102 formed on the substrate 101, an upright portion 103, and a coating 104 (not shown in FIG. 1A). .

  The substrate 101 is a silicon substrate. The stacked unit 102 includes an SiGe layer 122 (thickness 20 nm) and an undoped Si layer 123 (thickness 340 nm) stacked in order from the substrate 101 side. The standing portion 103 is formed by bending the SiGe layer (first layer) 122 and the Si layer (second layer) 123 toward the Si layer 123. Since the SiGe layer 122 has a larger lattice constant than the Si layer 123, these layers are bent away from the substrate 101 due to internal strain.

  The upright portion 103 is formed by bending the SiGe layer 122 and the Si layer 123, and has a U-shape. A boron-doped p-Si layer 124 is formed on the Si layer 123 of the standing portion 103. Both ends of the U-shaped p-Si layer 124 are connected to electrodes 125 and 126 formed on the stacked portion 102, respectively. The p-Si layer 124 functions as a piezoresistive element, and the resistance changes due to the deformation of the upright portion 103. That is, the p-Si layer 124 functions as a detection unit that detects the deformation of the upright portion 103.

  The coating 104 is made of, for example, silicone rubber. The film 104 is formed so as to cover the stacked portion 102 and the upright portion 103 and to fill the space between the substrate 101 and the upright portion 103. Note that a space may be provided between the substrate 101 and the upright portion 103.

  When the coating 104 is pushed, the standing portion 103 is deformed by the shearing force generated in the coating 104, and the resistance value of the p-Si layer 124 changes. This resistance value increases or decreases depending on the direction in which the film 104 is pressed. For example, the force in the direction A and the force in the direction B of FIG. Therefore, by measuring the direction of change (increase or decrease) in the resistance value of the p-Si layer 124 and the magnitude of the change, the direction and magnitude of the force pressed by the film 104 can be estimated.

[Another example of tactile sensor]
FIG. 2A shows a perspective view schematically showing the shape of another example of the tactile sensor used in the present invention, and FIG. 2B shows a cross-sectional view thereof. The tactile sensor 200 of FIG. 2 includes a substrate 101, a SiO ₂ layer 221 (thickness 400 nm) and a laminated portion 202 formed on the substrate 101, a standing portion 203, and a coating 104 (not shown in FIG. 2A). Z)).

  The stacked unit 202 includes a Si layer 222 (thickness 100 nm), a SiGe layer 223 (thickness 30 nm, Ge amount 21 atomic%), a Si layer 224 (thickness 200 nm), and a SiGe layer 225 ( 30 nm thick, Ge content 21 atomic%), and Si layer 226 (100 nm thick).

  The upright portion 203 includes a Si layer (first layer) 222 and a SiGe layer (second layer) 223 (and Si layer 224) in the first bent portion 231 located at the boundary between the stacked portion 202 and the upright portion 203. Is bent to the SiGe layer 223 side. The standing portion 203 is formed so as to be substantially perpendicular to the substrate surface, for example. A piezoresistive element (strain gauge) 232 is formed in the first bent portion 231. The piezoresistive element 232 can be formed of a known material such as p-type Si.

  The film 104 is formed so as to cover the upright portion 203. When force is applied to the coating 104, the upright portion 203 is deformed, and the resistance of the piezoresistive element 232 formed in the bent portion 231 changes. From the change in resistance, the magnitude and direction of the force applied to the coating 104 can be detected.

[Another example of tactile sensor]
FIG. 3A is a perspective view schematically showing the shape of another example of the tactile sensor used in the present invention. FIG. 3A shows an example in which two tactile sensors 300 are arranged to face each other. A cross-sectional view of one tactile sensor 300 is shown in FIG. The tactile sensor 300 includes a substrate 101, a SiO ₂ layer 221 and a laminated portion 202 formed on the substrate 101, an upright portion 303, and a coating 104 (not shown).

  The stacked unit 202 has the same structure as the stacked unit 202 described in the second embodiment. The Si layer (first layer) 222 and the SiGe layer (second layer) 223 include a first bent portion 331 located at the boundary between the stacked portion 202 and the upright portion 303, and a second in the upright portion 303. The bent portion 332 is bent at about 90 ° toward the SiGe layer 223 side (valley fold). The standing portion 303 includes a facing portion 303 a that is oriented so as to face the surface of the substrate 101 by the second bent portion 332.

  Further, the standing portion 303 includes a tip end portion 303b formed by bending the Si layer 222 and the SiGe layer 223 by about 90 ° (mountain fold) toward the Si layer 222 in the third bent portion 333. The tip 303b is disposed substantially perpendicular to the surface of the substrate 101. The tip portion 303b may be omitted.

  The detection means of the touch sensor 300 includes first and second electrodes 341 and 342 formed on the stacked portion 202 and the facing portion 303a so as to face each other. The two electrodes can be formed of a conductive material (for example, a metal such as Au).

  The film 104 is disposed so as to be in contact with at least the surface of the Si layer 222 of the upright portion 303. The coating 104 may be disposed so as to fill a space between the stacked portion 202 and the facing portion 303a, but a space may be provided between the two. By arranging a plurality of tactile sensors 300 to face each other as shown in FIG. 3A and using a coating material having a high viscosity, it is possible to make a space between the stacked portion 202 and the facing portion 303a. .

  When force is applied to the coating 104, the upright portion 303 is deformed. Depending on the direction of the force applied to the coating 104, the second electrode 342 is parallel to the surface of the substrate 101 (direction A in FIG. 3B) or perpendicular (FIG. 3B). (B direction). As a result, the capacitance between the first electrode 341 and the second electrode 342 changes. By monitoring this change in capacitance, the force applied to the coating 104 can be detected. Note that at least one of the first electrode 341 and the second electrode 342 may be divided into a plurality of strip electrodes (the same applies to the electrodes of Embodiment 4). By arranging the strip-shaped electrodes in the direction A in FIG. 3B, it becomes easy to separate the change in capacitance due to movement in the parallel direction and the change in capacitance due to movement in the vertical direction. In addition, regarding the change in capacity due to the movement in the parallel direction, the direction of the movement can be easily determined.

[Another example of tactile sensor]
FIG. 4A is a perspective view schematically showing the shape of the main part of another example of the tactile sensor used in the present invention. The tactile sensor 400 of FIG. 4A includes a substrate 101, a SiO ₂ layer 221 and a laminated portion 202 formed on the substrate 101, a standing portion 403, a cover 440 (shown in FIG. 5), and a coating 104 ( (Not shown). A partial cross-sectional view of the standing portion 403 of the touch sensor 400 is shown in FIG.

  The stacked unit 202 has the same structure as the stacked unit 202 described in the second embodiment. A part of the Si layer 222 and the SiGe layer 223 (and the Si layer 224) is bent toward the SiGe layer 223 at the first bent portion 431 and the second bent portion 432. Further, a part of the Si layer 222 and the SiGe layer 223 are bent toward the Si layer 222 at the third bent portion 433. The standing portion 403 includes a slide portion 403a formed by bending the Si layer 222 and the SiGe layer 223. The slide part 403 a slides on the stacked part 202 in accordance with the movement of the coating 104 that contacts the upright part 403.

  The coating 104 is formed so as not to contact the slide part 403a from the second bent part 432 as much as possible so that the slide part 403a can move on the substrate 101. Therefore, as shown in FIG. 5, the tactile sensor 400 includes a cover 440 arranged to cover the slide part 403a. The cover 440 can be formed by bending the Si layer 222 and the SiGe layer 223 toward the SiGe layer 223 at the three bent portions. By using a material having a high viscosity as the material of the film 104, it is possible to prevent the film 104 from adhering to the slide portion 403a.

The detection means of the tactile sensor 400 includes first and second electrodes 441 and 442 formed on the stacked portion 202 and the slide portion 403a so as to face each other. The two electrodes can be formed of a conductive material (for example, a metal such as Au). In order not to short-circuit the two electrodes, a dielectric layer 443 made of, for example, SiO ₂ is formed on the surface of the electrodes.

  When a force is applied to the coating 104, the upright portion 403 is deformed and the slide portion 403a slides. As a result, the capacitance between the first electrode 441 and the second electrode 442 changes. By monitoring this change in capacitance, the force applied to the coating 104 can be detected.

[Example of tactile sensor unit]
A top view of an example of the tactile sensor unit used in the present invention is shown in FIG. The tactile sensor unit 500 of FIG. 6A includes four tactile sensors 100 arranged so that the directions are shifted by 90 °, and a processor element 510 that processes information obtained from the tactile sensors 100. A perspective view showing the arrangement of the four tactile sensors 100 of the tactile sensor unit of FIG. 6A is shown in FIG.

  The four tactile sensors 100 are connected to the processor element 510. The processor element 510 is connected to the processor element of the tactile sensor unit adjacent to the tactile sensor unit 500 to which the processor element 510 belongs by a wiring 501. Further, the processor element 510 is connected to a control arithmetic unit to be described later via a clock bus and a general-purpose bus.

  When the tactile sensor 100 is formed on a semiconductor substrate such as a Si substrate, the tactile sensor 100 and the processor element 510 can be formed on the same substrate. The wiring connecting the tactile sensor 100 and the processor element 510 can be formed by a semiconductor process.

  According to the touch sensor unit 500, as shown in FIG. 6A, the biaxial shear force in the X-axis direction and the Y-axis direction and the pressing force can be detected. The tactile sensor device of the present invention includes a plurality of tactile sensor units 500.

[Example of tactile sensor device]
An example of the tactile sensor device of the present invention is shown in FIG. The tactile sensor device 600 of FIG. 7 may be formed on one substrate or may be formed using a plurality of substrates.

  A tactile sensor device 600 of FIG. 7 includes a plurality of tactile sensor units 500 arranged in a matrix, a calculator 701, a memory 702, and a control calculator 703. In FIG. 7, wiring (network) connecting the processor element 510, the calculator 701, the memory 702, and the control calculator 703 of the tactile sensor unit 500 is omitted. A wiring (network) connecting the processor element 510, the arithmetic unit 701, the memory 702, and the control arithmetic unit 703 is shown in FIG.

  The processor element 510 of the tactile sensor unit 500 arranged at a position other than the outer peripheral portion includes a total of eight tactile sensor units 500 adjacent in the row direction and the column direction and four tactile sensor units 500 adjacent in the diagonal direction. The tactile sensor unit 500 is connected to the processor element 510.

  Among the processor elements 510 arranged in a matrix, the processor elements 510 existing at the end of the column and the end of the row are connected to the computing unit 701 so as to correspond one-to-one. The plurality of computing units 701 store the data acquired from the processor element 510 and processed as necessary in the memory 702. The control computing unit 703 acquires and processes the data stored in the memory 702. The memory 702 may be used as a shift register. The memory 702 and the control computing unit 703, or the computing unit 701, the memory 702, and the control computing unit 703 may be integrated into one chip.

  For example, the following two methods can be used as a method for the control computing unit 703 to access the data of the processor element 510.

  The first method is an access method using a general-purpose bus. In this case, the matrix number (coordinate) of the processor element 510 is designated, and data of a specific processor element 510 is accessed. However, since this method requires access to each processor element 510, there is a problem that the burden on the control computing unit 703 is large.

  The second method is to access the result of data processing performed on the network formed by the processor element 510. First, a command is issued from the control computing unit 703 to the predetermined processor element 510 to form a predetermined network and perform data processing via the general-purpose bus. Upon receiving the command, the processor element 510 forms a specific network, for example, a specific row direction or a 45 ° diagonal network in accordance with the command. The processor element 510 on the network combines the data of the processor element 510 adjacent on the network with its own data, and outputs the combined data to the processor element 510 adjacent to the other on the network. The processor element 510 at the end of the network outputs data obtained by combining the data of the processor elements 510 on the network to the computing unit 701. The computing unit 701 processes the obtained data as necessary, and writes it in the memory 702. The control computing unit 703 accesses the memory 702 and acquires data written in the memory 702. The control computing unit 703 obtains information on the pressure and shear force in a predetermined area by processing data on a plurality of networks written in the memory 702.

  In the second method, since the control computing unit 703 does not need to access the individual processor elements 510, the burden on the control computing unit 703 is small. Note that the first method and the second method may be properly used depending on the situation.

  In addition, when the designated network and the computing unit 701 are not directly connected, that is, in order to connect the designated network and the computing unit 701, connection via the processor element 510 that is not a data collection target is performed. It may be necessary. In such a case, the processor element 510 may be controlled so that the processor element 510 that is not a data collection target performs processing to pass through the data.

  The control computing unit 703 is connected to the processor element 510 by a clock bus and a general-purpose bus (both not shown in FIG. 8). The control computing unit 703 controls the processor element 510 via these buses. Further, the control computing unit 703 processes the obtained data and outputs the processed data to the outside as necessary. In addition, when arrange | positioning a tactile sensor in a wide area | region, the some tactile sensor apparatus 600 may be arranged, and the data obtained from them may be processed with a high-order arithmetic unit.

  FIG. 9 schematically shows the relationship between the control calculator 703 and the processor element 510. 6 and 7, four tactile sensors 100 are connected to one processor element 510, but only one of the four is shown in FIG. The processor element 510 includes an A / D converter 511, a program memory 512, a computing unit 513, a memory 514, and a gate switch 515. The processor element 510 may include other circuits and elements as necessary.

  The A / D converter 511 digitizes the output of the tactile sensor 100 with reference to a predetermined threshold. Here, for example, digitization is performed in two stages (binarization), four stages, eight stages, sixteen stages, or more. The digitized data is input to the calculator 513.

  The computing unit 513 executes a program recorded in the program memory 512. The computing unit 513 can be connected to the memory 514 ′ of the adjacent eight processor elements 510 via the gate switch 515. The computing unit 513 opens and closes the gate of the gate switch 515 as necessary, and accesses the data recorded in the memory 514 ′ of the nearby processor element 510. The computing unit 513 opens and closes the gate of the gate switch 515 as necessary to construct a network of an arbitrary path. Note that the gate switches 515 of some of the processor elements 510 at the outer edge are connected to the computing unit 701. A program memory 701a (corresponding to the program memory 512 of the processor element 510) is connected to the computing unit 701. The computing unit 701 also processes the data received from the adjacent processor element 510 according to the program recorded in the program memory 701a and stores it in the memory 702.

  It should be noted that the programs executed by each computing unit 513 and computing unit 701 may all be the same or different.

  The control computing unit 703 is connected to the computing unit 513 and the computing unit 701 by a clock bus 704 and a general-purpose bus 705. The control arithmetic unit 703 outputs a clock signal via the clock bus 704, moves the program counter of the arithmetic unit, and causes the arithmetic unit to execute the program.

The control arithmetic unit 703 receives predetermined information such as a processing procedure for each clock, a gate opening / closing procedure, a threshold value for digitizing data, calibration data, and the like via the general-purpose bus 705. And input to the calculator 701. Examples of processing included in a program that is transferred in advance from the computing unit for control 703 to the computing unit 513 and the computing unit 701 include the following processes.
A. Get tactile sensor data.
B. Binarize the tactile sensor data.
C. Connect the left and right processor elements to the network.
D. Add the data of the processor element on the left and the data of its own.
E. Disconnect the network connection with the left and right processor elements.

The computing unit 513 properly interprets the transmitted program. For example, each process included in the transmitted program is replaced with a process according to a predetermined clock timing. When the process to be executed is “A. Acquire sensor data.”, The process may be executed according to a procedure interpreted at each predetermined clock timing as follows. The following processing and the number of clocks are examples.
A-1. Apply current to the sensor (4 clocks)
A-2. Measure the resistance value (8 clocks)
A-3. Discretize (40 clocks)
A-4. Calibrate (90 clocks)
A-5. Save to memory (2 clocks)

  The computing unit 701 processes the data input from the processor element 510 and stores it in the memory 702. Data stored in the memory 702 is read and processed by the control computing unit 703.

  In a tactile sensor device having a large number of tactile sensors, if data of all tactile sensor units are collected and processed one by one, there is a problem that it takes time to process and a great burden is placed on the control calculator. On the other hand, since the tactile sensor device 600 of the present invention can perform distributed processing, it is possible to obtain information on the pressure applied to the tactile sensor unit without collecting data of all the tactile sensor units. Is possible. For example, when pressure is applied to only a part of a large number of tactile sensor units, the processor element extracts only the data of the tactile sensor units to which pressure is applied, and the extracted data is input to the control computing unit. May be.

  Hereinafter, an example of data processing in the tactile sensor device 600 of the present invention will be described. First, the processor element 510 acquires resistance value data (analog data) corresponding to the pressure from each tactile sensor 100 of the tactile sensor unit 500, and digitizes the data by the A / D converter 511. Next, the digitized data is calibrated with the calibration data stored in advance in the memory 514, and the calibrated data is written in the memory 514. Such processing is performed for each of the four tactile sensors 100 included in the tactile sensor unit 500.

  Note that the pressing force, the shearing force in the X-axis direction, and the shearing force in the Y-axis direction shown in FIG. 6A may be calculated using data of the four tactile sensors 100. Here, the data obtained from the two tactile sensors 100 facing in the X-axis direction in FIG. 6A are respectively A and B, and the data obtained from the two tactile sensors 100 facing in the Y-axis direction in FIG. The data to be obtained are C and D, respectively. In one example, the average value of the data A, B, C, and D may be data corresponding to the pressing force shown in FIG. Further, the difference between the data A and B may be data corresponding to the shearing force in the X-axis direction. Similarly, the difference between the data C and D may be data corresponding to the shearing force in the Y-axis direction. In the data processing, the sign of the data may be inverted or a predetermined coefficient may be applied to the data as necessary. These processes may be performed by software, or may be performed by hardware by assembling a circuit for performing the process. The software processing method has the advantage of high flexibility. On the other hand, the method performed by hardware has an advantage that processing can be speeded up and an advantage that processor elements can be simplified.

  In one example, the tactile sensor device of the present invention may monitor only the pressure. In that case, you may process only the data of the tactile sensor unit 500 of the part which detected the press among the tactile sensor units 500. FIG. As a method for performing such a process, for example, there is a method of performing a centroid calculation of data of a group of tactile sensor units 500 that detect a press and a subsequent radon conversion.

The method for calculating the center of gravity is not particularly limited, and a known method may be applied. Hereinafter, an example of a method for calculating the center of gravity for an area where the tactile sensor units 500 are arranged in [Nx columns] × [Ny rows] will be described. In the following description, a case where the tactile sensor unit 500 includes only one tactile sensor 100 will be described for simplification. The center of gravity (C _x , C _y ) is obtained by the following equation.

  The center of gravity can be obtained by the above formula. Note that the area of [Nx column] × [Ny row] to be cut out is determined according to the situation. For example, when the area where the object contacts is known, only the vicinity thereof may be cut out. If the area touched at the previous time is known, an area slightly wider than the area touched at the previous time may be cut out on the assumption that the object is also in contact with the current time. . Moreover, you may perform previously the process which restricts a contact area from all the areas.

By using an arithmetic unit 513 and operation unit 701, it is possible to calculate the sum M _x of each row sum M _y, and each column of. Such processing can be performed by the following procedure, for example.
1. The data obtained from the tactile sensor is binarized.
2. A network connection is made in the row direction, and the number of units whose data value is “1” is counted. This process is performed for each row.
3. Network connection is made in the column direction, and the number of units whose data value is “1” is counted. This process is performed for each column.

The addition of data in the row direction can be performed by constructing an adder circuit in which Nx arithmetic units 513 included in one row take charge of each digit of the full adder. Using such adder circuit, by adding the data sequentially from the most distant calculator 513 from the arithmetic unit 701, the bit string representing the sum M _y in the row direction of the data is obtained. The total sum M _x of the data in the column direction can be calculated in the same procedure.

  The center of gravity is obtained as described above. It is also possible to recognize the shape of the pressed part by performing data processing following the calculation of the center of gravity. The shape recognition may be performed by applying a Radon transform used in, for example, a CT scan.

  In the Radon transform, data in a predetermined sampling direction passing through the center of gravity is added, and shape recognition is performed based thereon. For example, when the entire inside of the region 801 having the shape shown in FIG. 10A is pressed, data of a plurality of sampling directions (directions a, b and c in FIG. 10A) passing through the center of gravity are added. Thus, data as shown in FIG. 10B is obtained. It is possible to estimate the shape of the pressed area by comparing the data (feature feature amount) of FIG. 10B with the feature amounts of various shapes registered in advance. At this time, the accuracy of shape recognition increases as the number of sampling directions increases, but the processing load increases, so the number of sampling directions is set to an appropriate value.

  When adding data in the sampling direction, a network is formed along the sampling direction, and the data of the tactile sensor units on the network are added. The addition of data can be performed by the above-described method or a known method. If a network that completely matches the sampling direction cannot be formed, the network closest to the sampling direction may be formed for processing.

  In the tactile sensor device of the present invention, it is also possible to extract the contour of the pressed area or reduce noise by filtering the obtained data. For example, it is possible to extract a contour by constructing a network with nine tactile sensor units of 3 columns × 3 rows, and performing first-order differentiation (or difference) of data in the region. Such processing is well known in the field of image processing. In the filter processing, for example, a Sobel filter well-known in image processing or a Previtt filter may be applied.

  In the tactile sensor of the present invention, as shown in FIG. 11, all the tactile sensors 100 may be arranged in the same direction. Although the processor element is not shown in FIG. 11, the processor element is connected to one or more tactile sensors 100. In the above example, the tactile sensor device using the tactile sensor 100 has been described, but other sensors other than the tactile sensor 100 may be used.

  In the data processing in the tactile sensor device of the present invention, a method proposed in image processing such as an image sensor may be used. For example, JP 2001-195564 A, JP 2003-218338 A, and the IEICE Transactions on Dynamically Reconfigurable SIMD Processors for Vision Chips (Vol. J86-D-II, No. 11, pp. 1575-1585, November 2003) may be used.

  The tactile sensor device of the present invention can be used for artificial hands and prosthetic legs having a tactile sensor, and artificial skin of a robot having a tactile sensor. The tactile sensor device of the present invention can be embedded in a handle, a staircase, a passage or the like and used as a part of a device that senses and predicts human movement. As described above, it is also possible to extract the contour of the pressed area using the tactile sensor device of the invention. For example, when an object as shown in FIG. 12A is arranged on the tactile sensor device, the acquired data is processed to obtain the contour of the pressed area as shown in FIG. It is possible to obtain the data. Therefore, the tactile sensor device of the present invention can be used as a device for recognizing the shape and contour of an object placed thereon.

[Manufacturing method of tactile sensor]
The tactile sensor device of the present invention can be formed by combining known methods used for manufacturing semiconductor elements. Hereinafter, an example of the manufacturing method will be described using the tactile sensor 300 as an example. Note that the processor element of each tactile sensor unit may be formed directly on the same substrate as the tactile sensor, or may be arranged by mounting an element incorporating a circuit on the substrate.

  A manufacturing process is shown to Fig.13 (a)-(f). 13A, 13C, and 13E are top views, and FIGS. 13B, 13D, and 13F are cross-sectional views corresponding thereto.

First, an SOI substrate ((100) surface) in which a SiO ₂ layer 221 (thickness 400 nm) and a Si layer (thickness 20 nm) are formed on a substrate 101 (thickness 400 μm) made of Si single crystal is prepared. A Si layer (thickness 80 nm) is epitaxially grown on the substrate to form a Si layer 222 (thickness 100 nm). Next, the SiGe layer 223, the Si layer 224, the SiGe layer 225, and the Si layer 226 are sequentially epitaxially grown to form the stacked unit 202. Epitaxial growth can be performed under normal conditions. For example, it can be performed under conditions of a substrate temperature of 550 ° C. and a growth rate of 60 nm / h. Ge may be supplied by a Knudsen cell, and Si may be supplied by EB vapor deposition.

In this way, a multilayer film is formed on the substrate 101 as shown in FIGS. Next, as shown in FIGS. 13C and 13D, a part of the multilayer film is etched to form three types of grooves 11, 12, and 13. These grooves can be formed by combining photolithography and etching. The groove 11 is formed at a position corresponding to the first and second bent portions 331 and 332 (see FIG. 3B). The groove 11 can be formed by etching the Si layer 226 and the SiGe layer 225. The groove 12 is formed at a position corresponding to the third bent portion 333 (see FIG. 3B). The groove 12 can be formed by etching the Si layer 226, the SiGe layer 225, and the Si layer 224. The groove 13 is a groove for cutting the stacked portion 202 to form the upright portion 303. The groove 13 is formed so as to reach the SiO ₂ layer 221. The groove 13 can be formed by etching the Si layer 226, the SiGe layer 225, the Si layer 224, the SiGe layer 223, and the Si layer 222. The Si layer can be etched with, for example, an NH ₃ aqueous solution. The SiGe layer can be etched with, for example, a mixed aqueous solution of HF / H ₂ O ₂ / CH ₃ COOH. When forming the groove 13, the Si layer and the SiGe layer may be etched with a mixed aqueous solution of HF / HNO ₃ / CH ₃ COOH.

  Next, first and second electrodes 341 and 342 are formed at predetermined positions on the Si layer 226. At this time, a wiring (not shown) connected to the electrode is also formed. These electrodes and wiring can be formed by a known method such as vapor deposition.

Next, the SiO ₂ layer 221 present at the position corresponding to the portion that becomes the standing portion 303 is removed. The removal of the SiO ₂ layer 221 can be performed by wet etching through the groove 13. For example, hydrofluoric acid (HF aqueous solution) can be used as the etching solution. When the SiO ₂ layer 221 existing below the portion that becomes the upright portion 303 is removed, the semiconductor layer is bent at the three bent portions.

  In the groove 11 portion, the SiGe layer 223 is sandwiched between two Si layers 222 (thickness 100 nm) and Si layer 224 (thickness 200 nm) having a smaller lattice constant. Therefore, the stress that the thick Si layer 224 exerts on the SiGe 223 is larger than the stress that the thin Si layer 222 exerts on the SiGe layer 223, and therefore, these three layers bend toward the Si layer 224 side. On the other hand, since only the Si layer 222 and the SiGe layer 223 are present in the groove 12, the two layers are bent toward the Si layer 222 side. In this way, the semiconductor layer is bent at three locations, and the standing portion 303 as shown in FIG. 3 is formed.

  Next, the film 104 is formed so as to cover the substrate 101 and the upright portion 303. The coating 104 can be formed, for example, by applying the material of the coating 104 on the substrate 101 and then curing the material. The material of the film 104 can be applied by a known method such as a spin coating method. The method of curing the material of the coating 104 is selected according to the material. In this way, the tactile sensor 300 is formed.

Note that other tactile sensors other than the tactile sensor 300 can be formed by the same method. The groove patterns for forming the tactile sensors 100, 200 and 400 are shown in FIGS. 14A, 14B and 14C, respectively. In FIG. 14A, the groove 13 is a groove reaching the SiO ₂ layer. In FIG. 14B, two narrow grooves 11 are formed in the first bent portion, but a single wide groove 11 may be used.

  According to the manufacturing method, a plurality of tactile sensors can be simultaneously formed on the same substrate, and the tactile sensor unit of the present invention can be easily manufactured. In addition, the said manufacturing method is an example and you may form the tactile sensor of this invention by another method. For example, each layer on the substrate can be formed by vapor deposition, CVD, or other various epitaxial methods depending on the layer. As an etching method for forming the groove formed in the bent portion, an arbitrary method can be applied depending on the semiconductor layer constituting the stacked portion.

In the above-described embodiment, the example using the multilayer film in which the Si layer and the SiGe layer are stacked has been described. However, a similar sensor can be formed using a III-V group compound semiconductor. Note that the angle of the bent portion can be controlled by the thickness and composition of the semiconductor layer and the width of the groove. For example, in the case of a bent portion composed of an In _0.2 Ga _0.8 As layer and a GaAs layer, and when these are layers having the thicknesses described above, the radius of curvature of the GaAs layer is usually as follows: It is approximated by the formula.
R = (a / Δa) · {(t1 + t2) / 2}
Here, a is a lattice constant of the GaAs layer and is 5.6533 angstroms. Δa is the difference between the lattice constant (5.7343 Å) of the In _0.2 Ga _0.8 As layer and the lattice constant of the GaAs layer. T1 is the thickness of the In _0.2 Ga _0.8 As layer (unit: angstrom), and t2 is the thickness of the GaAs layer (unit: angstrom).

  The embodiments of the present invention have been described above by way of examples, but the present invention is not limited to the above-described embodiments, and can be applied to other embodiments based on the technical idea of the present invention.

  The tactile sensor device of the present invention can be applied to various fields as a device for sensing external pressure.

It is (a) perspective view and (b) sectional view showing typically an example of a tactile sensor used by the present invention. It is (a) perspective view and (b) sectional drawing which show typically another example of the tactile sensor used by this invention. It is the (a) perspective view and (b) sectional drawing which show typically another example of the tactile sensor used by this invention. It is (a) perspective view and (b) sectional view showing typically some other examples of a tactile sensor used by the present invention. It is a perspective view which shows typically the structure of the tactile sensor of FIG. 1A is a top view schematically showing a configuration of an example of a touch sensor unit used in the present invention, and FIG. 2B is a perspective view showing an arrangement of touch sensors. It is a top view which shows typically the structure of an example of the tactile sensor apparatus of this invention. It is a figure which shows typically the combined state of a processor element, a calculator, etc. about the tactile sensor apparatus shown in FIG. It is a figure which shows typically the relationship of each component about the tactile sensor apparatus shown in FIG. It is a figure which shows an example of the method of recognizing the shape of a press part using the tactile sensor unit of this invention. It is a perspective view which shows typically the other example of the tactile sensor apparatus of this invention. It is a figure which shows an example of the method of recognizing the outline of an object using the tactile sensor apparatus of this invention. It is sectional drawing which shows typically an example of the manufacturing method of the tactile sensor apparatus of this invention. It is a top view which shows typically the other example of the manufacturing method of the tactile sensor apparatus of this invention.

Explanation of symbols

11, 12, 13 Groove 100, 200, 300, 400 Tactile sensor 101 Substrate 102, 202 Laminated portion 103, 203, 303, 403 Standing portion 104 Coating 121, 221 SiO ₂ layer 122, 223, 225 SiGe layer 123, 222, 224, 226 Si layer 124 p-Si layer (piezoresistive element)
231, 331, 332, 333, 431, 432, 433 Bending part 232 Piezoresistive element 303a Opposing part 303b Tip part 341, 342, 441, 442 Electrode (detection means)
403a Slide unit 440 Cover 443 Dielectric layer 500 Touch sensor unit 510 Processor element 600 Touch sensor device 701 Calculator 702 Memory 703 Control calculator

Claims (5)

A tactile sensor device including a plurality of tactile sensor units,
The tactile sensor unit includes at least one tactile sensor and a processor element that processes output from the at least one tactile sensor;
The processor element of the tactile sensor unit is connected to the processor element of the adjacent tactile sensor unit so that data obtained from the tactile sensor can be transmitted;
The tactile sensor includes a substrate and a film disposed on the substrate, and detects a force applied to the film.
The tactile sensor includes a laminated portion including first and second layers arranged in order from the substrate side on the substrate, and bending the first and second layers to the second layer side. And further comprising a formed upright portion and a detecting means for detecting deformation of the upright portion,
Each of the stacked portion and the standing portion includes a plurality of layers having different lattice constants from each other,
The first and second layers are bent by a force generated by a difference in lattice constant in the plurality of layers,
The upright part and the coating are in contact with each other;
The tactile sensor device, wherein the upright portion is deformed when a force is applied to the coating.   The feature amount of the output data of the tactile sensor included in the plurality of tactile sensor units existing in one area is extracted and output by at least one processor element in the area. Tactile sensor device.   The tactile sensor device according to claim 2, wherein the feature amount is extracted when a plurality of the processor elements existing in the region form a network. The plurality of tactile sensor units are arranged in a matrix,
The processor element of the tactile sensor unit arranged at a position other than the outer edge can transmit data obtained from the tactile sensor to the processor elements of four or more adjacent tactile sensor units. The tactile sensor device according to claim 1, which is connected.   The tactile sensor device according to claim 1, further comprising a computing unit connected to the plurality of processor elements.

Images