Sensor, method for manufacturing sensor, and information processing method

Abstract

The invention provides a sensor having a simpler structure and high precision. According to one aspect of the present invention, a sensor for measuring an object may be provided. The sensor is provided with a plurality of light-emitting elements and at least one light-receiving element. The light-emitting elements are respectively arranged at different positions on the base material. The light receiving element is arranged on the base material. The light receiving element receives, as main light, one of the reflected light caused by each light emitting element, and receives, as crosstalk light, reflected light other than the main light distinguishably from the main light. The reflected light is emitted from each of the light-emitting elements and is reflected from each of the objects. Spatial physical quantities related to a reference plane and an object in the sensor are measured based on the differentially received main light and crosstalk light.

Description

Technical Field

[0001] The present invention relates to a sensor, a method for manufacturing the sensor, and an information processing method. Background Art

[0002] Patent Document 1 discloses a sensor. The sensor includes a substrate, a light-emitting element, a light-receiving element, and a synchronous detection circuit. A plurality of light-emitting elements are provided at different positions on the substrate and are configured to emit light using first and second modulation signals that are orthogonal to each other. The light-receiving element is arranged inside the substrate and is configured to receive external light through a pinhole and generate a photocurrent. The synchronous detection circuit is configured to detect the phase of a specific component in the photocurrent. Here, the specific component refers to a component caused by the synthetic reflected light irradiated from the light-emitting element and reflected from the object, and the phase is expressed as a function of the distance to the object. Prior art literature Patent Literature

[0003] Patent Document 1: Japanese Patent Application Laid-Open No. 2020-190552 Summary of the Invention Technical problem to be solved by the invention

[0004] On the other hand, a sensor that is different from the prior art disclosed in Patent Document 1 and has a simpler structure and can grasp spatial physical quantities such as distance with high precision is sought.

[0005] In view of the above circumstances, the present invention provides a sensor having a simpler structure and high precision. Technical means to solve the problem

[0006] According to one embodiment of the present invention, a sensor for measuring an object is provided. The sensor includes multiple light-emitting elements and at least one light-receiving element. The light-emitting elements are disposed at different positions on a substrate. The light-receiving element is disposed on the substrate. One of the reflected lights from each light-emitting element is received as main light, and reflected light other than the main light is received distinguishably from the main light as crosstalk light. The reflected light is light emitted from each light-emitting element and reflected from the object. Based on the distinguishably received main light and crosstalk light, a spatial physical quantity related to a reference surface and the object in the sensor is measured.

[0007] According to the present invention, a sensor having a simpler structure and high precision can be realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 It is a plan view showing the schematic configuration of the sensor 1 according to the present embodiment, mainly the detection unit 2 . Figure 2This is a block diagram showing the schematic configuration of the sensor 1 according to the present embodiment, mainly the information processing unit 4 . Figure 3 This is a block diagram showing functions implemented by the processor 43 and the like in the sensor 1 . Figure 4 Schematic diagram showing the structure of the neural network in the learned model 5. Figure 5 This is an activity diagram showing the flow of a measurement process of the object Ob using the sensor 1 . Figure 6 Schematic diagram showing the timing of each light emitting element 31 emitting light and the way in which the light receiving element 32 receives light. Figure 6 A. Figure 6 B. Figure 6 C and Figure 6 D's order of light. Figure 7 It is a schematic front view showing the structure of the sensor 1 according to a modified example. DETAILED DESCRIPTION

[0009] [Implementation Method] The following describes embodiments of the present invention using the accompanying drawings. Various features described in the following embodiments can be combined with each other.

[0010] However, the program for implementing the software appearing in this embodiment can be provided as a non-transitory computer-readable medium (Non-Transitory Computer-Readable Medium), can be provided by downloading from an external server, or can be provided in a manner such that the program is started in an external computer and the function is implemented on a client terminal (so-called cloud computing).

[0011] Furthermore, in this embodiment, a "unit" may also include, for example, a structure that combines hardware resources implemented by a broad circuit and information processing implemented specifically by software using these hardware resources. Furthermore, this embodiment processes a variety of information, such as physical values ​​representing voltage / current signal values, high / low signal values ​​represented by binary bit sets consisting of 0s and 1s, or quantum superpositions (so-called qubits), and performs communication / computation on a broad circuit.

[0012] In a broad sense, a circuit refers to a circuit implemented by appropriately combining at least circuits, circuitry, processors, and memory. This includes application-specific integrated circuits (ASICs), programmable logic devices (e.g., simple programmable logic devices (SPLDs), complex programmable logic devices (CPLDs), and field programmable gate arrays (FPGAs).

[0013] 1. Hardware structure This section describes the hardware structure of sensor 1 according to this embodiment. Sensor 1 is used to measure an object Ob. More specifically, sensor 1 is configured to measure spatial physical quantities related to a reference plane within sensor 1 and object Ob. These spatial physical quantities preferably include at least the distance d between the reference plane and object Ob, and the angles θ and φ (parameters indicating the three-dimensional posture) of object Ob relative to the reference plane. It is particularly preferred that the distance d be the closest distance. For example, the distance d is 50 mm or less. Specifically, for example, the distance d is 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, 21, 20, 19, 18, 17, 16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, or 0 mm. Alternatively, the distance d and the angles θ and φ between the object Ob and the object can be measured and applied to various situations such as object picking and inspection. Furthermore, the object Ob located at a very close distance can be accurately detected according to the embodiment.

[0014] Figure 1 It is a plan view showing the schematic configuration of the sensor 1 according to the present embodiment, mainly the detection unit 2 . Figure 2 1 is a block diagram showing the structure of the sensor 1 according to the present embodiment, mainly the information processing unit 4. Figure 1 and Figure 2 As shown, the sensor 1 mainly includes a detection unit 2 and an information processing unit 4 .

[0015] 1.1 Detection Unit 2 First, the detection unit 2 will be described. Figure 1 and Figure 2 As shown, the detection unit 2 includes a substrate 21, a light-receiving and light-emitting block 3 provided on the substrate 21, and an AI input unit 22 connected to the information processing unit 4. Each component will be further described below.

[0016] (Substrate 21) The substrate 21 forms the frame of the sensor 1. In this embodiment, the substrate 21 has a generally circular, flat plate-like shape, for example. However, this is merely an example and is not limiting. Preferably, the substrate 21 is formed of a black, matte material with low reflectivity. Furthermore, the substrate 21 includes the light-receiving and light-emitting blocks 3 disposed thereon.

[0017] (Light receiving block 3) like Figure 1 As shown, a plurality of light-receiving and light-emitting blocks 3 are provided on the substrate 21. Each light-receiving and light-emitting block 3 is provided with a light-emitting element 31 and a light-receiving element 32. Here, one light-receiving and light-receiving block 3 is provided with one light-emitting element 31 and one light-receiving element 32. In other words, a light-receiving element 32 and a corresponding light-emitting element 31 are adjacent to each other to form a light-receiving and light-emitting block 3. A plurality of light-receiving and light-emitting blocks 3 are provided on the substrate 21.

[0018] Here, the number of 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 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 values ​​exemplified here.

[0019] In addition, the number of the light receiving blocks 3 may be two or more, preferably three or more. In this embodiment, four light receiving blocks 3a, 3b, 3c, and 3d are provided. Figure 1As shown, it is further preferred that each of the light-receiving and light-emitting blocks 3 is arranged in a ring shape and at equal intervals on the substrate 21. In other words, the light-receiving and light-emitting blocks 3 are symmetrically arranged on the substrate 21. Specifically, the light-receiving and light-emitting blocks 3a are arranged with light-emitting elements 31a and light-receiving elements 32a; the light-receiving and light-emitting blocks 3b are arranged with light-emitting elements 31b and light-receiving elements 32b; the light-receiving and light-emitting blocks 3c are arranged with light-emitting elements 31c and light-receiving elements 32c; and the light-receiving and light-emitting blocks 3d are arranged with light-emitting elements 31d and light-receiving elements 32d. In this embodiment, symmetry is achieved in the photocurrents J (an example of an electrical quantity) generated by the multiple light-receiving elements 32, further improving the accuracy of the sensor 1. In other words, the number of light-receiving elements 32 is the same as that of the light-emitting elements 31, and each light-receiving element 31 has a one-to-one correspondence. In this embodiment, the photocurrents J generated by the multiple light-receiving elements 32 are obtained, thereby further improving the accuracy of the sensor 1. The photocurrents J will be described later.

[0020] like Figure 1 As shown, the light-emitting elements 31 are respectively arranged at different positions on the substrate 21. The light-emitting element 31 can be any element that emits diffuse light, such as a light-emitting diode (LED), preferably an infrared LED that emits infrared rays that are invisible to humans and harmless to the human body. Of course, this is not limited to this, and the light-emitting element 31 can be a red LED, a green LED, or a blue LED. By connecting the positive side of a power supply (not shown) to the anode side of such a light-emitting element 31, which is a light-emitting diode, current flows, and diffuse light of a specific frequency is irradiated on the object Ob.

[0021] like Figure 1 As shown, the light receiving element 32 is provided on the substrate 21. The light receiving element 32 is an element that detects the received light and generates a photocurrent J as an example of an electrical physical quantity based on the light received. In other words, the light receiving element 32 is configured to output the photocurrent J according to the illuminance of the received light. Preferably, the characteristics of the illuminance of the light and the photocurrent J are linear. As the main light receiving elements 32, for example, there can be cited: a phototube, a photomultiplier tube, a phototransistor utilizing the internal photoelectric effect of a semiconductor, a photodiode, an avalanche photodiode, a photoconductive cell, and an image sensor. Preferably, the light receiving element 32 is a photodiode having a wide directivity, or a phototransistor in which the photodiode and an amplifier are combined.

[0022] Here, the light receiving element 32 receives reflected light L emitted from the light emitting elements 31 belonging to the same light receiving block 3 and reflected from the object Ob as main light. Furthermore, the light receiving element 32 receives reflected light L emitted from the light emitting elements 31 belonging to adjacent light receiving blocks 3 and reflected from the object Ob as crosstalk light. For example, the light receiving element 32a receives reflected light L from the light emitting element 31a as main light, and reflected light L from the light emitting elements 31b and 31d as crosstalk light. In other words, the light receiving element 32 is configured to receive one of the reflected light L from each light emitting element 31 as main light, and to receive reflected light L other than the main light, distinguishable from the main light, as crosstalk light, thereby generating a photocurrent J (an example of an electrical quantity) corresponding 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, spatial physical quantities related to the reference surface and the object Ob on the sensor 1 are measured based on the photocurrents J based on the distinguishably received main light and crosstalk light. This will be described in further detail later.

[0023] (AI input unit 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 into the learned model 5 stored in the storage unit 42 of the information processing unit 4, which will be described later. Specifically, the AI ​​input unit 22 may be a conductive line 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 learned model 5 stored in the storage unit 42 via the AI ​​input unit 22 and the communication unit 41, which will be described later.

[0024] 1.2 Information Processing Unit 4 Next, the information processing unit 4 will be described. The information processing unit 4 is a processing circuit configured to control the operation of the sensor 1, and is, for example, a microcontroller. The information processing unit 4 includes a communication unit 41, a storage unit 42, and a processor 43. These components are electrically connected within the information processing unit 4 via a communication bus 40. Each component will be further described.

[0025] The communication unit 41 is configured to transmit various electrical signals from the information processing unit 4 to external components. Furthermore, the communication unit 41 is configured to receive various electrical signals from 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 an estimated spatial physical quantity based on the learned model 5. More preferably, the communication unit 41 has a network communication function, thereby enabling communication between the sensor 1 and external devices via a network such as the Internet.

[0026] The storage unit 42 stores the various information defined above. For example, it can be implemented as a storage device such as a solid-state drive (SSD) that stores various programs related to the sensor 1 and executed by the processor 43, or as a memory such as a random access memory (RAM) that stores temporarily required information (parameters, sequences) related to program operations. The storage unit 42 stores various programs and variables related to the sensor 1 executed by the processor 43. In particular, in this embodiment, the storage unit 42 stores the learned model 5.

[0027] The processor 43 is, for example, a central processing unit (CPU) (not shown). The processor 43 implements various functions related to the sensor 1 by reading predetermined programs stored in the storage unit 42. Specifically, the information processing performed by the software stored in the storage unit 42 can be specifically implemented by the processor 43, which is an example of hardware, and the information processing can be executed as each functional unit included in the processor 43. This will be described in detail in the following section. Furthermore, the processor 43 is not limited to a single one and can also be implemented with multiple processors 43 for each function. Furthermore, a combination of these is also possible.

[0028] According to such a configuration, 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 structure As described above, information processing by software stored in the storage unit 42 can be specifically implemented by the processor 43 as an example of hardware, and each functional unit included in the processor 43 is executed.

[0030] Figure 3 4 is a block diagram showing functions implemented 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] The acquisition unit 431 is configured to acquire, as an acquisition step, various types of 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] The conversion unit 432 is configured to convert the information by performing a predetermined operation on the various information acquired by the acquisition unit 431. For example, the conversion unit 432 may perform an operation using a predetermined operation value based on the multiple photocurrent J values ​​acquired by the acquisition unit 431 as the converted information.

[0033] The input processing unit 433 is configured to input input parameters to the learned 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 learned model 5, such as the distance d and the angles θ and φ.

[0035] 3. Complete the study of the structure of model 5 This section describes the learned model 5 stored in the storage unit 42. The learned model 5 is a model obtained by pre-machine learning of the relationship between the spatial physical quantity associated with the reference surface on the sensor 1 and the object Ob, and the value or calculated value of the photocurrent J. Preferably, the learned model 5 learns the relationship between the value or a predetermined calculated value of the photocurrent J and the spatial physical quantity between the object Ob and the reference surface of the sensor 1, depending on the material of the object Ob. Specifically, the learned model 5 includes objects Ob made of a specularly reflective material, diffusely reflective white materials, objects Ob made of green rigid polyvinyl chloride board, yellow rigid polyvinyl chloride board, blue rigid polyvinyl chloride board, aluminum board, rough-processed aluminum board, wire mesh, and leather products. This approach enables robust measurement of previously difficult-to-measure objects Ob, such as mirror-surface objects Ob, transparent objects Ob, and objects Ob with uneven surfaces.

[0036] Figure 4 Schematic diagram showing the structure of the neural network in the learning completed model 5. Figure 4 As shown, the learned model 5 includes an input layer 51 , an intermediate layer 52 composed of at least one layer, and an output layer 53 .

[0037] The input layer 51 receives as input signals the photocurrent J value output from each light-receiving element 32 or a calculated value obtained by performing a predetermined calculation on the photocurrent J. Figure 4 In the figure, there are at least eight input terminals (white circles in the figure) as the input layer 51, but this is just an example and is not limited to this.

[0038] The middle layer 52 is the layer that performs the main processing of the neural network, and the number of layers is not particularly limited. Multiple neurons on each layer (white circles in the figure) are connected to multiple neurons on the next layer. For example, a neuron in the first layer of the middle layer 52 receives input by being weighted (loaded) by each terminal of the input layer 51, and then weights each neuron in the second layer of the middle layer 52 and sends the calculation result. Then, a neuron in the second layer of the middle layer 52 receives input by being weighted by each neuron in the first layer of the middle layer 52, and then weights each neuron in the third layer of the middle layer 52 and sends the calculation result. Such processing is sequentially performed according to the number of intermediate layers 52 .

[0039] The output layer 53 is a layer that outputs the output result of the learned model 5. Figure 4 , two output terminals (white circles in the figure) are shown as output layer 53, but this is merely an example and is not limiting. Preferably, distance d and angles θ and φ are output as spatial physical quantities related to the reference surface and object Ob on sensor 1, respectively. Specifically, processor 43 in information processing unit 4 is configured to input the value or calculated value of photocurrent J as an input parameter to learned model 5 and measure spatial physical quantities based on the output from learned model 5.

[0040] 4. Measurement process including information processing methods In this section, the flow of a measurement process of the measurement object Ob using the sensor 1 is described. Figure 5 1 is an activity diagram showing the flow of a measurement process of the object Ob using the sensor 1 . Figure 6 Schematic diagram showing the timing of each light emitting element 31 emitting light and the way in which the light receiving element 32 receives light. Figure 6 A. Figure 6 B. Figure 6 C and Figure 6 The light is emitted in the order of D. Hereinafter, the flow of the measurement process will be described with reference to this activity diagram.

[0041] In this embodiment, it is assumed that the sensor 1 has four light receiving blocks 3, and each of the light emitting elements 31 in the light receiving blocks 3 emits light at different timings. According to this scheme, the reflected light L caused by each light emitting element 31 can be easily distinguished. Hereinafter, it is expressed as "the kth (k = 1, 2, 3, 4) light emitting element 31 emits light." That is, the light emitting elements 31a, 31b, 31c, and 31d emit light periodically at different timings. First, let k = 1 (activity A001), and the first light emitting element 31, for example, Figure 6 As shown in A, the light emitting element 31a emits light (activity A002).

[0042] Next, light emitted from light-emitting element 31a irradiates object Ob, and the resulting reflected light L enters multiple light-receiving elements 32. Specifically, light-receiving element 32a receives reflected light L from light-emitting element 31a as main light (activity A003). Light-receiving element 32a is a light-receiving element 32 that forms a pair with light-emitting element 31a and is located in the same light-receiving block 3a as light-emitting element 31a. Light-receiving element 32b receives reflected light L from light-emitting element 31a as crosstalk light (activity A004). Light-receiving element 32b is a light-receiving element 32 located in a light-receiving block 3b adjacent to the left of the light-receiving block 3a in which light-emitting element 31a is located. Light-receiving element 32d receives reflected light L from light-receiving element 31a as crosstalk light (activity A005). Light-receiving element 32d is a light-receiving element 32 located in a light-receiving block 3d adjacent to the right of the light-receiving block 3a in which light-emitting element 31a is located. Then, each of the light-receiving elements 32a, 32b, and 32d outputs a photocurrent J corresponding to the received reflected light L. Thereafter, 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 corresponding light-emitting element 31 as main light, and receives the reflected light L other than the main light as crosstalk light in a manner distinguishable from the main light.

[0043] Next, since k<4, the value of k is incremented (activity A006), e.g. Figure 6 As shown in FIG. 8B , the second light emitting element 31 , ie, the light emitting element 31 b emits light (activity A002 ).

[0044] Next, light emitted from light-emitting element 31b irradiates object Ob, and the resulting reflected light L enters multiple light-receiving elements 32. Specifically, light-receiving element 32b receives reflected light L from light-emitting element 31b as main light (activity A003). Light-receiving element 32c receives reflected light L from light-emitting element 31b as crosstalk light (activity A004). Light-receiving element 32a receives reflected light L from light-emitting element 31b as crosstalk light (activity A005). Each of light-receiving elements 32b, 32c, and 32a then outputs a photocurrent J corresponding to the received reflected light L. The acquisition unit 431 in the information processing unit 4 then acquires each photocurrent J.

[0045] Next, since k<4, the value of k is incremented (activity A006), e.g. Figure 6 As shown in FIG. 3C , the third light emitting element 31 , ie, the light emitting element 31 c emits light (activity A002 ).

[0046] Next, light emitted from light-emitting element 31c irradiates object Ob, and the resulting reflected light L enters multiple light-receiving elements 32. Specifically, light-receiving element 32c receives reflected light L from light-emitting element 31c as main light (activity A003). Light-receiving element 32d receives reflected light L from light-emitting element 31c as crosstalk light (activity A004). Light-receiving element 32b receives reflected light L from light-emitting element 31c as crosstalk light (activity A005). Then, photocurrents J corresponding to the received reflected light L are output from light-receiving elements 32c, 32d, and 32b, respectively. The acquisition unit 431 in the information processing unit 4 then acquires each photocurrent J.

[0047] Next, since k<4, the value of k is incremented (activity A006), e.g. Figure 6 As shown in FIG. 2D , the fourth light emitting element 31 , ie, the light emitting element 31 d emits light (activity A002 ).

[0048] Next, light emitted from light-emitting element 31d irradiates object Ob, and the reflected light L enters multiple light-receiving elements 32. Specifically, light-receiving element 32d receives reflected light L from light-emitting element 31 as main light (activity A003). Light-receiving element 32a receives reflected light L from light-emitting element 31d as crosstalk light (activity A004). Light-receiving element 32c receives reflected light L from light-emitting element 31d as crosstalk light (activity A005). Then, each of light-receiving elements 32d, 32a, and 32c outputs a photocurrent J corresponding to the received reflected light L. The acquisition unit 431 in the information processing unit 4 then acquires each photocurrent J. Through the four light-emission cycles described above, photocurrents J for 12 channels are acquired.

[0049] This time, because k = 4, activities A007 to A009 are executed. Specifically, first, preprocessing is performed by the conversion unit 432 to convert each photocurrent J acquired by the acquisition unit 431 in the information processing unit 4 into a predetermined calculated value (activity A007). The predetermined calculated value is not particularly limited as long as it is calculated based on each photocurrent J. The number of channels for this calculated value can be the same as the number of channels of the photocurrent J (in one example, 12), or it can be less.

[0050] Next, the input processing unit 433 in the information processing unit 4 inputs the calculated values ​​obtained in activity A007 into the learned model 5 as input parameters (activity A008).

[0051] Finally, the output unit 434 of the information processing unit 4 outputs the spatial physical quantities estimated from the learned model 5, in this case, the distance d and the angles θ and φ (activity A009). The above processing is performed continuously as long as the sensor 1 is operating. Therefore, the distance d and the angles θ and φ can be measured sequentially based on the operating frequency of the sensor 1. The time from the first light-emitting element 31 emitting light to the estimated measurement of the distance d and the angles θ and φ can be 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 is, 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, or 0.1 milliseconds, or may be within a range between any two of the values ​​exemplified here. In other words, the time from when the kth light-emitting element 31 emits light to when the k+1th light-emitting element 31 emits light is preferably less than or equal to one-quarter of the aforementioned measurement time. More generally, the time from when the kth light-emitting element 31 emits light to when the k+1th light-emitting element 31 emits light is preferably less than or equal to the time obtained by dividing the time from when the first light-emitting element 31 emits light to when the distance d and angles θ and φ are estimated to be measured by the number of light-emitting elements 31.

[0052] In summary, the sensor 1 involved in 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 respectively arranged at different positions on the substrate 21. The light-receiving element 32 is arranged 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 the main light, and to receive the reflected light L other than the main light as the crosstalk light that can be distinguished from the main light. The reflected light L is light emitted from each light-emitting element 31 and reflected from the object Ob. Based on the main light and crosstalk light that can be distinguished and received, the sensor 1 measures spatial physical quantities related to the reference surface and the object Ob in the sensor 1.

[0053] Furthermore, from a different perspective, the information processing method includes the following steps. In the acquisition step, the 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 into the learned model 5. The learned model 5 is a model obtained by previously performing machine learning on the relationship between the spatial physical quantity related to the reference plane and the object Ob in the sensor 1 and the value of the photocurrent J or the calculated value. In the output step, the spatial physical quantity is estimated and output based on the learned model 5.

[0054] This form enables a highly accurate sensor with a simpler structure. In particular, by including crosstalk light, previously considered interference, as an input parameter for measurement, rather than solely using reflected light L from the corresponding relationship between light-emitting element 31a and light-receiving element 32a, more diverse information can be obtained, improving measurement accuracy.

[0055] [other] The sensor 1 involved in this embodiment can also be further developed.

[0056] The light emitting elements 31 and the light receiving elements 32 may not be arranged in a one-to-one relationship. Figure 7 This is a schematic front view showing the structure of the sensor 1 involved in the modification. The sensor 1 involved in the modification has two light-emitting elements 31x and 31y and one light-receiving element 32x as a minimum structure. Light emitted from the light-emitting element 31x is reflected from the object Ob, and the reflected light L1 (an example of main light) is received by the light-receiving element 32x. In addition, light emitted from the light-emitting element 31y is reflected from the object Ob, and the reflected light L2 (an example of crosstalk light) is received by the light-receiving element 32x. Even in this form of sensor 1, it is possible to estimate the distance d and the angle θ (the angle in one direction) using the learned model 5 that has been pre-machined.

[0057] In the above embodiment, the light-receiving and light-emitting blocks 3 are illustrated as including light-emitting elements 31 and light-receiving elements 32 in a one-to-one manner, and each light-receiving and light-emitting block 3 is arranged in a ring shape and at equal intervals on the substrate 21. However, the light-emitting elements 31 and light-receiving elements 32 may also be arranged completely randomly on the substrate 21. In this case, the distance d and the angles θ and φ can also be estimated by preparing a learned model 5 that has been previously machine-learned.

[0058] While the sensor 1 has been described as including a detection unit 2 and an information processing unit 4, it is also possible to implement the system in which the sensor 1 includes only the detection unit 2 and outputs measurement results using a computer (not shown) connected thereto. In this case, the learned model 5 is preferably stored in a storage unit (not shown) built into the computer (not shown), and the input processing to the learned model 5 is performed by a processor (not shown). The connection method is not particularly limited and can be directly connected using wires, or the sensor 1 can be equipped with wireless communication capabilities, and the sensor 1 and the computer (not shown) can be connected via wireless communication. In this case, the system can also be implemented in which a single computer (not shown) is connected to multiple sensors 1.

[0059] The sensor 1 can also be manufactured using a predetermined manufacturing method. This manufacturing method preferably includes the steps of disposing a plurality of light-emitting elements 31 and at least one light-receiving element 32 on a substrate 21. For example, the light-emitting elements 31 may be disposed on the substrate 21, followed by disposing the light-receiving element 32, and then connecting the output of the light-receiving element 32 to an information processing unit 4, such as a microcontroller. This method allows for the manufacture of a sensor with a simpler structure and high precision.

[0060] In the above embodiment, each light-emitting element 31 emits light at different timings to separate each photocurrent J obtained, but instead of timing, each light-emitting element 31 can be flashed at a different frequency, or the wavelength of the light emitted by the light-emitting element 31 can be made different to separate each photocurrent J obtained.

[0061] This may be implemented so that not only the left and right adjacent light receiving elements 32 but also the opposing light receiving element 32 receive the crosstalk light. That is, the position of the light receiving element 32 using the crosstalk light can be freely determined according to the number of light receiving and emitting blocks 3 .

[0062] After learning, model 5 can replace the neural network, which can be a one-dimensional convolutional network or a two-dimensional convolutional network.

[0063] The electrical quantity generated from the light receiving element 32 is not limited to the photocurrent J, and may be, for example, a voltage.

[0064] Furthermore, it can also be provided in the various forms described below.

[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 arranged at different positions on a substrate, the light-receiving element being arranged on the substrate and receiving one of the reflected lights generated by each light-emitting element as main light. The reflected light is light emitted from each light-emitting element and reflected from an object, the sensor being capable of distinguishing the main light from the crosstalk light. A spatial physical quantity related to a reference surface and the object is measured on the sensor.

[0066] (2) The sensor according to (1) above, wherein the spatial physical quantity includes at least a distance between the reference plane and the object and an angle of the object relative to the reference plane.

[0067] (3) In the sensor described in paragraph (2) above, the distance is 50 mm or less.

[0068] (4) In the sensor according to any one of (1) to (3) above, the number of the photodetectors is the same as the number of the light-emitting elements, and the photodetectors are configured to be associated with each of the light-emitting elements on a one-to-one basis, and receive the reflected light caused by the light-emitting element associated therewith as the main light, and receive the reflected light other than the main light as the light, and the light is configured to receive the reflected light other than the main light as crosstalk light, distinguished from the main light.

[0069] (5) In the sensor according to (4) above, the light receiving element and the corresponding light emitting element are adjacent to each other to form a light receiving and light emitting block, and a plurality of light receiving and light emitting blocks are provided on the substrate.

[0070] (6) In the sensor described in (5) above, each of the light receiving blocks and the light emitting blocks is arranged in a ring shape at equal intervals on the substrate.

[0071] (7) The sensor according to any one of (1) to (6) above, wherein the number of light-emitting elements is 2 to 4.

[0072] (8) In the sensor according to any one of (1) to (7) above, each light emitting element emits light at different times from each other.

[0073] (9) In the sensor according to any one of (1) to (8) above, the light receiving element is configured to generate an electrical physical quantity based on the reflected light, and to measure a spatial physical quantity related to the object based on each of the electrical physical quantities based on the distinguishably received main light and the crosstalk light, a reference surface in the sensor, and each of the electrical physical quantities based on the distinguishably received main light and the crosstalk light.

[0074] (10) The sensor according to (9) above further includes an artificial intelligence input unit, wherein the artificial intelligence input unit inputs the value of the electric physical quantity or the calculated value based on the electric physical quantity as an input parameter into the learning model, wherein the learning model is a model in which the relationship between the spatial physical quantity related to the reference surface and the object in the sensor and the electric physical quantity value or the calculated value is obtained in advance through machine learning.

[0075] (11) The sensor according to (10) above further comprises a storage unit / memory and a processor, wherein the storage unit / memory stores the learning model, and the processor is configured to input the electrical physical quantity value or the calculated value as an input parameter into the learning model and measure the spatial physical quantity based on an output result of the learning model. The electrical physical quantity value or the calculated value is input as an input parameter into the learning model and is configured to measure the spatial physical quantity based on an output result of the learning model.

[0076] (12) In the sensor described in any one of (1) to (11) above, the light receiving element is a light receiving element having wide directivity.

[0077] (13) A method for manufacturing a sensor, comprising the steps of arranging a plurality of light-emitting elements and at least one light-receiving element on a base material, wherein the sensor is a sensor manufactured according to any one of (1) to (12) above.

[0078] (14) An information processing method comprises the following steps: an acquisition step of acquiring an electric physical quantity output from a light receiving element; an input processing step of inputting the value of the electric physical quantity or a calculated value based on the electric physical quantity as an input parameter into a learning model, wherein the learning model is a pre-machine learning model of the relationship between a spatial physical quantity related to a reference surface and an object at a sensor and the value or calculated value of the electric physical quantity, and wherein an output step infers the output spatial physical quantity based on the learning model. Of course, it is not limited to this.

[0079] Finally, although various embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These new embodiments may be implemented in various other ways, and various omissions, substitutions, and modifications may be made without departing from the spirit of the invention. These embodiments and their variations are intended to be within the scope and spirit of the invention and are intended to be included in the invention recited in the claims and their equivalents. Description of Reference Numerals

[0080] 1: sensor; 2: detection unit; 21: substrate; 22: AI input unit; 3: light receiving block; 3a: light receiving block; 3b: light receiving block; 3c: light receiving block; 3d: light receiving 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: light receiving element; 32a: light receiving element; 32b: light receiving element; 32c: light receiving element; 32 d: light-receiving element; 32x: light-receiving element; 4: information processing unit; 40: communication bus; 41: communication unit; 42: storage unit; 43: processor; 431: acquisition unit; 432: conversion unit; 433: input processing unit; 434: output unit; 5: learned model; 51: input layer; 52: intermediate layer; 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, characterized in that: A plurality of light emitting elements and at least one light receiving element are provided. The light emitting elements are respectively arranged at different positions on the substrate. The light receiving element is disposed on the substrate, receives one of the reflected lights caused by each light emitting element as main light, and receives reflected light other than the main light distinguishable from the main light as crosstalk light, wherein the reflected light is light emitted from each of the light emitting elements and reflected from the object, respectively. Based on the distinguishably received main light and the crosstalk light, a spatial physical quantity related to a reference surface in the sensor and the object is measured.

2. The sensor according to claim 1, characterized in that The spatial physical quantity includes at least a distance between the reference plane and the object and an angle of the object relative to the reference plane.

3. The sensor according to claim 2, characterized in that The distance is less than 50 mm.

4. The sensor according to any one of claims 1 to 3, characterized in that: The light receiving element is composed of: The number of the light emitting elements is the same as that of the light emitting elements, and the light emitting elements correspond one-to-one to each of the light emitting elements. The reflected light caused by the corresponding light emitting element is received as the main light, and the reflected light other than the main light is received as the crosstalk light in a manner distinguishable from the main light.

5. The sensor according to claim 4, characterized in that The light receiving element and the corresponding light emitting element are adjacent to each other to form a light receiving and light emitting block. A plurality of the light receiving and emitting blocks are disposed on the substrate.

6. The sensor according to claim 5, characterized in that Each of the light receiving and emitting blocks is arranged in a ring shape and at equal intervals on the substrate.

7. The sensor according to any one of claims 1 to 6, characterized in that: The number of the light emitting elements is 2 to 4.

8. The sensor according to any one of claims 1 to 7, characterized in that: Each of the light emitting elements emits light at a timing different from each other.

9. The sensor according to any one of claims 1 to 8, characterized in that: The light receiving element is configured to generate an electrical physical quantity corresponding to the reflected light. A spatial physical quantity related to a reference surface and the object in the sensor is measured based on the electrical physical quantities based on the distinguishably received main light and the crosstalk light.

10. The sensor according to claim 9, characterized in that It also has an AI input unit. The AI ​​input unit inputs the value of the electrical physical quantity or the calculated value based on the electrical physical quantity as an input parameter into the learned model, wherein the learned model is a model obtained by pre-machine learning of the relationship between the spatial physical quantity related to the reference plane in the sensor and the object and the value of the electrical physical quantity or the calculated value.

11. The sensor according to claim 10, characterized in that It also includes a storage unit and a processor, The storage unit stores the learned model. The processor is configured to input the value of the electrical physical quantity or the calculated value as the input parameter into the learned model, and measure the spatial physical quantity based on an output result from the learned model.

12. The sensor according to any one of claims 1 to 11, characterized in that: The light receiving element is a light receiving element having wide directivity.

13. A method for manufacturing a sensor, characterized in that: The method comprises the steps of disposing a plurality of light emitting elements and at least one light receiving element on a substrate, wherein the sensor is the sensor according to any one of claims 1 to 12.

14. An information processing method, characterized in that: The following steps are involved: 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 the calculated value based on the electrical physical quantity is input as an input parameter into a learned model, wherein the learned model is a model obtained by mechanically learning in advance the relationship between a spatial physical quantity related to a reference plane and an object in a sensor 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 learned model.

Images