Optical arithmetic device

Abstract

An optical computation device (1) is provided with: an optical computation unit (11) including an optical modulation element (11ai); and an image sensor (12) that generates an electric signal indicating the intensity distribution of the signal light output from the light calculation unit (11). The phase modulation amount of each unit constituting the light modulation element (11ai) is set so that the direction in which the signal light (SLi) is emitted differs from the direction in which the noise light (NSi) is emitted. As a result, an optical arithmetic device is realized in which the influence of noise light on an electrical signal generated by an image sensor is reduced.

Description

Technical Field

[0001] The present invention relates to an optical computing device that performs optical computing using an optical modulation element, wherein the optical modulation element has a plurality of units, and the plurality of units independently set phase modulation amounts or can independently set phase modulation amounts. Background Art

[0002] A light modulator is known that is designed to have a plurality of units and optically performs a predetermined operation by interfering the signal lights that pass through each unit. Optical operations using such light modulators have the advantages of high speed and low power consumption compared to electrical operations using a processor. In addition, by allowing a plurality of light modulators to act on the signal light in sequence, it is possible to easily achieve a high level of optical operations that are difficult to achieve with a single light modulator.

[0003] Patent Document 1 discloses an optical neural network having an input layer, an intermediate layer, and an output layer. The above-mentioned light modulation element can be used as, for example, the intermediate layer of such an optical neural network.

[0004] Prior art literature

[0005] Patent Literature

[0006] Patent Document 1: U.S. Patent No. 7847225 Summary of the invention

[0007] 1. Technical issues to be resolved

[0008] However, when the signal light before the operation is made incident on the optical modulator, in addition to the signal light representing the operation result emitted from the optical modulator, various noise lights are also emitted from the optical modulator. In the case where the optical modulator is a transmission-type optical modulator, the transmitted light phase-modulated by each unit constituting the optical modulator interferes, thereby generating a signal light representing the operation result. The noise light generated because the signal light before the operation is not phase-modulated by each unit and is transmitted (normally transmitted) by the optical modulator is emitted in the same direction as the signal light. In addition, in the case where the optical modulator is a reflection-type optical modulator, the reflected light phase-modulated by the above-mentioned multiple units interferes, thereby generating a signal light representing the operation result. The noise light generated because the signal light before the operation is not phase-modulated by each unit and is reflected (normally reflected) by the optical modulator is emitted in the same direction as the signal light.

[0009] Among the noise light emitted from each light modulator, the noise light emitted in the same direction as the signal light enters the image sensor together with the signal light. Therefore, the electrical signal generated by the image sensor is affected by the noise light. Furthermore, when the influence of the noise light is large, it is difficult to correctly read the result of the optical operation based on the electrical signal generated by the image sensor.

[0010] One aspect of the present invention is made in view of the above-mentioned problem, and an object of the present invention is to realize an optical computing device that reduces the influence of noise light on an electrical signal generated by a light sensor.

[0011] (II) Technical solution

[0012] The optical computing device of the first aspect of the present invention is characterized in that it comprises: an optical computing unit including an optical modulator, the optical modulator having a plurality of units for which phase modulation amounts are independently set or for which phase modulation amounts can be independently set; and an optical sensor for detecting signal light output from the optical computing unit and generating an electrical signal representing a detection result, wherein the optical modulator is (1) a transmissive optical modulator that emits signal light generated by interference of transmitted light that has been phase-modulated by the plurality of units, respectively, and noise light that has not been phase-modulated by the plurality of units and has transmitted the optical modulator, or (2) The optical modulator is a reflective optical modulator, which emits signal light generated by interference of reflected light that has been phase-modulated by the multiple units, and noise light that has not been phase-modulated by the multiple units and has been reflected by the optical modulator. In the optical modulator, the phase modulation amount of each of the multiple units is set so that the direction in which the signal light is emitted from the optical modulator is different from the direction in which the noise light is emitted from the optical modulator. Compared with the case in which the direction in which the signal light is emitted from the optical modulator is the same as the direction in which the noise light is emitted from the optical modulator, the influence of the noise light on the electrical signal is smaller.

[0013] (III) Beneficial effects

[0014] According to one aspect of the present invention, it is possible to realize an optical computing device that reduces the influence of noise light on an electrical signal generated by an image sensor. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a side view showing the structure of the optical computing device according to the first embodiment of the present invention.

[0016] Figure 2 Yes means Figure 1 A side view of a variation of the optical computing device shown.

[0017] Figure 3 Yes means Figure 1 FIG. 1 is a plan view and a cross-sectional view of a first configuration example of a light modulation element included in the optical computation device shown.

[0018] Figure 4 Yes means Figure 1 FIG. 2 is a cross-sectional view of a second configuration example of a light modulation element included in the optical computation device shown.

[0019] Figure 5 Yes means Figure 1 FIG. 2 is a perspective view of a third configuration example of a light modulation element included in the optical computation device shown.

[0020] Figure 6 It is a side view showing the structure of an optical computation device according to a second embodiment of the present invention.

[0021] Figure 7 Yes means Figure 6 FIG. 4 is a side view of a first variation of the optical computation device shown.

[0022] Figure 8 Yes means Figure 6 The top view, side view, and plan view of a second variation of the optical computation device are shown.

[0023] Fig. 9 Yes means Figure 6 FIG. 1 is a plan view and a cross-sectional view of a first configuration example of a light modulation element included in the optical computation device shown.

[0024] Fig.10 Yes means Figure 6 FIG. 2 is a cross-sectional view of a second configuration example of a light modulation element included in the optical computation device shown.

[0025] Fig.11 Yes means Figure 6 FIG. 2 is a perspective view of a third configuration example of a light modulation element included in the optical computation device shown. DETAILED DESCRIPTION

[0026] [First embodiment]

[0027] (Structure of optical computing device)

[0028] Reference Figure 1 The structure of the optical computing device 1 according to the first embodiment of the present invention will be described. Figure 1 It is a side view showing the structure of the optical computation device 1 .

[0029] like Figure 1 As shown, the optical computation device 1 includes an optical computation unit 11 and an image sensor 12 .

[0030] The optical operation unit 11 is a set of at least one optical modulator 11a1 to 11an (n is any natural number greater than 1). In this embodiment, in order to realize multi-stage optical operations, a set of three optical modulators 11a1 to 11a3 is used as the optical operation unit 11. In addition, the optical modulators 11a1 to 11an can be integrated. For example, n light diffraction layers formed in a structure that transmits signal light such as a dried gel can be used as the optical modulators 11a1 to 11an.

[0031] Each optical modulator 11ai (i is a natural number greater than 1 and less than n) is composed of a plurality of units whose phase modulation amounts are set independently of each other. Each optical modulator 11ai is a transmission type optical modulator, and performs optical calculation by causing the transmitted light phase-modulated by each unit to interfere with each other. Here, performing optical calculation means: converting the two-dimensional intensity distribution of the signal light from the two-dimensional intensity distribution representing the information before the calculation to the two-dimensional intensity distribution representing the information after the calculation. The structural example of the transmission type optical modulator will be described later with reference to different drawings.

[0032] In addition to emitting the above-mentioned signal light, the light modulator 11ai also emits noise light that has not been phase-modulated by each unit and has passed through the light modulator 11ai. For example, when there is a gap between the units constituting the light modulator 11ai, the light passing through the gap is noise light. The optical axis of the noise light emitted from the light modulator 11ai is located in the plane where the optical axis of the signal light incident on the light modulator 11ai and the normal of the incident surface of the light modulator 11ai are located. The emission angle of the noise light emitted from the light modulator 11ai is the same as the incident angle of the signal light incident on the light modulator 11ai. That is, the optical axis of the noise light is consistent with the optical axis of the transmitted light when the signal light is passing through the light modulator 11ai.

[0033] In order to separate the noise light from the signal light, the phase modulation amount of each unit constituting the optical modulator 11ai is set so that the direction in which the signal light is emitted from the optical modulator 11ai is different from the direction in which the noise light is emitted from the optical modulator 11ai. In a transmission type optical modulator, the emission angle of the noise light is the same as the incident angle of the incident light. Therefore, the phase modulation amount of each unit is set so that the emission angle of the signal light is different from the incident angle of the incident light.

[0034] In the present embodiment, the signal light SL0 is perpendicularly incident on the incident surface of the first light modulator 11a1. The first light modulator 11a1 causes the noise light NL1 to be emitted perpendicularly from the exit surface, and causes the signal light SL1 to be emitted obliquely from the exit surface. The second light modulator 11a2 is configured such that (1) its incident surface is parallel to the exit surface of the first light modulator 11a1, (2) its incident surface intersects with the signal light SL1 emitted from the first light modulator 11a1, and (3) its incident surface does not intersect with the optical axis of the noise light NL1 emitted from the first light modulator 11a1. In addition, the following condition may be set instead of condition (3), namely, the optical axis of the noise light NL1 emitted from the first light modulator 11a1 does not intersect with the intersection point where the incident surface of the second light modulator 11a2 intersects with the signal light SL1 emitted from the first light modulator 11a1. Alternatively, a condition may be set such that at least a part of the noise light NL1 emitted from the first light modulation element 11 a 1 does not enter the incident surface of the second light modulation element 11 a 2 .

[0035] In addition, in the present embodiment, the second light modulator 11a2 emits the noise light NL2 obliquely and emits the signal light SL2 vertically. The third light modulator 11a3 is configured such that (1) its incident surface is parallel to the emission surface of the second light modulator 11a2, (2) its incident surface intersects with the signal light SL2 emitted from the second light modulator 11a2, and (3) its incident surface does not intersect with the optical axis of the noise light NL2 emitted from the second light modulator 11a2. In addition, instead of condition (3), the following condition may be set, that is, the optical axis of the noise light NL2 emitted from the second light modulator 11a2 does not intersect with the intersection point where the incident surface of the third light modulator 11a3 and the signal light SL2 emitted from the second light modulator 11a2 intersect. Alternatively, the following condition may be set, that is, at least a part of the noise light NL2 emitted from the second light modulator 11a2 does not enter the incident surface of the third light modulator 11a3. The third light modulation element 11a3 emits the noise light NL3 vertically and emits the signal light SL3 obliquely. In the present embodiment, the optical axes of the signal lights SL0 to SL3 and the noise lights NL1 to NL3 are located in the same plane.

[0036] The image sensor 12 is an example of an optical sensor, and is used to detect the signal light SL3 output from the optical calculation unit 11 and generate an electrical signal indicating the detection result. The image sensor 12 is composed of a plurality of photoelectric conversion units to generate an electrical signal indicating the intensity distribution of the signal light SL3 emitted from the optical calculation unit 11. The image sensor 12 is configured such that (1) its incident surface is parallel to the exit surface of the third optical modulation element 11a3, (2) its incident surface intersects with the optical axis of the signal light SL3 emitted from the optical calculation unit 11, and (3) its incident surface does not intersect with the optical axis of the noise light NL1 to NL3 emitted from the optical calculation unit 11. In addition, instead of condition (3), the following condition may be set, that is, the optical axis of the noise light NL1 to NL3 emitted from the optical calculation unit 11 does not intersect with the intersection point where the incident surface of the image sensor 12 intersects with the optical axis of the signal light SL3 emitted from the optical calculation unit 11.

[0037] By configuring the light modulators 11a1 to 11a3 and the image sensor 12 in the above manner, the influence of the noise light NL1 to NL3 on the electric signal generated by the image sensor 12 is small compared to the conventional optical computing device in which the signal light SLi and the noise light NLi are emitted from each light modulator 11ai in the same direction. Instead of the condition that "the incident surface (any point) of the image sensor 12 does not intersect with the optical axis of the noise light NL1 to NL3", the following condition may be set: the point intersecting with the optical axis of the signal light SL1 to SL3 on the incident surface of the image sensor 12 is different from the point intersecting with the optical axis of the noise light NL1 to NL3 on the incident surface of the image sensor 12. In this case, the intensity of the noise light NL1 to NL3 incident on the image sensor 12 can be reduced compared to the conventional structure in which the center of the incident surface of the image sensor 12 intersects with the optical axis of the noise light NL1 to NL3. As a result, the influence of the noise light NL1 to NL3 on the electric signal generated by the image sensor 12 can be suppressed. It can be said that the same effect is achieved when the condition that “at least a part of the noise light NL1 to NL3 does not enter the incident surface of the image sensor 12 ” is set.

[0038] In addition, in the present embodiment, a configuration is adopted for all the photoelectric conversion units constituting the image sensor 12 so that the noise light NL1 to NL3 is not incident with an intensity exceeding the detection limit, but the present invention is not limited to this. For example, a configuration may be adopted for the photoelectric conversion unit constituting the image sensor 12 to which the signal light SL3 is incident with an intensity exceeding the detection limit so that the noise light NL1 to NL3 is not incident with an intensity exceeding the detection limit. In this case, the influence of the noise light NL1 to NL3 on the electrical signal generated by the image sensor 12 can be made smaller than that of the conventional optical computing device.

[0039] In addition, in the present embodiment, a structure is adopted in which the signal light SL0 subjected to two-dimensional intensity modulation according to the input data is input to the optical operation unit 11, but the present invention is not limited thereto. That is, the following structure may be adopted, that is, the carrier light is input to the optical operation unit 11, and the signal light SL1 subjected to two-dimensional modulation according to the input data is generated by the optical modulation element 11a1. In addition, the optical operation device 1 may further include: a display that generates the signal light SL0, or a light source that generates the carrier light.

[0040] In addition, for example, machine learning can be used to achieve: setting of the phase modulation amount of each unit of each light modulator 11ai. As a model used in the machine learning, for example, a model can be used that takes the two-dimensional intensity distribution of the signal light SL0 incident on the light modulator 11a1 as input and takes the two-dimensional intensity distribution of the signal light SL3 incident on the image sensor 12 as output, that is, a model that includes the phase modulation amount in each unit of each light modulator 11ai as a parameter. Here, the two-dimensional intensity distribution of the signal light SL0 incident on the light modulator 11a1, for example, refers to: the set of intensities of the signal light SL0 incident on each unit constituting the light modulator 11a1. In addition, the two-dimensional intensity distribution of the signal light SL3 incident on the image sensor 12, for example, refers to: the set of intensities of the signal light SL3 incident on each photoelectric conversion unit constituting the image sensor 12.

[0041] When the optical computing unit 11 includes the optical modulator 11ai whose phase modulation amount of each cell is variable, the optical computing device 1 may include a control unit that sets the phase modulation amount of each cell of the optical modulator 11ai. This allows the content of the optical computing performed in the optical computing device 1 to be changed.

[0042] Furthermore, when n light diffraction layers formed in a dried gel are used as the light modulators 11a1 to 11an, it is preferable to use a gel that shrinks while maintaining a similar shape by dehydration, such as a gel used in an implosion fabrication method. Thus, by drying a swollen gel on which n light diffraction layers are formed, it is possible to easily manufacture the optical computing unit 11 on which n light diffraction layers are arranged with high precision.

[0043] (Modification of the optical computing device)

[0044] Reference Figure 2 A modification of the optical computing device 1 will be described. Figure 2 It is a side view showing the structure of an optical computation device 1 (hereinafter referred to as an optical computation device 1A) according to this modification.

[0045] Similar to the optical computation device 1 , the optical computation device 1A includes an optical computation unit 11 and an image sensor 12 . The optical computation device 1A differs from the optical computation device 1 in the positions of the light modulation elements 11 a 1 to 11 a 3 and the image sensor 12 .

[0046] In this variation, the signal light SL0 is incident obliquely on the incident surface of the first light modulator 11a1. The first light modulator 11a1 causes the noise light NL1 and the signal light SL1 to be emitted obliquely in different directions. The second light modulator 11a2 is configured such that (1) its incident surface is parallel to the emission surface of the first light modulator 11a1, (2) its incident surface intersects with the signal light SL1 emitted from the first light modulator 11a1, and (3) its incident surface does not intersect with the optical axis of the noise light NL1 emitted from the first light modulator 11a1. The second light modulator 11a2 causes the noise light NL2 and the signal light SL2 to be emitted obliquely in different directions. The third light modulator 11a3 is configured such that (1) its incident surface is parallel to the exit surface of the second light modulator 11a2, (2) its incident surface intersects with the signal light SL2 emitted from the second light modulator 11a2, and (3) its incident surface does not intersect with the optical axis of the noise light NL2 emitted from the second light modulator 11a2. The third light modulator 11a3 causes the noise light NL3 and the signal light SL3 to be emitted obliquely in different directions. The image sensor 12 is configured such that (1) its incident surface is parallel to the exit surface of the third light modulator 11a3, (2) its incident surface intersects with the optical axis of the signal light SL3 emitted from the optical calculation unit 11, and (3) its incident surface does not intersect with the optical axes of the noise lights NL1 to NL3 emitted from the optical calculation unit 11. In addition, the optical axes of the signal lights SL0 to SL3 and the noise lights NL1 to NL3 are located in the same plane.

[0047] By configuring the optical modulators 11a1 to 11a3 and the image sensor 12 in the above manner, compared with a conventional optical computing device in which the signal light SLi and the noise light NLi are emitted from each optical modulator 11ai in the same direction, the noise light NL1 to NL3 has less influence on the electrical signal generated by the image sensor 12.

[0048] (Structural Example of Transmissive Light Modulation Element)

[0049] Reference Figure 3 A configuration example of the light modulation element 11 ai included in the optical computation device 1 will be described. Figure 3 (a) is a top view of the light modulation element 11ai of this specific example. Figure 3 (b) is a cross-sectional view of a cell C constituting the light modulation element 11 ai of this specific example.

[0050] like Figure 3As shown in (a), the optical modulator 11ai is composed of a plurality of cells C whose phase modulation amounts are independently set. When the signal light SLi-1 enters the optical modulator 11ai, the signal lights SLi-1 transmitted through each cell C interfere with each other while being phase modulated, thereby forming a signal light SLi representing the calculation result.

[0051] The cell C constituting the light modulating element 11ai is a microcell. Here, a "microcell" refers to a cell having a cell size of less than 10 μm. In addition, a "cell size" refers to the square root of the area of ​​the cell. For example, when the top view shape of the microcell is a square, the cell size of the microcell refers to the length of one side of the microcell. The lower limit of the cell size of the microcell is, for example, 1 nm.

[0052] Figure 3 The light modulator 11ai illustrated in (a) is composed of 200×200 cells C arranged in a matrix. The top view shape of each cell C is a square of 500nm×500nm, and the top view shape of the light modulator 11ai is a square of 100μm×100μm. In this case, the signal light SLi-1 that is not phase modulated by the cell C and passes through the gaps between the cells C or the peripheral parts of the cells C is noise light.

[0053] like Figure 3 As shown in (b), each cell C constituting the light modulation element 11 ai can be composed of, for example, a polarizing plate C11 , a polarizing plate C12 , a first electrode C13 , a magnetization free layer C14 , an insulating layer C15 , a magnetization fixed layer C16 , and a second electrode C17 .

[0054] The polarizing plate C11 and the polarizing plate C12 are arranged in a mutually opposed manner. The first electrode C13, the magnetization free layer C14, the insulating layer C15, the magnetization fixed layer C16 and the second electrode C17 are stacked in sequence and sandwiched between the polarizing plate C11 and the polarizing plate C12. Here, the stacking direction of the first electrode C13, the magnetization free layer C14, the insulating layer C15, the magnetization fixed layer C16 and the second electrode C17 is orthogonal to the stacking direction of the polarizing plate C11 and the reflector C12. Therefore, the first side surface of the magnetization free layer C14 is in surface contact with one main surface of the polarizing plate C11, and the second side surface of the magnetization free layer C14 opposite to the first side surface is in surface contact with one main surface of the reflector C12. The signal light SLi- 1 (1) enters the interior of the magnetization free layer C14 via the polarization plate C11 , (2) passes through the magnetization free layer C14 , and (3) exits the exterior of the magnetization free layer C14 via the polarization plate C12 .

[0055] The magnetization free layer C14 is composed of, for example, a soft magnetic material having electrical conductivity and light transmittance (e.g., CoFeB). In addition, the magnetization fixed layer C16 is composed of, for example, a hard magnetic material having electrical conductivity (e.g., Permalloy). In addition, as the polarizer C11 and the polarizer C12, a polarizer that selectively transmits a polarization component whose polarization direction P is parallel to the magnetization direction M of the magnetization fixed layer C16 can be selected. Figure 3 (b) shows an example in which the magnetization direction M and the polarization direction P are parallel to both the principal surface of the polarizing plate C11 and the principal surface of the magnetization fixed layer C16 .

[0056] When a potential difference is applied between the first electrode C13 and the second electrode C17, a spin current (the flow of spin-polarized electrons) is injected from the magnetization fixed layer C16 to the magnetization free layer C14 via the insulating layer C15 due to the tunnel effect, and magnetization occurs in the magnetization free layer C14. Here, the magnetization occurring in the magnetization free layer C14 is parallel to the magnetization direction M of the magnetization fixed layer C16, that is, parallel to the polarization direction P of the signal light SLi-1 incident on the magnetization free layer C14 via the polarizing plate C11. Therefore, in the process of the signal light SLi-1 propagating in the magnetization free layer C14, the phase of the signal light SLi-1 is delayed due to the transverse Kerr effect.

[0057] Here, the phase change amount of the signal light SLi-1 in the cell C is determined according to the magnitude of the magnetization generated in the magnetization free layer C14. In addition, the magnitude of the magnetization generated in the magnetization free layer C14 is determined according to the magnitude of the spin current injected into the magnetization free layer C14. In addition, the magnitude of the spin current injected into the magnetization free layer C14 is determined according to the potential difference applied between the first electrode C13 and the second electrode C17. Therefore, by controlling the potential difference applied between the first electrode C13 and the second electrode C17, the phase modulation amount of the cell C can be set to a desired value.

[0058] In addition, in this structural example, a cell C having the same structure as an MRAM (Magnetoresistive Random Access Memory) of an STT (Spin Transfer Torque) method is described, but the present invention is not limited thereto. For example, a cell C having the same structure as an MRAM of an SOT (Spin Orbit Torque) method may also be used. In addition, such a cell C can be obtained, for example, by Figure 3The structure shown in (b) is realized by removing the insulating layer C15, the magnetization fixed layer C16, and the second electrode C17. In this case, for example, by including a heavy metal in the first electrode C13 and applying a pulse voltage or a pulse current to the first electrode C13, a spin current can be efficiently injected into the magnetization free layer C14.

[0059] Alternatively, the unit C may be configured similarly to a unit of an LCD (Liquid Crystal Display). In this case, Figure 4 As shown, the cell C is composed of, for example, a first glass substrate C21a, a first transparent electrode C22a, a liquid crystal layer C23, a second transparent electrode C22b, and a second glass substrate C21b. The first glass substrate C21a, the first transparent electrode C22a, the liquid crystal layer C23, the second transparent electrode C22b, and the second glass substrate C21b are stacked in order from the incident side of the signal light SLi-1.

[0060] The signal light SLi-1 (1) passes through the first glass substrate C21a and the first transparent electrode C22a, (2) passes through the liquid crystal layer C23, and (3) passes through the second transparent electrode C22b and the second glass substrate C21b. The liquid crystal layer C23 is composed of liquid crystal molecules oriented in a direction parallel to the main surface of the glass substrate C21, and has a refractive index corresponding to the potential difference between the first transparent electrode C22a and the second transparent electrode C22b. Therefore, the signal light SLi-1 is phase modulated when passing through the liquid crystal layer C23. By controlling the potential difference applied between the first transparent electrode C22a and the second transparent electrode C22b, the phase modulation amount of the unit C can be set to a desired value.

[0061] Furthermore, the light modulation element 11ai can also be composed of a plurality of cells C having thicknesses independently set from each other. Figure 5 It is a perspective view showing an enlarged portion of the light modulation element 11 ai composed of a plurality of cells C having thicknesses independently set.

[0062] Figure 5 The light modulator 11ai shown is composed of a transparent substrate C31 and a plurality of columns formed on the upper surface of the transparent substrate C31. Each column is a quadrangular prism-shaped structure having a square bottom surface with each side having the same length as the unit size, and functions as a unit C. In this case, the signal light SLi-1 that is not phase modulated by the unit C and is transmitted between the columns constituting each unit C (the peripheral portion of each unit where no column is formed) is noise light.

[0063] The signal light SLi (1) enters from the upper surface of the column, (2) passes through the column, (3) passes through the transparent substrate C31, and (4) exits from the lower surface of the transparent substrate C31. The phase modulation amount of the signal light SLi passing through each unit C is determined by the height of the column constituting each unit C. That is, the phase modulation amount of the signal light SLi passing through the unit C composed of the column with a higher height is larger, and the phase modulation amount of the signal light SLi passing through the unit C composed of the column with a lower height is smaller. In addition, the phase modulation amount of each unit C is fixed.

[0064] [Second embodiment]

[0065] (Structure of optical computing device)

[0066] Reference Figure 6 The structure of an optical computing device 2 according to a second embodiment of the present invention will be described. Figure 6 It is a side view showing the structure of the optical computing device 2 .

[0067] like Figure 6 As shown, the optical computation device 2 includes an optical computation unit 21 , an image sensor 22 , and a reflection mirror 23 .

[0068] The optical operation unit 21 is a set of at least one optical modulator 21a1 to 21an (n is any natural number greater than 1). In this embodiment, in order to realize multi-stage optical operations, a set of three optical modulators 21a1 to 21a3 is used as the optical operation unit 21. In addition, the optical modulators 21a1 to 21an can be integrated. For example, n light diffraction layers formed in a structure that transmits signal light such as a dry gel can be used as optical modulators 21a1 to 21an. Alternatively, the optical modulators 21a1 to 21an can be embedded in a single substrate, or n regions of a single optical modulator can be used as optical modulators 21a1 to 21an.

[0069] Each optical modulator 21ai (i is a natural number greater than 1 and less than n) is composed of a plurality of units that can set the phase modulation amount independently of each other. Each optical modulator 21ai is a reflective optical modulator that causes the reflected light phase-modulated by each unit to interfere with each other to perform optical calculations. Here, performing optical calculations means converting the two-dimensional intensity distribution of the signal light from a two-dimensional intensity distribution representing information before the calculation to a two-dimensional intensity distribution representing information after the calculation. The structural example of the reflective optical modulator will be described later with reference to different drawings.

[0070] In addition to emitting the above-mentioned signal light, the optical modulator 21ai also emits noise light that has not been phase-modulated by each unit and has passed through the optical modulator 21ai. For example, light that has not been incident on the inside of the unit constituting the optical modulator 21ai and has been reflected by the surface of the optical modulator 21ai is noise light. The optical axis of the noise light emitted from the optical modulator 21ai is located in the plane where the optical axis of the signal light incident on the optical modulator 21ai and the normal of the incident surface of the optical modulator 21ai are located, and the emission angle of the noise light emitted from the optical modulator 21ai is the same as the incident angle of the signal light incident on the optical modulator 21ai. That is, the optical axis of the noise light is consistent with the optical axis of the reflected light when the signal light is positively reflected by the optical modulator 21ai.

[0071] In order to separate such noise light from the signal light, the phase modulation amount of each unit constituting the optical modulator 21ai is set so that the direction in which the signal light is emitted from the optical modulator 21ai is different from the direction in which the noise light is emitted from the optical modulator 21ai. In a reflective optical modulator, for example, the emission angle of the noise light is the same as the incident angle of the incident light. Therefore, the phase modulation amount of each unit is set so that the emission angle of the signal light is different from the incident angle of the incident light.

[0072] The reflective mirror 23 is used to reflect the signal lights SL1 to SL3 emitted from the light modulator elements 21 a 1 to 21 a 3 , and is arranged such that its reflective surface faces the incident and emitted surfaces of the light modulator elements 21 a 1 to 21 a 3 .

[0073] In the present embodiment, the signal light SL0 is incident obliquely on the first light modulator 21a1. The first light modulator 21a1 causes the noise light NL1 and the signal light SL1 to be emitted obliquely in different directions. The noise light NL1 and the signal light SL1 emitted from the first light modulator 21a1 are respectively reflected by the reflector 23. The second light modulator 21a2 is configured such that (1) its incident and incident emission surfaces are located on the same plane as the incident and incident emission surfaces of the first light modulator 21a1, (2) its incident and incident emission surfaces intersect with the signal light SL1 reflected by the reflector 23, and (3) its incident and incident emission surfaces do not intersect with the optical axis of the noise light NL1 reflected by the reflector 23. In addition, instead of condition (3), the following condition may be set, that is, the optical axis of the noise light NL1 emitted from the first light modulator 21a1 does not intersect with the intersection point where the incident and incident emission surfaces of the second light modulator 21a2 intersect with the signal light SL1 emitted from the first light modulator 21a1. Alternatively, a condition may be set such that at least a part of the noise light NL1 emitted from the first light modulation element 21 a 1 does not enter the incident / exit surface of the second light modulation element 21 a 2 .

[0074] In addition, in the present embodiment, the second light modulator 21a2 causes the noise light NL2 and the signal light SL2 to be emitted obliquely in different directions. The noise light NL2 and the signal light SL2 emitted from the second light modulator 21a2 are respectively reflected by the reflector 23. The third light modulator 21a3 is configured such that (1) its incident and output surfaces are located on the same plane as the incident and output surfaces of the second light modulator 21a2, (2) its incident and output surfaces intersect with the signal light SL2 reflected by the reflector 23, and (3) its incident and output surfaces do not intersect with the optical axis of the noise light NL2 reflected by the reflector 23. In addition, the following condition may be set instead of condition (3), namely: the optical axis of the noise light NL2 emitted from the second light modulator 21a2 does not intersect with the intersection point where the incident and output surfaces of the third light modulator 21a3 intersect with the signal light SL2 emitted from the second light modulator 21a2. Alternatively, the following condition may be set: at least a portion of the noise light NL2 emitted from the second light modulator 21a2 does not enter the incident surface of the third light modulator 21a3. The third light modulator 21a3 causes the noise light NL3 and the signal light SL3 to be emitted in different directions. In addition, the optical axes of the signal lights SL0 to SL3 and the noise lights NL1 to NL3 are located in the same plane.

[0075] The image sensor 22 is an example of an optical sensor, and is used to detect the signal light SL3 output from the optical calculation unit 21 and generate an electrical signal indicating the detection result. The image sensor 22 is composed of a plurality of photoelectric conversion units to generate an electrical signal indicating the intensity distribution of the signal light SL3 emitted from the optical calculation unit 21. The image sensor 22 is configured such that (1) its incident surface and the reflection surface of the reflector 23 are located on the same plane, (2) its incident surface intersects with the optical axis of the signal light SL3 emitted from the optical calculation unit 21, and (3) its incident surface does not intersect with the optical axes of the noise lights NL1 to NL3 emitted from the optical calculation unit 21.

[0076] By configuring the light modulators 21a1 to 21a3, the image sensor 22, and the reflector 23 in the above manner, it is possible to avoid the situation where the noise light NL1 to NL3 is incident on the photoelectric conversion unit with an intensity exceeding the detection limit of the photoelectric conversion unit for all the photoelectric conversion units constituting the image sensor 22. Therefore, compared with the conventional optical computing device in which the signal light SLi and the noise light NLi are emitted in the same direction from each light modulator 21ai, the influence of the noise light NL1 to NL3 on the electrical signal generated by the image sensor 22 is small. Instead of the condition that "the incident surface (any point) of the image sensor 22 does not intersect with the optical axis of the noise light NL1 to NL3", the following condition may be set: the point intersecting with the optical axis of the signal light SL1 to SL3 on the incident surface of the image sensor 22 is different from the point intersecting with the optical axis of the noise light NL1 to NL3 on the incident surface of the image sensor 22 (relatively weak condition). In this case, the intensity of the noise light NL1 to NL3 incident on the image sensor 22 can be reduced compared to the conventional structure in which the center of the incident surface of the image sensor 22 intersects with the optical axis of the noise light NL1 to NL3. As a result, the influence of the noise light NL1 to NL3 on the electrical signal generated by the image sensor 22 can be suppressed. When the condition that "at least a part of the noise light NL1 to NL3 is not incident on the incident surface of the image sensor 22" is set, it can be said that the same effect is achieved.

[0077] In addition, in the present embodiment, a configuration is adopted for all the photoelectric conversion units constituting the image sensor 22 so that the noise light NL1 to NL3 is not incident with an intensity exceeding the detection limit, but the present invention is not limited thereto. For example, a configuration may be adopted for the photoelectric conversion unit constituting the image sensor 22 to which the signal light SL3 is incident with an intensity exceeding the detection limit so that the noise light NL1 to NL3 is not incident with an intensity exceeding the detection limit. In this case, the influence of the noise light NL1 to NL3 on the electrical signal generated by the image sensor 22 can be made smaller than that of the conventional optical computing device.

[0078] In addition, in the present embodiment, a structure is adopted in which the signal light SL0 subjected to two-dimensional intensity modulation according to the input data is input to the optical operation unit 21, but the present invention is not limited thereto. That is, the following structure may be adopted, that is, the carrier light is input to the optical operation unit 21, and the signal light SL1 subjected to two-dimensional modulation according to the input data is generated by the light modulator 21a1. In addition, the optical operation device 2 may further include: a display that generates the signal light SL0, or a light source that generates the carrier light.

[0079] In addition, for example, machine learning can be used to achieve: setting of the phase modulation amount of each unit of each light modulator 21ai. As a model used in the machine learning, for example, a model can be used that takes the two-dimensional intensity distribution of the signal light SL0 incident on the light modulator 21a1 as input and the two-dimensional intensity distribution of the signal light SL3 incident on the image sensor 22 as output, that is, a model that includes the phase modulation amount in each unit of each light modulator 21ai as a parameter. Here, the two-dimensional intensity distribution of the signal light SL0 incident on the light modulator 21a1, for example, refers to: the set of intensities of the signal light SL0 incident on each unit constituting the light modulator 21a1. In addition, the two-dimensional intensity distribution of the signal light SL3 incident on the image sensor 22, for example, refers to: the set of intensities of the signal light SL3 incident on each photoelectric conversion unit constituting the image sensor 22.

[0080] When the optical computing unit 21 includes the optical modulator 21ai whose phase modulation amount of each cell is variable, the optical computing device 2 may include a control unit for setting the phase modulation amount of each cell of the optical modulator 21ai. This allows the content of the optical computing performed in the optical computing device 1 to be changed.

[0081] In addition, as described above, n regions of a single optical modulator can be used as the optical modulators 21a1 to 21an. That is, n operation regions A1, A2, ..., AN can be set in a single optical modulator, and the operation region A1 performs optical operation by modulating and reflecting incident light, and each operation region Ai (i is a natural number greater than 2 and less than N) other than the operation region A1 performs optical operation by modulating and reflecting the signal light reflected by the reflector 23 after being modulated and reflected by the operation region Ai-1.

[0082] Furthermore, when n light diffraction layers formed in a dried gel are used as the light modulators 21a1 to 21an, it is preferable to use a gel that shrinks while maintaining a similar shape by syneresis, such as a gel used in an implosion fabrication method. Thus, by drying a swollen gel on which n light diffraction layers are formed, it is possible to easily manufacture the optical computing unit 21 on which n light diffraction layers are arranged with high precision.

[0083] (Variation 1 of Optical Computing Device)

[0084] Reference Figure 7 A first modification of the optical computation device 2 will be described. Figure 7 It is a side view showing the structure of an optical computation device 2 (hereinafter referred to as an optical computation device 2A) according to this modification.

[0085] Similar to the optical computing device 2, the optical computing device 2A includes an optical computing unit 21, an image sensor 22, and a reflector 23. The optical computing device 2A differs from the optical computing device 2 in the positions of the light modulators 21a1 to 21a3, the position and orientation of the image sensor 22, and the orientation of the reflector 23.

[0086] In the optical computing device 2, the reflector 23 is configured so that its reflection surface is parallel to the incident and output surfaces of the light modulators 21a1 to 21a3. In contrast, in the optical computing device 2A, the reflector 23 is configured so that its reflection surface is not parallel to the incident and output surfaces of the light modulators 21a1 to 21a3. In addition, in the optical computing device 2, the image sensor 22 is configured so that its incident surface is parallel to the incident and output surfaces of the light modulators 21a1 to 21a3. In contrast, in the optical computing device 2A, the image sensor 22 is configured so that its incident surface is not parallel to the incident and output surfaces of the light modulators 21a1 to 21a3. In addition, a point common to the optical computing device 2 and the optical computing device 2A is that the incident surface of the image sensor 22 and the reflection surface of the reflector 23 are located on the same plane.

[0087] In this variation, the signal light SL0 is also incident obliquely on the first light modulator 21a1. The first light modulator 21a1 causes the noise light NL1 and the signal light SL1 to be emitted obliquely in different directions. The noise light NL1 and the signal light SL1 emitted from the first light modulator 21a1 are respectively reflected by the reflector 23. The second light modulator 21a2 is configured such that (1) its incident and incident surfaces are located on the same plane as the incident and incident surfaces of the first light modulator 21a1, (2) its incident and incident surfaces intersect with the signal light SL1 reflected by the reflector 23, and (3) its incident and incident surfaces do not intersect with the optical axis of the noise light NL1 reflected by the reflector 23. The second light modulator 21a2 causes the noise light NL2 and the signal light SL2 to be emitted obliquely in different directions. The noise light NL2 and the signal light SL2 emitted from the second light modulator 21a2 are respectively reflected by the reflector 23. The third light modulator 21a3 is configured such that (1) its incident and incident surfaces are located on the same plane as the incident and incident surfaces of the second light modulator 21a2, (2) its incident and incident surfaces intersect with the signal light SL2 reflected by the reflector 23, and (3) its incident and incident surfaces do not intersect with the optical axis of the noise light NL2 reflected by the reflector 23. The third light modulator 21a3 causes the noise light NL3 and the signal light SL3 to be emitted obliquely in different directions. The image sensor 22 is configured such that (1) its incident surface is located on the same plane as the reflection surface of the reflector 23, (2) its incident surface intersects with the optical axis of the signal light SL3 emitted from the optical calculation unit 21, and (3) its incident surface does not intersect with the optical axes of the noise lights NL1 to NL3 emitted from the optical calculation unit 21. In addition, the optical axes of the signal lights SL0 to SL3 and the noise lights NL1 to NL3 are located in the same plane.

[0088] By configuring the light modulators 21a1 to 21a3, the image sensor 22, and the reflector 23 in the above manner, it is possible to avoid the situation where the noise light NL1 to NL3 is incident on the photoelectric conversion unit with an intensity exceeding the detection limit of the photoelectric conversion unit for all the photoelectric conversion units constituting the image sensor 22. Therefore, compared with the existing optical computing device in which the signal light SLi and the noise light NLi are emitted from each light modulator 21ai in the same direction, the influence of the noise light NL1 to NL3 on the electrical signal generated by the image sensor 22 is small. In addition, in this modified example, the incident angle of the signal light SL0 to SL2 incident on the light modulators 21a1 to 21a3 and the emission angle of the signal light SL1 to SL3 emitted from the light modulators 21a1 to 21a3 can be made common. Therefore, it is easy to design and manufacture the light modulators 21a1 to 21a3.

[0089] (Variation 2 of Optical Computing Device)

[0090] Reference Figure 8 A second modification of the optical computing device 2 is described. Figure 8 In the figure, (a) is a top view of the optical computing device 2 (hereinafter referred to as the optical computing device 2B) of this modification, (b) is a side view of the optical computing device 2B, and (c) is a front view of the optical computing device 2B. Figure 8 In (a), the reflection mirror 23 is omitted from illustration.

[0091] Similar to the optical computing device 2, the optical computing device 2B includes an optical computing unit 21, an image sensor 22, and a reflector 23. The first difference between the optical computing device 2A and the optical computing device 2 is the positions of the light modulators 21a1 to 21a3, the position and orientation of the image sensor 22, and the position and orientation of the reflector 23. The second difference between the optical computing device 2A and the optical computing device 2 is the emission direction of the signal light SLi emitted from each light modulator 21ai.

[0092] In the optical computing device 2, the reflector 23 is configured so that its reflection surface is parallel to the incident and outgoing surfaces of the light modulators 21a1 to 21a3. In contrast, in the optical computing device 2B, the reflector 23 is configured so that its reflection surface is not parallel to the incident and outgoing surfaces of the light modulators 21a1 to 21a3. In addition, in the optical computing device 2, the image sensor 22 is configured so that its incident surface and the reflection surface of the reflector 23 are located on the same plane. In contrast, in the optical computing device 2B, the image sensor 22 is configured so that its incident surface and the incident and outgoing surfaces of the light modulators 21a1 to 21a3 are located on the same plane.

[0093] In addition, in the optical computing device 2, the optical axis of the signal light SLi emitted from each optical modulator 21ai is arranged in a plane including the optical axis of the signal light SLi-1 incident on the optical modulator 21ai and the optical axis of the noise light NLi emitted from the optical modulator 21ai. In contrast, in the optical computing device 2, the optical axis of the signal light SLi emitted from each optical modulator 21ai is arranged outside the plane including the optical axis of the signal light SLi-1 incident on the optical modulator 21ai and the optical axis of the noise light NLi emitted from the optical modulator 21ai.

[0094] In this variation, the signal light SL0 is also incident obliquely on the first light modulator 21a1. The first light modulator 21a1 causes the noise light NL1 and the signal light SL1 to be emitted obliquely in different directions. The noise light NL1 and the signal light SL1 emitted from the first light modulator 21a1 are respectively reflected by the reflector 23. The second light modulator 21a2 is configured such that (1) its incident and incident surfaces are located on the same plane as the incident and incident surfaces of the first light modulator 21a1, (2) its incident and incident surfaces intersect with the signal light SL1 reflected by the reflector 23, and (3) its incident and incident surfaces do not intersect with the optical axis of the noise light NL1 reflected by the reflector 23. The second light modulator 21a2 causes the noise light NL2 and the signal light SL2 to be emitted obliquely in different directions. The noise light NL2 and the signal light SL2 emitted from the second light modulator 21a2 are respectively reflected by the reflector 23. The third light modulator 21a3 is configured such that (1) its incident and incident surfaces are located on the same plane as the incident and incident surfaces of the second light modulator 21a2, (2) its incident and incident surfaces intersect with the signal light SL2 reflected by the reflector 23, and (3) its incident and incident surfaces do not intersect with the optical axis of the noise light NL2 reflected by the reflector 23. The third light modulator 21a3 causes the noise light NL3 and the signal light SL3 to be emitted obliquely in different directions. The noise light NL3 and the signal light SL3 emitted from the third light modulator 21a3 are respectively reflected by the reflector 23. The image sensor 22 is configured such that (1) its incident surface is located on the same plane as the reflection surface of the reflector 23, (2) its incident surface intersects with the optical axis of the signal light SL3 reflected by the reflector 23, and (3) its incident surface does not intersect with the optical axes of the noise lights NL1 to NL3 reflected by the reflector 23.

[0095] By configuring the light modulators 21a1 to 21a3, the image sensor 22, and the reflector 23 in the above manner, it is possible to avoid the situation where the noise light NL1 to NL3 is incident on the photoelectric conversion unit with an intensity exceeding the detection limit of the photoelectric conversion unit for all the photoelectric conversion units constituting the image sensor 22. Therefore, compared with the existing optical computing device in which the signal light SLi and the noise light NLi are emitted from each light modulator 21ai in the same direction, the influence of the noise light NL1 to NL3 on the electrical signal generated by the image sensor 22 is small. In addition, in this modified example, the incident angle of the signal light SL0 to SL2 incident on the light modulators 21a1 to 21a3 and the emission angle of the signal light SL1 to SL3 emitted from the light modulators 21a1 to 21a3 can be made common. Therefore, it is easy to design and manufacture the light modulators 21a1 to 21a3.

[0096] (Structural Example of Reflective Light Modulation Element)

[0097] Reference Fig. 9 A configuration example of the reflective light modulation element 21 ai included in the optical computation device 1 will be described. Fig. 9 (a) is a top view of the light modulation element 21ai of this specific example. Fig. 9 (b) is a cross-sectional view of a cell C constituting the light modulation element 21ai of this specific example.

[0098] like Fig. 9 As shown in (a), the optical modulator 21ai is composed of a plurality of cells C whose phase modulation amounts are independently set. When the signal light SLi-1 enters the optical modulator 21ai, the signal lights SLi-1 reflected by the cells C interfere with each other while being phase modulated, thereby forming a signal light SLi representing the calculation result.

[0099] The cell C constituting the light modulator 21ai is a microcell. Here, a "microcell" refers to a cell having a cell size of less than 10 μm. In addition, a "cell size" refers to the square root of the area of ​​the cell. For example, when the top view shape of the microcell is a square, the cell size of the microcell refers to the length of one side of the microcell. The lower limit of the cell size of the microcell is, for example, 1 nm.

[0100] Fig. 9 The light modulation element 21 ai illustrated in (a) is composed of 200×200 cells C arranged in a matrix. The plan view shape of each cell C is a square of 500 nm×500 nm, and the plan view shape of the light modulation element 21 ai is a square of 100 μm×100 μm.

[0101] like Fig. 9As shown in (b), each unit C constituting the light modulator 21ai can be composed of, for example, a polarizing plate C11, a reflecting plate C18, a first electrode C13, a magnetization free layer C14, an insulating layer C15, a magnetization fixed layer C16, and a second electrode C17. In this case, the signal light SLi-1 reflected by the surface of the polarizing plate C11 is noise light.

[0102] The polarizing plate C11 and the reflecting plate C18 are arranged in a mutually opposed manner. The first electrode C13, the magnetization free layer C14, the insulating layer C15, the magnetization fixed layer C16, and the second electrode C17 are stacked in sequence and sandwiched between the polarizing plate C11 and the reflecting plate C18. Here, the stacking direction of the first electrode C13, the magnetization free layer C14, the insulating layer C15, the magnetization fixed layer C16, and the second electrode C17 is orthogonal to the stacking direction of the polarizing plate C11 and the reflecting plate C18. Therefore, the first side surface of the magnetization free layer C14 is in surface contact with one main surface of the polarizing plate C11, and the second side surface of the magnetization free layer C14 opposite to the first side surface is in surface contact with one main surface of the reflecting plate C18. Regarding the signal light SLi-1, (1) it is incident on the inside of the magnetization free layer C14 via the polarizing plate C11, (2) it is reflected by the reflecting plate C18, and (3) it is emitted to the outside of the magnetization free layer C14 via the polarizing plate C11.

[0103] The magnetization free layer C14 is composed of, for example, a soft magnetic material having electrical conductivity and light transmittance (e.g., CoFeB). In addition, the magnetization fixed layer C16 is composed of, for example, a hard magnetic material having electrical conductivity (e.g., Permalloy). In addition, as the polarizer C11, a polarizer that selectively transmits a polarization component whose polarization direction P is parallel to the magnetization direction M of the magnetization fixed layer C16 can be selected. Fig. 9 (b) shows an example in which the magnetization direction M and the polarization direction P are parallel to both the principal surface of the polarizing plate C11 and the principal surface of the magnetization fixed layer C16 .

[0104] When a potential difference is applied between the first electrode C13 and the second electrode C17, a spin current (the flow of spin-polarized electrons) is injected from the magnetization fixed layer C16 to the magnetization free layer C14 via the insulating layer C15 due to the tunnel effect, and magnetization occurs in the magnetization free layer C14. Here, the magnetization occurring in the magnetization free layer C14 is parallel to the magnetization direction M of the magnetization fixed layer C16, that is, parallel to the polarization direction P of the signal light SLi-1 incident on the magnetization free layer C14 via the polarizing plate C11. Therefore, in the process of the signal light SLi-1 propagating in the magnetization free layer C14, the phase of the signal light SLi-1 is delayed due to the transverse Kerr effect.

[0105] Here, the phase change amount of the signal light SLi-1 in the cell C is determined according to the magnitude of the magnetization generated in the magnetization free layer C14. In addition, the magnitude of the magnetization generated in the magnetization free layer C14 is determined according to the magnitude of the spin current injected into the magnetization free layer C14. In addition, the magnitude of the spin current injected into the magnetization free layer C14 is determined according to the potential difference applied between the first electrode C13 and the second electrode C17. Therefore, by controlling the potential difference applied between the first electrode C13 and the second electrode C17, the phase modulation of the cell C can be set to a desired value.

[0106] In addition, in this structural example, a cell C having the same structure as an MRAM (Magnetoresistive Random Access Memory) of an STT (Spin Transfer Torque) method is described, but the present invention is not limited thereto. For example, a cell C having the same structure as an MRAM of an SOT (Spin Orbit Torque) method may also be used. In addition, such a cell C can be obtained, for example, by Fig. 9 The structure shown in (b) is realized by removing the insulating layer C15, the magnetization fixed layer C16, and the second electrode C17. In this case, for example, by including a heavy metal in the first electrode C13 and applying a pulse voltage or a pulse current to the first electrode C13, a spin current can be efficiently injected into the magnetization free layer C14.

[0107] Alternatively, the cell C may be constructed in the same manner as a cell of LCOS (Liquid Crystal On Silicon). Fig.10 As shown, the cell C is composed of, for example, a glass substrate C21, a transparent electrode C22, a liquid crystal layer C23, and a reflective electrode C24. The glass substrate C21, the transparent electrode C22, the liquid crystal layer C23, and the reflective electrode C24 are stacked in order from the incident side of the signal light SLi-1.

[0108] Regarding the signal light SLi-1, (1) it passes through the glass substrate C21 and the transparent electrode C22, (2) it passes through the liquid crystal layer C23, (3) it is reflected by the reflective electrode C24, (4) it passes through the liquid crystal layer C23, and (5) it passes through the transparent electrode C22 and the glass substrate C21. The liquid crystal layer C23 is composed of liquid crystal molecules oriented in a direction parallel to the main surface of the glass substrate C21, and has a refractive index corresponding to the potential difference between the transparent electrode C22 and the reflective electrode C24. Therefore, the signal light SLi-1 is phase modulated when passing through the liquid crystal layer C23. By controlling the potential difference applied between the transparent electrode C22 and the reflective electrode C24, the phase modulation amount of the unit C can be set to a desired value.

[0109] Furthermore, the reflective light modulation element 21ai can also be composed of a plurality of cells C having thicknesses independently set. Fig.11 It is a perspective view showing an enlarged portion of the light modulation element 21ai composed of a plurality of cells C having thicknesses independently set.

[0110] Fig.11 The light modulator 21ai shown is composed of a reflector C32 and a plurality of columns formed on the upper surface of the reflector C32. Each column is a quadrangular prism-shaped structure having a square bottom surface with each side having the same length as the unit size, and functions as a unit C. In this case, in the area where the column is formed, the signal light SLi-1 reflected by the upper surface of the column is noise light. In addition, in the area where the column is not formed, the signal light SLi-1 reflected by the upper surface of the reflector 32 is noise light.

[0111] Regarding the signal light SLi-1, (1) it is incident from the upper surface of the column, (2) it is transmitted through the column, (3) it is reflected by the reflector C32, (4) it is transmitted through the column, and (5) it is emitted from the upper surface of the column. The phase modulation amount of the signal light SLi-1 reflected by each unit C is determined according to the height of the column constituting each unit C. That is, the phase modulation amount of the signal light SLi-1 reflected by the unit C composed of the column with a higher height is larger, and the phase modulation amount of the signal light reflected by the unit C composed of the column with a lower height is smaller. In addition, the phase modulation amount of each unit C is fixed.

[0112] [Summarize]

[0113] The optical computing device of the first aspect of the present invention is characterized in that it comprises: an optical computing unit including an optical modulator, the optical modulator having a plurality of units for which phase modulation amounts are independently set or for which phase modulation amounts can be independently set; and an optical sensor for detecting signal light output from the optical computing unit and generating an electrical signal representing a detection result, wherein the optical modulator is (1) a transmissive optical modulator that emits signal light generated by interference of transmitted light that has been phase-modulated by the plurality of units, respectively, and noise light that has not been phase-modulated by the plurality of units and has transmitted the optical modulator, or (2) The optical modulator is a reflective optical modulator, which emits signal light generated by interference of reflected light that has been phase-modulated by the multiple units, and noise light that has not been phase-modulated by the multiple units and has been reflected by the optical modulator. In the optical modulator, the phase modulation amount of each of the multiple units is set so that the direction in which the signal light is emitted from the optical modulator is different from the direction in which the noise light is emitted from the optical modulator. Compared with the case in which the direction in which the signal light is emitted from the optical modulator is the same as the direction in which the noise light is emitted from the optical modulator, the influence of the noise light on the electrical signal is smaller.

[0114] According to the above configuration, it is possible to reduce the influence of noise light on the electric signal generated by the optical sensor.

[0115] In the optical computing device of mode 2 of the present invention, based on the structure of mode 1, the following structure is also adopted, that is, the optical sensor is an image sensor having a plurality of photoelectric conversion units, and the image sensor is configured so that the noise light does not enter the photoelectric conversion unit to which the signal light enters with an intensity above the detection limit with an intensity above the detection limit.

[0116] According to the above configuration, it is possible to further reduce the influence of noise light on the electric signal generated by the image sensor.

[0117] In the optical computing device of mode 3 of the present invention, in addition to the structure of mode 2, the following structure is further adopted, that is, the image sensor is configured so that the noise light does not enter all the photoelectric conversion units with an intensity exceeding the detection limit. Alternatively, in the optical computing device of mode 3 of the present invention, in addition to the structure of mode 1, the following structure is further adopted, that is, the optical sensor is configured so that (1) its incident surface intersects with the optical axis of the signal light emitted from the optical computing unit, and (2) the intersection point where its incident surface intersects with the optical axis of the signal light emitted from the optical computing unit does not intersect with the optical axis of the noise light emitted from the optical computing unit.

[0118] According to the above configuration, it is possible to further reduce the influence of noise light on the electric signal generated by the image sensor.

[0119] In the optical computing device of mode 4 of the present invention, on the basis of the structure of any one of modes 1 to 3, the following structure is further adopted, that is, the optical computing unit at least includes: a first transmission type optical modulator, and a second transmission type optical modulator, as the optical modulator, the second transmission type optical modulator is configured so that (1) its incident surface intersects with the optical axis of the signal light emitted from the first transmission type optical modulator, and (2) the intersection point of its incident surface and the optical axis of the signal light emitted from the first transmission type optical modulator does not intersect with the optical axis of the noise light emitted from the first transmission type optical modulator. In addition, in the optical computing device of mode 5 of the present invention, on the basis of the structure of any one of modes 1 to 3, the following structure is further adopted, that is, the optical computing unit includes: a first transmission type light modulator, and a second transmission type light modulator, as the light modulator, the second transmission type light modulator is configured so that (1) its incident surface intersects with the optical axis of the signal light emitted from the first transmission type light modulator, and (2) at least a part of the noise light emitted from the first transmission type light modulator is not incident on the incident surface of the second transmission type light modulator.

[0120] According to the above configuration, when the light modulation element is a transmissive light modulation element, the influence of noise light on the electric signal generated by the optical sensor can be effectively reduced.

[0121] In the optical modulator of mode 6 of the present invention, on the basis of the structure of any one of modes 1 to 3, the following structure is further adopted, that is, a reflector is also provided, and the optical operation unit at least includes: a first reflective optical modulator, and a second reflective optical modulator, as the optical modulator, the second reflective optical modulator is configured so that (1) its incident and output surfaces intersect with the optical axis of the signal light emitted from the first reflective optical modulator and reflected by the reflector, and (2) the intersection of its incident and output surfaces with the optical axis of the signal light emitted from the first reflective optical modulator and reflected by the reflector does not intersect with the optical axis of the noise light emitted from the first reflective optical modulator and reflected by the reflector or the optical axis of the noise light emitted from the first reflective optical modulator. In addition, in the optical modulation device of mode 7 of the present invention, on the basis of the structure of any one of modes 1 to 3, the following structure is further adopted, that is, a reflector is also provided, and the optical operation unit includes: a first reflective optical modulation element, and a second reflective optical modulation element, as the optical modulation element, the second reflective optical modulation element is configured so that (1) its incident and output surfaces intersect with the optical axis of the signal light emitted from the first reflective optical modulation element and reflected by the reflector, and (2) the noise light emitted from the first reflective optical modulation element and reflected by the reflector, or at least a part of the noise light emitted from the first reflective optical modulation element, is not incident on the incident and output surfaces of the second reflective optical modulation element.

[0122] According to the above configuration, when the light modulation element is a reflective light modulation element, the influence of noise light on the electric signal generated by the optical sensor can be effectively reduced.

[0123] In the optical computing device of mode 8 of the present invention, based on the structure of mode 6 or 7, the following structure is also adopted, that is, the reflecting surface of the reflector is not parallel to the incident and emitting surfaces of the first reflective light modulator and the second reflective light modulator.

[0124] According to the above structure, the incident angle of the signal light incident on the first reflective light modulator and the incident angle of the signal light incident on the second reflective light modulator can be made common, and the exit angle of the signal light emitted from the first reflective light modulator and the exit angle of the signal light emitted from the second reflective light modulator can be made common. Therefore, it is easy to design and manufacture the optical computing device.

[0125] In the optical computing device of mode 9 of the present invention, based on the structure of mode 6 or 7, the following structure is also adopted, namely: the optical computing unit is composed of a single optical modulator, and the single optical modulator is set with an area that functions as the first reflective optical modulator and the second reflective optical modulator.

[0126] According to the above configuration, it is not necessary to move the light modulators individually to optimize the positional relationship of the light modulators, and thus it is possible to easily manufacture a highly accurate optical computation unit.

[0127] In the optical computing device of mode 10 of the present invention, in addition to the structure of any one of modes 1 to 9, the following structure is also adopted, namely: the optical computing unit is a structure that transmits signal light and has a light diffraction layer formed inside that functions as the optical computing element.

[0128] According to the above configuration, since it is not necessary to move the light modulators individually to optimize the positional relationship of the light modulators, it is possible to easily manufacture a highly accurate optical computation unit.

[0129] [Additional Notes]

[0130] The present invention is not limited to the above-described embodiments, and various modifications can be made within the scope of the claims. Embodiments obtained by appropriately combining the technical means included in the above-described embodiments are also included in the technical scope of the present invention.

[0131] Description of Reference Numerals

[0132] 1, 2-optical computing device; 11, 21-optical computing unit; 11a1~11a3-optical modulation element (transmissive optical modulation element); 21a1~21a3-optical modulation element (reflective optical modulation element); 12, 22-image sensor.

Claims

1. An optical computing device, characterized in that: have: an optical computing unit, comprising an optical modulation element having a plurality of units for which phase modulation amounts are independently set or for which phase modulation amounts can be independently set; and an optical sensor for detecting signal light output from the optical computing unit and generating an electrical signal indicating a detection result, Regarding the light modulator, (1) the light modulator is a transmission type light modulator, which emits signal light generated by interference of transmitted light that has been phase-modulated by the multiple units, and noise light that has not been phase-modulated by the multiple units and has transmitted the light modulator, or (2) the light modulator is a reflection type light modulator, which emits signal light generated by interference of reflected light that has been phase-modulated by the multiple units, and noise light that has not been phase-modulated by the multiple units and has been reflected by the light modulator, In the optical modulation element, the phase modulation amount of each of the plurality of units is set so that the direction in which the signal light is emitted from the optical modulation element is different from the direction in which the noise light is emitted from the optical modulation element. Compared with a case where the direction in which the signal light is emitted from the light modulation element is consistent with the direction in which the noise light is emitted from the light modulation element, the influence of the noise light on the electrical signal is smaller.

2. The optical computing device according to claim 1, wherein: The optical sensor is an image sensor having a plurality of photoelectric conversion units, The image sensor is configured so that the noise light does not enter the photoelectric conversion unit into which the signal light enters with an intensity greater than a detection limit with an intensity greater than the detection limit.

3. The optical computing device according to claim 1, wherein: The optical sensor is configured such that (1) its incident surface intersects with the optical axis of the signal light emitted from the optical calculation unit, and (2) an intersection point where the incident surface intersects with the optical axis of the signal light emitted from the optical calculation unit does not intersect with the optical axis of the noise light emitted from the optical calculation unit.

4. The optical computing device according to any one of claims 1 to 3, characterized in that: The optical computing unit includes: a first transmission type optical modulation element and a second transmission type optical modulation element, as the optical modulation element, The second transmission-type light modulator is configured such that (1) its incident surface intersects with the optical axis of the signal light emitted from the first transmission-type light modulator, and (2) the intersection point of its incident surface with the optical axis of the signal light emitted from the first transmission-type light modulator does not intersect with the optical axis of the noise light emitted from the first transmission-type light modulator.

5. The optical computing device according to any one of claims 1 to 3, characterized in that: The optical computing unit includes: a first transmission type optical modulation element and a second transmission type optical modulation element, as the optical modulation element, The second transmission-type light modulator is configured such that (1) its incident surface intersects with the optical axis of the signal light emitted from the first transmission-type light modulator, and (2) at least a portion of the noise light emitted from the first transmission-type light modulator does not enter the incident surface of the second transmission-type light modulator.

6. The optical computing device according to any one of claims 1 to 3, characterized in that: It also has a reflector. The optical computing unit includes: a first reflective optical modulation element and a second reflective optical modulation element, as the optical modulation elements, The second reflective light modulator is configured such that (1) its incident and output surfaces intersect with the optical axis of the signal light emitted from the first reflective light modulator and reflected by the reflector, and (2) the intersection of its incident and output surfaces with the optical axis of the signal light emitted from the first reflective light modulator and reflected by the reflector does not intersect with the optical axis of the noise light emitted from the first reflective light modulator and reflected by the reflector or the optical axis of the noise light emitted from the first reflective light modulator.

7. The optical computing device according to any one of claims 1 to 3, characterized in that: It also has a reflector. The optical computing unit includes: a first reflective optical modulation element and a second reflective optical modulation element, as the optical modulation elements, The second reflective light modulator is configured such that (1) its incident and output surfaces intersect with the optical axis of the signal light emitted from the first reflective light modulator and reflected by the reflector, and (2) the noise light emitted from the first reflective light modulator and reflected by the reflector, or at least a portion of the noise light emitted from the first reflective light modulator, does not enter the incident and output surfaces of the second reflective light modulator.

8. The optical computing device according to claim 6 or 7, characterized in that: The reflection surface of the reflection mirror is not parallel to the incident and emitting surfaces of the first reflective light modulation element and the second reflective light modulation element.

9. The optical computing device according to any one of claims 6 to 8, characterized in that: The optical calculation unit is composed of a single light modulation element, and the single light modulation element has regions set to function as the first reflective light modulation element and the second reflective light modulation element.

10. The optical computing device according to any one of claims 1 to 5, characterized in that: The optical calculation unit is a structure that transmits signal light and has a light diffraction layer that functions as the light modulation element formed therein.

Images