Sensor, method for manufacturing sensor, and information processing method
Patent Information
- Application Number
- JP2022124366
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2022-08-03
- Publication Date
- 2025-08-06
- Estimated Expiration
- 2042-08-03
Smart Images
Figure 00000000_0000_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to a sensor, a method for manufacturing a sensor, and an information processing method. [Background technology]
[0002] Patent Document 1 discloses a sensor. This sensor includes a substrate, a light-emitting element, a light-receiving element, and a synchronous detection circuit, where the light-emitting element is provided at multiple different positions on the substrate and configured to emit light using first and second modulated signals that are orthogonal to each other, the light-receiving element is disposed inside the substrate and configured to receive external light through a pinhole and generate a photocurrent, and the synchronous detection circuit is configured to detect the phase of a specific component of the photocurrent, where the specific component is a component resulting from a composite reflected light that is irradiated from the light-emitting element and reflected from an object, and the phase is expressed as a function of the distance to the object. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] JP 2020-190552 A Summary of the Invention [Problem to be solved by the invention]
[0004] Meanwhile, unlike the prior art disclosed in Patent Document 1, there is a demand for a sensor that can grasp spatial physical quantities such as distance with a simpler structure and with high accuracy.
[0005] In view of the above circumstances, the present invention provides a sensor having a simpler structure and high accuracy. [Means for solving the problem]
[0006] According to one aspect of the present invention, a sensor for measuring an object is provided. The sensor includes a plurality of light-emitting elements and at least one light-receiving element. The light-emitting elements are provided at different positions on a substrate. The light-receiving element is provided on the substrate. One of the reflected lights caused by each light-emitting element is received as a main light, and the reflected light other than the main light is received as crosstalk light in a manner distinguishable from the main light. The reflected light is light emitted from each light-emitting element and reflected from the object. A spatial physical quantity related to a reference surface and the object in the sensor is measured based on the main light and the crosstalk light that are received in a manner distinguishable from the main light.
[0007] According to the present disclosure, a sensor having a simpler structure and high accuracy is realized. [Brief description of the drawings]
[0008] [Figure 1] FIG. 1 is a plan view mainly showing the outline of the configuration of a detection unit 2 of a sensor 1 according to the present embodiment. [Diagram 2] 2 is a block diagram showing mainly an outline of the configuration of an information processing unit 4 of a sensor 1 according to the present embodiment. FIG. [Diagram 3] 2 is a block diagram showing functions realized by a processor 43 and the like in the sensor 1. FIG. [Figure 4] This is an overview diagram showing the configuration of the neural network in the trained model 5. [Diagram 5] 1 is an activity diagram showing the flow of a measurement process of an object Ob using a sensor 1. FIG. [Figure 6] 6A, 6B, 6C, and 6D are schematic diagrams showing the timing at which each light-emitting element 31 emits light and the manner in which the light-receiving element 32 receives light, in which the light-emitting elements 31 emit light in the order of FIG. 6A, FIG. 6B, FIG. 6C, and FIG. 6D. [Figure 7] FIG. 11 is a schematic front view showing the configuration of a sensor 1 according to a modified example. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0009] [Embodiment] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described with reference to the drawings. Various characteristic features shown in the following embodiments can be combined with each other.
[0010] Incidentally, the program for realizing the software appearing in this embodiment may be provided as a non-transitory computer-readable recording medium, or may be provided so as to be downloadable from an external server, or may be provided so that the program is started on an external computer and its functions are realized on a client terminal (so-called cloud computing).
[0011] In addition, in this embodiment, the term "unit" may include, for example, a combination of hardware resources implemented by a circuit in the broad sense and software information processing that can be specifically realized by these hardware resources. In addition, in this embodiment, various information is handled, and this information is represented, for example, by physical values of signal values representing voltage and current, high and low signal values as a binary bit collection consisting of 0 or 1, or quantum superposition (so-called quantum bits), and communication and calculation can be performed on the circuit in the broad sense.
[0012] In addition, a circuit in the broad sense is a circuit realized by at least appropriately combining a circuit, circuitry, a processor, a memory, etc. In other words, it includes an application specific integrated circuit (ASIC), a programmable logic device (e.g., a simple programmable logic device (SPLD), a complex programmable logic device (CPLD), and a field programmable gate array (FPGA)), etc.
[0013] 1. Hardware Configuration In this section, a hardware configuration of the sensor 1 according to this embodiment will be described. The sensor 1 is a sensor for measuring an object Ob. More specifically, the sensor 1 is configured to be able to measure spatial physical quantities related to a reference plane in the sensor 1 and the object Ob. The spatial physical quantities here may include at least a distance d between the reference plane and the object Ob, and angles θ, φ (parameters indicating a three-dimensional posture) of the object Ob with respect to the reference plane. Particularly preferably, the distance d is a close distance, for example, the distance d is 50 mm or less, specifically, for example, 50, 49, 48, 47, 46, 45, 44, 43, 42, 41, 40, 39, 38, 37, 36, 35, 34, 33, 32, 31, 30, 29, 28, 27, 26, 25, 24, 23, 22, 21, 20, 19, 18, 17, 16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0 mm, and may be within a range between any two of the numerical values exemplified here. According to such an embodiment, the distance d and angles θ and φ to the object Ob are measured, and it can be applied to various situations such as picking and inspection of objects. Also, according to such an embodiment, it is possible to accurately grasp the object Ob located at a very close distance.
[0014] Fig. 1 is a plan view showing the general configuration of a sensor 1 according to this embodiment, mainly of a detection unit 2. Fig. 2 is a block diagram showing the general configuration of a sensor 1 according to this embodiment, mainly of an information processing unit 4. As shown in Figs. 1 and 2, the sensor 1 mainly includes a detection unit 2 and an information processing unit 4.
[0015] 1.1 Detection section 2 First, the detection unit 2 will be described. As shown in Fig. 1 and Fig. 2, the detection unit 2 includes a base material 21, a light receiving and emitting block 3 provided on the base material 21, and an AI input unit 22 connected to an information processing unit 4. Each component will be further described below.
[0016] (Base material 21) The base material 21 constitutes the housing of the sensor 1. In this embodiment, the base material 21 has, for example, a substantially circular shape and is configured in a flat plate shape, but this is merely an example and is not limited to this. Preferably, the base material 21 is formed of a material that is, for example, black, has no gloss, and has low reflectance. The base material 21 also includes a light receiving and emitting block 3 provided thereon.
[0017] (Light receiving and emitting block 3) 1, a plurality of light receiving and emitting blocks 3 are provided on the substrate 21. A light emitting element 31 and a light receiving element 32 are arranged in each light receiving and emitting block 3, and here, one light emitting element 31 and one light receiving element 32 are arranged in one light receiving and emitting block 3. In other words, the light receiving element 32 and the associated light emitting element 31 are adjacent to each other to form the light receiving and emitting block 3. A plurality of light receiving and emitting blocks 3 are provided on the substrate 21.
[0018] Here, the number of the light-emitting elements 31 is preferably 2 to 30, more preferably 2 to 10, and most preferably 2 to 4. Specifically, for example, the number of the light-emitting elements 31 is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30, and may be within a range between any two of the numerical values exemplified here.
[0019] The number of the light receiving and emitting blocks 3 may be 2 or more, preferably 3 or more, and in this embodiment, four light receiving and emitting blocks 3a, 3b, 3c, and 3d are provided. As shown in FIG. 1, more preferably, the light receiving and emitting blocks 3 are arranged on the substrate 21 in an annular shape and at equal intervals. In other words, the light receiving and emitting blocks 3 are arranged symmetrically with respect to the substrate 21. Specifically, the light receiving and emitting block 3a is provided with a light emitting element 31a and a light receiving element 32a, the light receiving and emitting block 3b is provided with a light emitting element 31b and a light receiving element 32b, the light receiving and emitting block 3c is provided with a light emitting element 31c and a light receiving element 32c, and the light receiving and emitting block 3d is provided with a light emitting element 31d and a light receiving element 32d. According to this embodiment, symmetry occurs in each of the photocurrents J (an example of an electrical physical quantity) based on the multiple light receiving elements 32, and the accuracy of the sensor 1 can be further improved. In other words, the number of light receiving elements 32 is the same as the number of light emitting elements 31, and they are in one-to-one correspondence with each light emitting element 31. According to this embodiment, a photocurrent J based on a plurality of light receiving elements 32 can be obtained, thereby further improving the accuracy of the sensor 1. The photocurrent J will be described later.
[0020] As shown in Fig. 1, the light-emitting elements 31 are provided at different positions on the substrate 21. The light-emitting elements 31 may be any element that emits diffuse light, such as a light-emitting diode (LED), and preferably an infrared LED that emits infrared light that is invisible to humans and harmless to the human body. Of course, the light-emitting elements 31 are not limited to this, and may be a red LED, a green LED, or a blue LED. By connecting the positive side of a power source (not shown) to the anode side of the light-emitting element 31, which is such a light-emitting diode, a current flows and diffuse light of a specific frequency is irradiated onto the object Ob.
[0021] As shown in FIG. 1, the light receiving element 32 is provided on the substrate 21. The light receiving element 32 is an element that detects received light, and generates a photocurrent J, which is an example of an electrical physical quantity, in response to the detection of the received light. In other words, the light receiving element 32 is configured to be able to output a photocurrent J according to the illuminance of the received light. Preferably, the characteristic between the illuminance of light and the photocurrent J is linear. Main light receiving elements 32 include, for example, phototubes, photomultiplier tubes, phototransistors that utilize the internal photoelectric effect of semiconductors, photodiodes, avalanche photodiodes, photoconductive cells, image sensors, and the like. Preferably, the light receiving element 32 is a photodiode having wide directivity, or a phototransistor that combines the photodiode with an amplifier.
[0022] Here, the light receiving element 32 receives the reflected light L emitted from the light emitting element 31 belonging to the same light receiving / emitting block 3 and reflected from the object Ob as the main light. Also, the light receiving element 32 receives the reflected light L emitted from the light emitting element 31 belonging to the adjacent light receiving / emitting block 3 and reflected from the object Ob as the crosstalk light. For example, the light receiving element 32a receives the reflected light L caused by the light emitting element 31a as the main light, and receives the reflected light L caused by the light emitting elements 31b and 31d as the crosstalk light. In other words, the light receiving element 32 receives one of the reflected lights L caused by each light emitting element 31 as the main light, and receives the reflected light L other than the main light as the crosstalk light that can be distinguished from the main light, and generates a photocurrent J (an example of an electrical physical quantity) according to the reflected light L. Here, the reflected light L is light emitted from each light emitting element 31 and reflected from the object Ob. In this way, based on the photocurrent J based on the distinguishably received main light and the crosstalk light, the spatial physical quantity related to the reference surface and the object Ob in the sensor 1 is measured. This will be described in more detail later.
[0023] (AI input section 22) The AI input unit 22 inputs the value of the photocurrent J or a calculated value based on the photocurrent J as an input parameter to the trained model 5 stored in the memory unit 42 of the information processing unit 4 described later. That is, the AI input unit 22 may be a conductor extending from each light receiving element 32 in the detection unit 2 to the information processing unit 4. The photocurrent J output from the light receiving element 32 is input to the trained model 5 stored in the memory unit 42 via the AI input unit 22 and a communication unit 41 described later.
[0024] 1.2 Information Processing Section 4 Next, the information processing unit 4 will be described. The information processing unit 4 is a processing circuit, such as a microcomputer, configured to control the operation of the sensor 1. The information processing unit 4 has a communication unit 41, a storage unit 42, and a processor 43, and these components are electrically connected via a communication bus 40 inside the information processing unit 4. Each component will be further described.
[0025] The communication unit 41 is configured to be able to transmit various electrical signals from the information processing unit 4 to external components. The communication unit 41 is also configured to be able to receive various electrical signals from the external components to the information processing unit 4. Specifically, the communication unit 41 receives the photocurrent J output from the AI input unit 22 in the detection unit 2. The communication unit 41 outputs a spatial physical quantity estimated based on the trained model 5. More preferably, the communication unit 41 has a network communication function, which allows various information to be communicated between the sensor 1 and an external device via a network such as the Internet.
[0026] The storage unit 42 stores various information defined by the above description. This can be implemented, for example, as a storage device such as a solid state drive (SSD) that stores various programs and the like related to the sensor 1 executed by the processor 43, or as a memory such as a random access memory (RAM) that stores temporarily required information (arguments, arrays, etc.) related to the calculation of the program. The storage unit 42 stores various programs, variables, etc. related to the sensor 1 executed by the processor 43. In particular, in this embodiment, the storage unit 42 stores the trained model 5.
[0027] The processor 43 is, for example, a central processing unit (CPU) not shown. The processor 43 realizes various functions related to the sensor 1 by reading out a predetermined program stored in the storage unit 42. That is, information processing by software stored in the storage unit 42 is specifically realized by the processor 43, which is an example of hardware, and can be executed as each functional unit included in the processor 43. These will be described in more detail in the next section. The processor 43 is not limited to being single, and may be implemented with multiple processors 43 for each function. Also, a combination of these may be used.
[0028] According to this embodiment, the spatial physical quantity of the object Ob can be measured by the sensor 1 alone without using a separate computer or the like.
[0029] 2. Functional configuration In this section, the functional configuration of the present embodiment will be described. As described above, information processing by the software stored in the storage unit 42 is specifically realized by the processor 43, which is an example of hardware, so that each functional unit included in the processor 43 can be executed.
[0030] 3 is a block diagram showing functions realized by the processor 43 and the like in the sensor 1. Specifically, the processor 43 includes an acquisition unit 431, a conversion unit 432, an input processing unit 433, and an output unit 434 as functional units.
[0031] As an acquisition step, the acquisition unit 431 is configured to acquire various information received from the outside via the communication unit 41 or stored in advance in the storage unit 42. For example, the acquisition unit 431 acquires the photocurrent J output from the light receiving element 32.
[0032] As a conversion step, the conversion unit 432 is configured to convert the information by performing a predetermined calculation on various pieces of information acquired by the acquisition unit 431. For example, the conversion unit 432 may calculate a predetermined calculation value based on the values of the multiple photocurrents J acquired by the acquisition unit 431 as converted information.
[0033] The input processing unit 433 is configured to input input parameters to the trained model 5 stored in the storage unit 42 as an input processing step.
[0034] The output unit 434 is configured to output various information as an output step. Specifically, the output unit 434 may output spatial physical quantities estimated based on the trained model 5, such as the distance d and the angles θ and φ.
[0035] 3. Construction of trained model 5 In this section, the trained model 5 stored in the storage unit 42 will be described. The trained model 5 is a model that has previously been machine-learned to learn the relationship between the spatial physical quantity related to the reference surface of the sensor 1 and the object Ob, and the value or calculated value of the photocurrent J. Preferably, the trained model 5 is trained to learn the relationship between the value of the photocurrent J or a predetermined calculated value and the spatial physical quantity between the object Ob and the reference surface of the sensor 1 for each material of the object Ob. Specifically, examples of the object Ob include a specularly reflective object Ob, a diffusely reflective white object Ob, a green / hard PVC board object Ob, a yellow / hard PVC board object Ob, a blue / hard PVC board object Ob, an aluminum board object Ob, a roughly finished aluminum board object Ob, a wire mesh object Ob, and a leather product object Ob. According to this embodiment, it is possible to perform robust measurement even on an object Ob that has been difficult to measure in the past, such as an object Ob having a mirror surface, a transparent object Ob, or an object Ob having an uneven surface.
[0036] 4 is a schematic diagram showing the configuration of a neural network in the trained model 5. As shown in FIG. 4, the trained model 5 includes an input layer 51, an intermediate layer 52 consisting of at least one layer, and an output layer 53.
[0037] To the input layer 51, input signals are input, for example, values of photocurrent J output from each light receiving element 32 or values calculated by performing a predetermined operation on the photocurrent J. In FIG. 4, at least eight input terminals (white circles in the figure) are shown as the input layer 51, but this is merely an example and is not limited to this.
[0038] The intermediate layer 52 is a layer that performs the main processing of the neural network, and the number of layers is not particularly limited. A plurality of neurons (white circles in the figure) in each layer are connected to a plurality of neurons in the next layer. For example, one neuron in the first layer of the intermediate layer 52 receives inputs weighted (loaded) from each terminal of the input layer 51, weights each neuron in the second layer of the intermediate layer 52, and transmits the calculation results. Next, one neuron in the second layer of the intermediate layer 52 receives inputs weighted from each neuron in the first layer of the intermediate layer 52, weights each neuron in the first layer of the intermediate layer 52, and transmits the calculation results to each neuron in the third layer of the intermediate layer 52. Such processing is executed sequentially according to the number of layers of the intermediate layer 52.
[0039] The output layer 53 is a layer that outputs the output result of the trained model 5, and in FIG. 4, two output terminals (white circles in the figure) are illustrated as the output layer 53, but this is merely an example and is not limited to this. Preferably, a distance d and angles θ and φ are output as spatial physical quantities relating to a reference plane and an object Ob in the sensor 1. That is, in relation to the trained model 5, the processor 43 in the information processing unit 4 is configured to input the value or calculated value of the photocurrent J to the trained model 5 as an input parameter and measure the spatial physical quantity based on the output result from the trained model 5.
[0040] 4. Measurement process flow including information processing method In this section, the flow of measurement processing of an object Ob using the sensor 1 will be described. Fig. 5 is an activity diagram showing the flow of measurement processing of an object Ob using the sensor 1. Fig. 6 is a schematic diagram showing the timing at which each light-emitting element 31 emits light and the manner in which the light-receiving element 32 receives light, with the light-emitting element 31 emitting light in the order of Fig. 6A, Fig. 6B, Fig. 6C, and Fig. 6D. Below, the flow of measurement processing will be described according to this activity diagram.
[0041] In this embodiment, a sensor 1 having four light receiving and emitting blocks 3 is assumed, and each of the light emitting elements 31 in the light receiving and emitting blocks 3 emits light separately at different timings. According to this embodiment, the reflected light L caused by each light emitting element 31 can be easily distinguished. Hereinafter, it will be expressed as "the kth (k=1, 2, 3, 4th) light emitting element 31 emits light." In other words, the light emitting elements 31a, 31b, 31c, and 31d emit light periodically at different timings. First, k=1 (activity A001), and the first light emitting element 31, for example, the light emitting element 31a as shown in FIG. 6A, emits light (activity A002).
[0042] Next, the light emitted from the light-emitting element 31a is irradiated to the object Ob, and the reflected light L is incident on the multiple light-receiving elements 32. Specifically, the light-receiving element 32a receives the reflected light L caused by the light-emitting element 31a as the main light (activity A003). The light-receiving element 32a is a light-receiving element 32 that is paired with the light-emitting element 31a, and is located in the same light-receiving and light-receiving block 3a as the light-emitting element 31a here. The light-receiving element 32b receives the reflected light L caused by the light-emitting element 31a as crosstalk light (activity A004). The light-receiving element 32b is a light-receiving element 32 that is located in the light-receiving and light-receiving block 3b that is located to the left of the light-emitting and light-receiving block 3a in which the light-emitting element 31a is located. The light-receiving element 32d receives the reflected light L caused by the light-emitting element 31a as crosstalk light (activity A005). The light receiving element 32d is a light receiving element 32 arranged in the light receiving / emitting block 3d located to the right of the light receiving / emitting block 3a in which the light emitting element 31a is arranged. Then, a photocurrent J corresponding to the received reflected light L is output from each of the light receiving elements 32a, 32b, and 32d. Then, the acquisition unit 431 in the information processing unit 4 acquires each photocurrent J. In other words, the light receiving element 32 receives the reflected light L caused by the associated light emitting element 31 as a main light, and receives the reflected light L other than the main light as a crosstalk light that can be distinguished from the main light.
[0043] Next, since k<4, the value of k is incremented (activity A006), and the second light-emitting element 31, for example the light-emitting element 31b as shown in FIG. 6B, emits light (activity A002).
[0044] Next, the light emitted from the light-emitting element 31b is irradiated onto the object Ob, and the reflected light L is incident on the multiple light-receiving elements 32. Specifically, the light-receiving element 32b receives the reflected light L caused by the light-emitting element 31b as the main light (activity A003). The light-receiving element 32c receives the reflected light L caused by the light-emitting element 31b as crosstalk light (activity A004). The light-receiving element 32a receives the reflected light L caused by the light-emitting element 31b as crosstalk light (activity A005). Then, a photocurrent J corresponding to the received reflected light L is output from each of the light-receiving elements 32b, 32c, and 32a. After that, the acquisition unit 431 in the information processing unit 4 acquires each photocurrent J.
[0045] Next, since k<4, the value of k is incremented (activity A006), and the third light-emitting element 31, for example the light-emitting element 31c as shown in FIG. 6C, emits light (activity A002).
[0046] Next, the light emitted from the light-emitting element 31c is irradiated onto the object Ob, and the reflected light L is incident on the multiple light-receiving elements 32. Specifically, the light-receiving element 32c receives the reflected light L caused by the light-emitting element 31c as the main light (activity A003). The light-receiving element 32d receives the reflected light L caused by the light-emitting element 31c as crosstalk light (activity A004). The light-receiving element 32b receives the reflected light L caused by the light-emitting element 31c as crosstalk light (activity A005). Then, a photocurrent J corresponding to the received reflected light L is output from each of the light-receiving elements 32c, 32d, and 32b. After that, the acquisition unit 431 in the information processing unit 4 acquires each photocurrent J.
[0047] Next, since k<4, the value of k is incremented (activity A006), and the fourth light-emitting element 31, for example the light-emitting element 31d as shown in FIG. 6D, emits light (activity A002).
[0048] Next, the light emitted from the light-emitting element 31d is irradiated onto the object Ob, and the reflected light L is incident on the multiple light-receiving elements 32. Specifically, the light-receiving element 32d receives the reflected light L caused by the light-emitting element 31d as the main light (activity A003). The light-receiving element 32a receives the reflected light L caused by the light-emitting element 31d as crosstalk light (activity A004). The light-receiving element 32c receives the reflected light L caused by the light-emitting element 31d as crosstalk light (activity A005). Then, the light-receiving elements 32d, 32a, and 32c each output a photocurrent J corresponding to the received reflected light L. Then, the acquisition unit 431 in the information processing unit 4 acquires each photocurrent J. By the above-mentioned four times of emission, the photocurrent J of 12 channels is acquired.
[0049] Since k=4 this time, the processes of activities A007 to A009 are executed. Specifically, first, a pre-processing is executed in which the conversion unit 432 converts each photocurrent J acquired by the acquisition unit 431 in the information processing unit 4 into a predetermined calculation value (activity A007). The predetermined calculation value is not particularly limited as long as it is a value calculated based on each photocurrent J. The number of channels of such calculation values may be the same as the number of channels of the photocurrent J (12 in one example), or may be less than this.
[0050] Next, the input processing unit 433 in the information processing unit 4 inputs each calculation value obtained in activity A007 to the trained model 5 as an input parameter (activity A008).
[0051] Finally, the output unit 434 in the information processing unit 4 outputs the spatial physical quantities, here the distance d and the angles θ and φ, estimated from the trained model 5 (activity A009). The above process is continuously executed as long as the sensor 1 is operating, so the distance d and the angles θ and φ are measured sequentially according to the operating frequency of the sensor 1. The time from the emission of the first light-emitting element 31 to the estimated measurement of the distance d and the angles θ and φ is preferably 100 milliseconds or less, preferably 10 milliseconds or less, more preferably 1 millisecond or less, and even more preferably 0.7 milliseconds or less. Specifically, the time until measurement may be, for example, 100, 95, 90, 85, 80, 75, 70, 65, 60, 55, 50, 45, 40, 35, 30, 25, 20, 15, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.1 milliseconds, and may be within a range between any two of the numerical values exemplified here. That is, the time from the emission of the kth light-emitting element 31 to the emission of the k+1th light-emitting element 31 may be 1 / 4 or less of the above measurement time. More generally, the time from the emission of the kth light-emitting element 31 to the emission of the k+1th light-emitting element 31 may be less than the time obtained by dividing "the time from the emission of the first light-emitting element 31 to the estimation of the distance d and the angles θ and φ" by "the number of light-emitting elements 31".
[0052] To summarize the above, the sensor 1 according to this embodiment is a sensor for measuring an object Ob. The sensor 1 includes a plurality of light-emitting elements 31, a light-receiving element 32, and an AI input unit 22. The light-emitting elements 31 are provided at different positions on the substrate 21. The light-receiving element 32 is provided on the substrate 21. The light-receiving element 32 is configured to receive one of the reflected lights L caused by each light-emitting element 31 as a main light, and to receive the reflected lights L other than the main light as crosstalk light distinguishable from the main light. The reflected lights L are lights emitted from each light-emitting element 31 and reflected from the object Ob. The sensor 1 measures spatial physical quantities related to the reference surface and the object Ob in the sensor 1 based on the main lights and the crosstalk light that are received in a distinguishable manner.
[0053] From another perspective, this information processing method includes the following steps. In the acquisition step, a photocurrent J (an example of an electrical physical quantity) output from the light receiving element 32 is acquired. In the input processing step, the value of the photocurrent J or a calculated value based on the photocurrent J is input as an input parameter to the trained model 5. The trained model 5 is a model that has previously been machine-learned to determine the relationship between the spatial physical quantity relating to the reference surface and the object Ob in the sensor 1 and the value or calculated value of the photocurrent J. In the output step, the spatial physical quantity is estimated and output based on the trained model 5.
[0054] According to this embodiment, a sensor having a simpler structure and high accuracy can be realized. In particular, by using the crosstalk light, which has been considered as a disturbance in the past, as an input parameter for measurement judgment, instead of only using the reflected light L between the corresponding elements such as the light emitting element 31a and the light receiving element 32a, more multi-dimensional information can be obtained and the accuracy of measurement can be improved.
[0055] [others] The sensor 1 according to this embodiment may be further improved.
[0056] The light emitting element 31 and the light receiving element 32 do not have to be arranged in a one-to-one relationship. FIG. 7 is a schematic front view showing the configuration of the sensor 1 according to the modified example. The sensor 1 according to the modified example includes, as a minimum configuration, two light emitting elements 31x and 31y and one light receiving element 32x. Light emitted from the light emitting element 31x is reflected by the object Ob, and the reflected light L1 (an example of the main light) is received by the light receiving element 32x. In addition, light emitted from the light emitting element 31y is reflected by the object Ob, and the reflected light L2 (an example of the crosstalk light) is received by the light receiving element 32x. Even with the sensor 1 of this aspect, the distance d and the angle θ (angle in one direction) can be estimated by using the trained model 5 that has been machine-learned in advance.
[0057] In the above embodiment, the light emitting and receiving blocks 3 include the light emitting elements 31 and the light receiving elements 32 in a one-to-one relationship, and the light emitting and receiving blocks 3 are arranged annularly and at equal intervals on the base material 21, but the light emitting elements 31 and the light receiving elements 32 may be arranged on the base material 21 without any regularity. Even in such a case, the distance d and the angles θ and φ can be estimated by preparing a trained model 5 that has been machine-learned in advance.
[0058] Although the sensor 1 has been described as including the detection unit 2 and the information processing unit 4, the sensor 1 may have only the detection unit 2, and the measurement results may be output using a computer (not shown) connected thereto. In such a case, the trained model 5 may be stored in a storage unit (not shown) built into the computer (not shown), and an input process to the trained model 5 may be executed by a processor (not shown). The manner of connection is not particularly limited, and the sensors may be directly connected by a conductor, or the sensor 1 may have a wireless communication function and be connected to the computer (not shown) via wireless communication. In such a case, one computer (not shown) may be used for multiple sensors 1.
[0059] The sensor 1 described above may be manufactured by a predetermined manufacturing method. The manufacturing method may include a step of arranging a plurality of light-emitting elements 31 and at least one light-receiving element 32 on a substrate 21. For example, the light-emitting element 31 may be arranged on the substrate 21, followed by the light-receiving element 32, and then the output of the light-receiving element 32 may be connected to an information processing unit 4 such as a microcomputer. According to such a method, a sensor having a simpler structure and high accuracy may be manufactured.
[0060] In the above-described embodiment, each light-emitting element 31 emits light at a different timing, thereby distinguishing between the obtained photocurrents J. However, instead of the timing, each light-emitting element 31 may be made to blink at a different frequency, or the wavelength of light emitted by the light-emitting elements 31 may be made different, thereby distinguishing between the obtained photocurrents J.
[0061] The crosstalk light may be received not only by the light receiving elements 32 adjacent to the left and right, but also by the opposing light receiving elements 32. In other words, the position of the light receiving elements 32 that employ the crosstalk light can be freely determined according to the number of light receiving and emitting blocks 3.
[0062] The trained model 5 may be a one-dimensional convolutional network or a two-dimensional convolutional network instead of a neural network.
[0063] The electrical physical quantity generated by the light receiving element 32 is not limited to the photocurrent J, but may be, for example, a voltage.
[0064] Furthermore, it may be provided in the following aspects:
[0065] (1) A sensor for measuring an object, comprising a plurality of light-emitting elements and at least one light-receiving element, the light-emitting elements being provided at different positions on a substrate, the light-receiving element being provided on the substrate, and receiving one of the reflected lights caused by each light-emitting element as a main light, and receiving reflected light other than the main light as crosstalk light distinguishable from the main light, the reflected light being light emitted from each of the light-emitting elements and reflected from the object, and measuring a spatial physical quantity relating to a reference surface in the sensor and the object based on the distinguishably received main light and the crosstalk light.
[0066] According to this embodiment, a highly accurate sensor having a simpler structure can be realized.
[0067] (2) In the sensor described in (1) above, the spatial physical quantity includes at least the distance between the reference plane and the object, and the angle of the object with respect to the reference plane.
[0068] According to this embodiment, the distance and angle to an object can be measured, and the invention can be applied to various situations such as picking and inspecting objects.
[0069] (3) In the sensor described in (2) above, the distance is 50 mm or less.
[0070] According to this embodiment, it is possible to accurately grasp an object located very close by.
[0071] (4) In the sensor described in any one of (1) to (3) above, the light receiving elements are the same number as the light emitting elements, are in one-to-one correspondence with each of the light emitting elements, and are configured to receive the reflected light caused by the corresponding light emitting element as the main light, and to receive the reflected light other than the main light as crosstalk light that can be distinguished from the main light.
[0072] According to this embodiment, light is obtained from a plurality of light receiving elements, so that the accuracy of the sensor can be further improved.
[0073] (5) In the sensor described in (4) above, the light receiving element and the associated light emitting element are adjacent to each other to form a light receiving and emitting block, and a plurality of the light receiving and emitting blocks are provided on the substrate.
[0074] According to this embodiment, light is obtained from a plurality of light receiving elements, so that the accuracy of the sensor can be further improved.
[0075] (6) In the sensor described in (5) above, the light receiving and emitting blocks are arranged on the base material in a circular shape and at equal intervals.
[0076] According to this embodiment, symmetry occurs in the physical quantities corresponding to the light obtained from the plurality of light receiving elements, and the accuracy of the sensor can be further improved.
[0077] (7) The sensor according to any one of (1) to (6) above, wherein the number of the light-emitting elements is 2 to 4.
[0078] According to this embodiment, a highly accurate sensor having a simpler structure can be realized.
[0079] (8) In the sensor according to any one of (1) to (7) above, the light-emitting elements each emit light at a different timing from each other.
[0080] According to this embodiment, the reflected light caused by each light emitting element can be easily distinguished.
[0081] (9) In the sensor described in any one of (1) to (8) above, the light receiving element is configured to generate an electrical physical quantity corresponding to the reflected light, and a spatial physical quantity relating to a reference surface and the object in the sensor is measured based on each of the electrical physical quantities based on the main light and the crosstalk light that are received in a distinguishable manner.
[0082] According to this embodiment, electrical handling essential for information processing becomes possible, and a highly accurate sensor is realized.
[0083] (10) The sensor described in (9) above, further comprising an AI input unit, which inputs the value of the electrical physical quantity or a calculated value based on the electrical physical quantity as an input parameter to a trained model, wherein the trained model is a model that has been machine-learned in advance to determine the relationship between a spatial physical quantity relating to a reference plane and the target object in the sensor and the value of the electrical physical quantity or the calculated value.
[0084] According to this embodiment, a highly accurate sensor based on machine learning is realized.
[0085] (11) The sensor described in (10) above, further comprising a memory unit and a processor, wherein the memory unit stores the trained model, and the processor is configured to input the value of the electrical physical quantity or the calculated value to the trained model as the input parameter, and to measure the spatial physical quantity based on the output result from the trained model.
[0086] According to this aspect, it is possible to measure the spatial physical quantity of an object using only the sensor without using a separate computer or the like.
[0087] (12) In the sensor according to any one of (1) to (11) above, the light receiving element is a light receiving element having wide directivity.
[0088] According to this aspect, both the main light and the crosstalk light can be received reliably, and the accuracy of the sensor can be improved.
[0089] (13) A method for manufacturing a sensor, comprising the step of arranging a plurality of light-emitting elements and at least one light-receiving element on a substrate, the sensor being the sensor described in any one of (1) to (12) above.
[0090] According to this method, a sensor having a simpler structure and high accuracy can be manufactured.
[0091] (14) An information processing method comprising the following steps: in an acquisition step, a photocurrent output from a light receiving element is acquired; in an input processing step, the value of the electrical physical quantity or a calculated value based on the electrical physical quantity is input as an input parameter to a trained model, wherein the trained model is a model that has been machine-learned in advance to determine the relationship between a spatial physical quantity relating to a reference surface and an object in the sensor and the value of the electrical physical quantity or the calculated value; and in an output step, the spatial physical quantity is estimated and output based on the trained model.
[0092] According to this embodiment, a highly accurate sensor having a simpler structure can be realized. Of course, this is not the case.
[0093] Finally, although various embodiments of the present invention have been described, these are presented as examples and are not intended to limit the scope of the invention. The novel embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. The embodiments and their modifications are included within the scope and spirit of the invention, and are included in the scope of the invention and its equivalents described in the claims. [Explanation of symbols]
[0094] 1: Sensor 2: Detection section 21: Base material 22:AI input section 3: Light receiving and emitting block 3a: Light receiving and emitting block 3b: Light receiving and emitting block 3c: Light receiving and emitting block 3d: Light receiving and emitting block 31: Light emitting element 31a: Light emitting element 31b: Light emitting element 31c: Light emitting element 31d: Light emitting element 31x: Light emitting element 31y: Light-emitting element 32: Photodetector 32a: Light receiving element 32b: Light receiving element 32c: Photodetector 32d: Light receiving element 32x: Photodetector 4: Information processing section 40: Communication bus 41: Communications Department 42: Storage section 43: Processor 431: Acquisition Department 432: Conversion section 433: Input processing section 434: Output section 5: Trained model 51: Input layer 52: Middle class 53: Output layer J : Photocurrent L : reflected light L1: reflected light L2: reflected light Ob: object d : distance θ : angle φ : Angle
Claims
1. A sensor for measuring an object, A plurality of light-emitting elements and at least one light-receiving element are provided, The light-emitting elements are provided at different positions on a substrate, The light receiving element is provided on the substrate, One of the reflected lights caused by each light-emitting element is received as a main light, and reflected lights other than the main light are received as crosstalk light in a manner distinguishable from the main light, wherein the reflected lights are lights emitted from each of the light-emitting elements and reflected from the object, A spatial physical quantity relating to a reference surface in the sensor and the object is measured based on the distinguishably received main light and the crosstalk light.
2. 2. The sensor of claim 1, The spatial physical quantity includes at least the distance between the reference plane and the object, and the angle of the object relative to the reference plane.
3. 3. The sensor according to claim 2, The distance is 50 mm or less.
4. 2. The sensor of claim 1, The light receiving element is The number of the light emitting elements is the same as the number of the light emitting elements, and the light emitting elements are respectively associated with each other one-to-one, The reflected light caused by the associated light-emitting element is received as the main light, and the reflected light other than the main light is received as crosstalk light that can be distinguished from the main light.
5. 5. The sensor according to claim 4, the light receiving element and the associated light emitting element are adjacent to each other to form a light receiving and emitting block, A plurality of the light receiving and emitting blocks are provided on the base material.
6. 6. The sensor according to claim 5, The light emitting and receiving blocks are arranged on the base material in a circular pattern at equal intervals.
7. 2. The sensor of claim 1, The number of the light-emitting elements is 2 to 4.
8. 2. The sensor of claim 1, The light emitting elements emit light at different times.
9. 2. The sensor of claim 1, the light receiving element is configured to generate an electrical physical quantity in accordance with the reflected light, A spatial physical quantity relating to a reference surface in the sensor and the object is measured based on the electrical physical quantities based on the distinguishably received main light and the crosstalk light.
10. 10. The sensor of claim 9, Further comprising an AI input unit, The AI input unit inputs the value of the electrical physical quantity or a calculated value based on the electrical physical quantity as an input parameter to a trained model, wherein the trained model is a model that has previously been machine-learned to determine the relationship between a spatial physical quantity relating to a reference plane in the sensor and the object, and the value of the electrical physical quantity or the calculated value.
11. 11. The sensor of claim 10, Further comprising a storage unit and a processor, The storage unit stores the trained model, The processor is configured to input the value of the electrical physical quantity or the calculated value as the input parameter to the trained model, and measure the spatial physical quantity based on the output result from the trained model.
12. 2. The sensor of claim 1, The light receiving element is a light receiving element having wide directivity.
13. A method for manufacturing a sensor, comprising: A method comprising the step of arranging a plurality of light-emitting elements and at least one light-receiving element on a substrate, wherein the sensor is a sensor according to any one of claims 1 to 12.
14. An information processing method, comprising: It includes the following steps: In the acquisition step, an electrical physical quantity output from the light receiving element is acquired; In the input processing step, the value of the electrical physical quantity or a calculated value based on the electrical physical quantity is input as an input parameter to a trained model, wherein the trained model is a model that has previously undergone machine learning to learn a relationship between a spatial physical quantity relating to a reference plane in the sensor and an object, and the value of the electrical physical quantity or the calculated value; In the output step, the spatial physical quantity is estimated and output based on the trained model.