Control device, light-receiving device, learning device, control method, and program

Abstract

This control device controls a light-receiving device having a deformable mirror and a photoelectric conversion unit, and includes: a monitor unit that monitors a voltage value of each cell in the photoelectric conversion unit which receives a beam reflected by the deformable mirror; a prediction unit that predicts, on the basis of the voltage value of each of the cells, wavefront information necessary for obtaining a desired beam pattern in the photoelectric conversion unit; and a mirror control unit that controls the deformable mirror by using the wavefront information predicted by the prediction unit.

Description

Control device, light receiving device, learning device, control method, and program

[0001] The present invention relates to optical wireless power supply technology.

[0002] Non-Patent Document 1 discloses optical wireless power supply technology. In optical wireless power supply technology, for example, a laser is used as an energy medium. Laser light (also called a laser beam or a beam) is transmitted from the laser toward a target to be powered, and the target converts the laser light into electricity using a photoelectric conversion element such as a solar cell. Here, a light source (light source element) that outputs laser light is called a "laser." A "laser" may also be called a laser medium, a laser oscillator, or the like.

[0003] Furthermore, Non-Patent Document 2 discloses a wavefront compensation technique. The wavefront compensation technique is a technique for improving spatial resolution in astronomical observation by compensating for wavefront disturbances caused by atmospheric turbulence, etc. In addition to astronomical observation, the wavefront compensation technique is also effective in situations where wavefront disturbances deteriorate performance, such as optical microscopes and optical wireless communication.

[0004] To supply high-power electricity, it is expected that a multi-cell type photoelectric conversion unit equipped with multiple cells (photoelectric conversion elements) will be used as the photoelectric conversion unit that converts the received beam into electricity. In a multi-cell type photoelectric conversion unit, it is desirable that the beam (light) is uniformly incident on each cell so as not to reduce the efficiency of photoelectric conversion.

[0005] Ke Jin, Weiyang Zhou, "Wireless Laser Power Transmission: A Review of Recent Progress", IEEE TRANSACTIONS ON POWER ELECTRONICS, APRIL 2019, VOL. 34, NO. 4 https: / / ieeexplore.ieee.org / stamp / stamp.jsp?tp=&arnumber=8404085&tag=1 Hideki Takami, "Wavefront Adaptive Optics (AO)", Journal of the Japan Society for Precision Engineering, 2001, Vol. 67, No. 10, 1584 https: / / www.jstage.jst.go.jp / article / jjspe1986 / 67 / 10 / 67_10_1584 / _pdf

[0006] In particular, when a beam output from a light source element propagates through the atmosphere, such as in outdoor optical wireless power supply, atmospheric turbulence disrupts the wavefront of the beam, causing uneven intensity distribution or positional deviation of the beam irradiating the photoelectric conversion unit, which leads to a decrease in photoelectric conversion efficiency.

[0007] Therefore, it is conceivable to use the wavefront compensation technology disclosed in Non-Patent Document 1 to compensate the wavefront of the beam before photoelectric conversion of the beam, thereby suppressing the effects of atmospheric turbulence and irradiating the photoelectric conversion element with a beam having a clear beam pattern.

[0008] However, when using conventional wavefront compensation techniques, there is a problem in that energy loss occurs when obtaining wavefront information.

[0009] The present invention has been made in consideration of the above points, and aims to provide a technique for irradiating a photoelectric conversion unit with a beam of a desired beam pattern in an optical wireless power supply system without causing energy loss for obtaining wavefront information.

[0010] According to the disclosed technology, there is provided a control device that controls a light receiving device having a deformable mirror and a photoelectric conversion unit, comprising: a monitor unit that monitors the voltage value of each cell in the photoelectric conversion unit that receives the beam reflected by the deformable mirror; a prediction unit that predicts wavefront information required to obtain a desired beam pattern in the photoelectric conversion unit based on the voltage value of each cell; and a mirror control unit that controls the deformable mirror using the wavefront information predicted by the prediction unit.

[0011] According to the disclosed technology, a technology is provided for irradiating a photoelectric conversion unit with a beam of a desired beam pattern without causing energy loss for obtaining wavefront information in an optical wireless power supply system.

[0012] FIG. 1 is a diagram for explaining the problem. FIG. 2 is a diagram for explaining an example of cell connection. FIG. 3 is a diagram for explaining the problem. FIG. 4 is a diagram for explaining an example of a configuration of an optical wireless power supply system according to an embodiment of the present invention. FIG. 5 is a diagram for explaining an example of a configuration of a light transmitting device 100. FIG. 6 is a flowchart for explaining an operation flow. FIG. 7 is a diagram for explaining an example of a configuration of a control device 100. FIG. 8 is a diagram for explaining an example of a configuration of a learning device 200. FIG. 9 is a diagram for explaining an example of wavefront compensation. FIG. 10 is a diagram for explaining an example of an example of wavefront compensation. FIG. 11 is a diagram for explaining an example of a hardware configuration of a device.

[0013] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. The embodiment described below is merely an example, and the embodiment to which the present invention is applied is not limited to the following embodiment.

[0014] In the embodiment described below, it is assumed that a laser is used as the light source in the light transmitting device, but using a laser as the light source is merely an example, and the light source to which the technology according to the present invention can be applied is not limited to a specific type of light source.

[0015] In the following, first, the problems associated with the technology of the present embodiment will be described in more detail, and then the technology of the present embodiment will be described. Note that the content of the following description of the problems is not publicly known.

[0016] (About the Issues) Optical wireless power transfer technology is available in two types: a single-cell type with a single photoelectric conversion element cell as the photoelectric conversion unit that converts received light into electricity, and a multi-cell type with multiple cells. The voltage that can be extracted from a single cell is determined by the band gap of the element, and is 0.5 V for a typical solar cell. Therefore, when attempting to supply high-power electricity, the current value becomes large, making the single-cell type unsuitable.

[0017] On the other hand, in the case of a multi-cell type, connecting the cells in series makes it possible to increase the voltage and thus the power that can be extracted. However, if there is a cell among the multiple cells that does not receive enough light, the current value of that cell becomes a bottleneck, limiting the overall current value and preventing high power from being obtained. Furthermore, the light energy that enters the other cells cannot be extracted as power, but rather turns into heat, reducing the photoelectric conversion efficiency. Therefore, in optical wireless power transfer, it is desirable for light to be uniformly incident on each cell in the photoelectric conversion section.

[0018] In particular, when a beam output from a light source element propagates through the atmosphere, such as when optical wireless power transmission is performed outdoors, atmospheric turbulence disrupts the wavefront of the beam, causing uneven intensity distribution or positional deviation of the beam irradiating the photoelectric conversion unit. This leads to a decrease in photoelectric conversion efficiency. Note that a wavefront is a surface on which the phase of light (electric field) is the same.

[0019] The above-mentioned problem will be explained in more detail with reference to FIG. 1. In the configuration shown in FIG. 1, a laser 1, which is a light source element, and a photoelectric conversion unit 2 on the light receiving side are shown. As for the laser 1, a surface 3 (light emitting surface) on the side from which the laser 1 outputs light is also shown. As for the photoelectric conversion unit 2, a surface (light receiving surface) on the side from which light is received is also shown. As shown in FIG. 1, the photoelectric conversion unit 2 has a plurality of cells (photoelectric conversion elements) arranged in a lattice pattern. The plurality of cells are connected in series, for example, as shown in FIG. 2.

[0020] As described above, atmospheric turbulence (disturbance) disturbs the wavefront of the beam, causing non-uniformity in the intensity distribution (which may also be called the beam pattern) of the beam irradiated onto the photoelectric conversion unit 2. An image of a beam pattern disturbed by disturbance is shown on the photoelectric conversion unit 2 in Figure 1.

[0021] Therefore, it is effective to use wavefront compensation technology to compensate the wavefront of the beam before photoelectric conversion, thereby suppressing the effects of atmospheric turbulence and irradiating the photoelectric conversion element with a clear beam pattern. Note that wavefront compensation means correcting the wavefront to obtain a desired wavefront.

[0022] An example configuration of a light receiving device 20 equipped with a mechanism for compensating the wavefront of a beam before photoelectric conversion of the beam using wavefront compensation technology is shown in Fig. 3. The configuration in Fig. 3 is a configuration that is assumed when conventional wavefront compensation technology is applied to optical wireless power transfer, and the configuration in Fig. 3 is not publicly known.

[0023] The light receiving device 20 shown in Fig. 3 includes a photoelectric conversion unit 2, a wavefront sensor 4, a deformable mirror 5, and a beam splitter 6. A beam transmitted from a light transmitting device is incident on the deformable mirror 5, reflected by the deformable mirror 5, and reaches the photoelectric conversion unit 2 via the beam splitter 6. As the beam passes through the beam splitter 6, a portion of the beam is reflected and reaches the wavefront sensor 4. The wavefront sensor 4 monitors the wavefront of the beam and, based on the monitoring results, provides feedback to the deformable mirror 5 to correct the wavefront. Based on the feedback, the deformable mirror 5 deforms, so that the beam pattern of the beam reaching the photoelectric conversion unit 2 becomes a desired beam pattern.

[0024] 3 can compensate for the wavefront of the beam incident on the photoelectric conversion unit 2. However, in a configuration using general adaptive optics technology as shown in FIG. 3, a portion of the beam is extracted for use by the wavefront sensor 4 in order to obtain wavefront information, which causes an energy loss in the beam that reaches the photoelectric conversion unit 2.

[0025] In the fields of astronomical observation and optical wireless communication, energy loss is often not a problem, but in the field of optical wireless power transfer, energy loss affects performance, so it is necessary to minimize energy loss as much as possible.

[0026] Furthermore, when powering flying objects such as drones, miniaturization and weight reduction of the system are required, and it is undesirable to equip the system with a wavefront sensor for wavefront compensation in addition to the photoelectric conversion unit.Furthermore, in long-distance optical wireless power transfer systems such as space solar power generation, the beam diameter becomes large to suppress the laser divergence angle, and a large-diameter wavefront sensor is required to obtain beam wavefront information, which increases costs.

[0027] The following describes an optical wireless power transfer system that can solve the above problems and achieve wavefront compensation with low loss. The optical receiving device in this system can be made small and lightweight.

[0028] (Configuration and operation of system / device) Fig. 4 shows a configuration example of an optical wireless power supply system according to this embodiment. As shown in Fig. 4, the optical wireless power supply system according to this embodiment includes a light transmitting device 10 and a light receiving device 20. The light transmitting device 10 is a general light transmitting device equipped with a laser or the like as a light source. In addition, this embodiment uses a learning device 200 for learning a machine learning model, which will be described later.

[0029] Fig. 5 shows an example of the configuration of the light receiving device 20. As shown in Fig. 5, the light receiving device 20 according to this embodiment includes a photoelectric conversion unit 2, a deformable mirror 5, and a control device 100. The control device 100 may also be called a control unit. As shown in Fig. 5, unlike the configuration in Fig. 3, the wavefront sensor 4 and the beam splitter 6 are not present. As in the example shown in Fig. 1, the photoelectric conversion unit 2 has a plurality of cells (photoelectric conversion elements), and the plurality of cells are connected in series, for example.

[0030] The beam transmitted from the light transmitting device 10 is incident on the deformable mirror 5 , reflected by the deformable mirror 5 , and then incident on the photoelectric conversion unit 2 .

[0031] Each cell is equipped with a voltage sensor and a current sensor. In the following description, the voltage and current values ​​of each cell are monitored; however, if multiple cells are connected in series, the current values ​​of each cell will be the same, so only the voltage value may be monitored. In this case, only a voltage sensor is attached to each cell. When only the voltage value is monitored, wavefront information required to obtain a desired beam pattern is predicted based on the voltage value of each cell. The operation of the control device 100 will be described with reference to the flowchart of FIG. 6.

[0032] In S101 (step 101), the control device 100 monitors the voltage and current values ​​of each cell. This allows the control device 100 to reproduce a beam pattern (the intensity distribution of light on the photoelectric conversion unit 2) and, based on the beam pattern, predict wavefront information required to obtain a desired beam pattern (i.e., a uniform intensity distribution) (S102). This wavefront information is, for example, information on a disturbed wavefront (such as the shape of the wavefront) incident on the deformable mirror 5. Knowing the information on the disturbed wavefront allows the amount of deformation of the deformable mirror 5 required to obtain a wavefront without disturbance to be determined. Alternatively, the wavefront information predicted in S102 may be the amount of deformation of the deformable mirror 5 required to obtain a desired beam pattern.

[0033] In S103, the control device 100 deforms the deformable mirror 5 using the wavefront information predicted in S102.

[0034] The control of steps S101 to S103 is performed continuously (always) while optical wireless power feeding is being performed. Alternatively, the control of steps S101 to S103 may be performed at certain time intervals. Furthermore, monitoring may be performed continuously, and the deformation control of the deformable mirror 5 may be performed only when a disturbance in the intensity distribution (for example, a disturbance of a magnitude equal to or greater than a threshold) is detected.

[0035] The above control by the control device 100 allows the photoelectric conversion unit 2 to be irradiated with a desired beam with reduced disturbance in intensity distribution, thereby improving the photoelectric conversion efficiency.

[0036] (Configuration example of control device 100) The control device 100 may be located inside the light receiving device 200 or outside the light receiving device 200. Fig. 7 shows an example of the functional configuration of the control device 100. As shown in Fig. 7, the control device 100 includes a monitor unit 110, a prediction unit 120, a mirror control unit 130, and a data storage unit 140.

[0037] The monitor unit 110 monitors the voltage value and current value of each cell of the photoelectric conversion unit 2. As described above, the monitor unit 110 may monitor only the voltage value of each cell of the photoelectric conversion unit 2.

[0038] For example, trained parameters of a machine learning model (described later) are stored in the data storage unit 140. The prediction unit 120 reads the trained parameters from the data storage unit 140, and predicts wavefront information from the voltage and current values ​​(or only the voltage values) of each cell using the machine learning model to which the trained parameters are set, and passes the wavefront information to the mirror control unit 130.

[0039] The mirror control unit 130 uses the wavefront information predicted by the prediction unit 120 to deform the deformable mirror 5 so as to generate a desired wavefront.

[0040] (Method for Predicting Wavefront Information) The method for predicting wavefront information from the beam pattern (the intensity distribution of the beam on the surface of the photoelectric conversion unit) etc. is not limited to a specific method, but in this embodiment, the wavefront information is predicted using machine learning. Note that in the following description, voltage values ​​and current values ​​are used, but it is also possible to use only voltage values.

[0041] In this embodiment, a simulation is used in advance to calculate the beam pattern in the photoelectric conversion unit 2 for each of the multiple beams when affected by disturbance, and to calculate the voltage value and current value of each cell in the photoelectric conversion unit 2 corresponding to each beam pattern. These calculations may be performed by the learning device 200 or another device.

[0042] The learning device 200 stores each beam pattern and the corresponding "voltage value and current value of each cell" as learning data (correct answer data), and learns a machine learning model that outputs a beam pattern from the "voltage value and current value of each cell."

[0043] In this learning, the learning device 200 inputs the "voltage and current values ​​of each cell" into a machine learning model such as a neural network, and optimizes the parameters of the machine learning model so that the beam pattern output from the machine learning model becomes the correct beam pattern. This machine learning model is referred to as machine learning model A.

[0044] Furthermore, the learning device 200 learns wavefront information corresponding to the beam pattern. This wavefront information may be information about the wavefront before compensation, or may be control information (e.g., the amount of deformation) to be given to the deformable mirror 5 so that the beam propagating from the deformable mirror 5 to the photoelectric conversion unit 2 will have a desired beam pattern (i.e., a beam pattern without distortion) on the photoelectric conversion unit 2.

[0045] In this learning, a large amount of learning data is prepared by, for example, simulation, which includes beam patterns before compensation and wavefront information for obtaining a desired beam pattern.

[0046] The learning device 200 holds the above learning data, inputs the uncompensated beam pattern to a machine learning model such as a neural network, and optimizes the parameters of the machine learning model so that the wavefront information output from the machine learning model becomes correct wavefront information (wavefront information necessary to obtain a desired beam pattern). This machine learning model is referred to as machine learning model B.

[0047] The prediction unit 120 of the control device 100 includes a machine learning model A and a machine learning model B. By using the machine learning model A, the prediction unit 120 can reproduce a beam pattern from the observed values ​​of the photoelectric conversion unit 2 (the voltage value and current value of each cell), and by using the machine learning model B, can predict wavefront information required to compensate for the wavefront of this beam pattern.

[0048] In the above example, two machine learning models, machine learning model A and machine learning model B, are used, but it is also possible to use a single machine learning model, which will be referred to as machine learning model C.

[0049] In this case, for each of the multiple beams, a simulation is used to calculate the voltage and current values ​​of each cell of the photoelectric conversion unit 2 when affected by disturbance. Also, a large amount of training data (correct answer data) is prepared, for example, by simulation, which includes the voltage and current values ​​of each cell before compensation and wavefront information for obtaining a desired beam pattern.

[0050] Next, the machine learning model C is trained so as to obtain wavefront information for the "voltage value and current value of each cell." Specifically, the learning device 200 holds the above-mentioned training data, inputs the "voltage value and current value of each cell" before compensation to the machine learning model C such as a neural network, and optimizes the parameters of the machine learning model C so that the wavefront information output from the machine learning model becomes the correct wavefront information.

[0051] In this case, the prediction unit 120 of the control device 100 includes a machine learning model C. By using the machine learning model C, the prediction unit 120 can predict wavefront information necessary for wavefront correction from the observed values ​​of the photoelectric conversion unit 2 (the voltage value and current value of each cell).

[0052] Note that machine learning model C may be a combination of machine learning model A and machine learning model B.

[0053] 8 shows an example of the functional configuration of the learning device 200. As shown in FIG. 8, the learning device 200 includes an input unit 210, a learning unit 220, an output unit 230, and a data storage unit 240.

[0054] Learning data (correct answer data) to be used for learning is input from the input unit 210 and stored in the data storage unit 240. The learning unit 220 holds the machine learning model (including parameters) to be learned, and learns the machine learning model using the learning data read from the data storage unit 240. In the case of a neural network, learning can be performed using, for example, backpropagation.

[0055] The output unit 230 outputs the trained machine learning model (specifically, the trained parameters). The trained machine learning model (specifically, the trained parameters) is stored in the data storage unit 140 of the control device 100.

[0056] The learning device 200 and the control device 100 may be the same device. For example, the control device 100 may include a learning unit 220, so that the control device 100 can predict wavefront information using a machine learning model that the control device 100 has learned.

[0057] Next, an example will be described. In this example, the flow of processing related to wavefront compensation in the light receiving device 20 will be described more specifically with reference to FIGS.

[0058] In this embodiment, 16 cells are connected in series in the photoelectric conversion unit 2 of the light receiving device 20 .

[0059] First, let us assume that a beam such as that shown in Fig. 9 (image) is irradiated onto the photoelectric conversion unit 2. Because multiple cells are connected in series, the current value flowing through each cell is equal, but because the amount of light received by each cell differs, the voltage value of cells that receive a large amount of light is high, and the voltage value of cells that receive a small amount of light is low.

[0060] When the control device 100 reproduces the intensity distribution (beam pattern) of the beam irradiated to the photoelectric conversion unit 2 using machine learning model A from the voltage and current values ​​of each cell, the beam pattern shown in Figure 10 is obtained.

[0061] When the control device 100 performs wavefront compensation on this beam pattern using machine learning model B to make the intensity distribution uniform, a beam pattern with a nearly uniform intensity distribution can be obtained, as shown in Fig. 11. By irradiating the photoelectric conversion unit 2 with a beam having such a beam pattern, the photoelectric conversion efficiency can be improved compared to when a beam having a non-uniform intensity distribution is irradiated.

[0062] (Example of Hardware Configuration of Device) Both the control device 100 and the learning device 200 described in this embodiment can be realized, for example, by causing a computer to execute a program. This computer may be a physical computer or a virtual machine on the cloud.

[0063] That is, the device can be realized by executing a program corresponding to the processing performed by the device using hardware resources such as a CPU, GPU, and memory built into a computer. The program can be recorded on a computer-readable recording medium (such as a portable memory) and stored or distributed. The program can also be provided via a network such as the Internet or email.

[0064] Fig. 12 is a diagram showing an example of the hardware configuration of the computer. The computer in Fig. 12 includes a drive device 1000, an auxiliary storage device 1002, a memory device 1003, a CPU 1004, an interface device 1005, a display device 1006, an input device 1007, an output device 1008, and the like, all of which are interconnected by a bus BS. The computer may further include a GPU.

[0065] The program that realizes the processing on the computer is provided by a recording medium 1001, such as a CD-ROM or a memory card. When the recording medium 1001 storing the program is set in the drive device 1000, the program is installed from the recording medium 1001 to the auxiliary storage device 1002 via the drive device 1000. However, the program does not necessarily have to be installed from the recording medium 1001, but may be downloaded from another computer via a network. The auxiliary storage device 1002 stores the installed program as well as necessary files, data, etc.

[0066] The memory device 1003 reads and stores a program from the auxiliary storage device 1002 when an instruction to start the program is received. The CPU 1004 realizes functions related to the device in accordance with the program stored in the memory device 1003. The interface device 1005 is used as an interface for connecting to a network, etc. The display device 1006 displays a GUI (Graphical User Interface) or the like according to the program. The input device 1007 is composed of a keyboard, mouse, buttons, a touch panel, etc., and is used to input various operation instructions. The output device 1008 outputs the results of calculations.

[0067] (Summary of the embodiment) In this embodiment, the control device 300 reproduces a beam pattern from the output result of the photoelectric conversion unit 2, and compensates the wavefront based on the beam pattern so as to obtain a desired beam pattern. In conventional wavefront compensation technology, a part of the beam is extracted to obtain wavefront information, which causes energy loss that is undesirable in the field of optical wireless power transfer.

[0068] In contrast, the technology according to this embodiment does not lose energy in obtaining wavefront information, and a highly efficient power supply system can be realized. Furthermore, since a wavefront sensor or beam splitter for wavefront compensation is not required, the light-receiving device 20 can be made small and lightweight. This makes it effective in situations where space and payload are limited, such as power supply to automobiles or flying objects. In long-distance optical wireless power supply, where it is difficult to prepare a large wavefront sensor or beam splitter to match the beam diameter, the technology according to this embodiment is also effective in terms of cost reduction.

[0069] (Effects of the Technology According to the Embodiments) As described above, the technology according to the present embodiment uses the photoelectric conversion unit 2 instead of a wavefront sensor, thereby eliminating the need for power to obtain wavefront information and suppressing the energy loss that occurs in conventional wavefront compensation technology. Furthermore, because components such as a beam splitter and a wavefront sensor are not required, the light-receiving device 20, which is composed of the photoelectric conversion unit 2, the deformable mirror 5, etc., can be made smaller and lighter.

[0070] In conventional methods of wavefront compensation, compensation is mainly performed to cancel out the disturbed wavefront observed by a wavefront sensor. However, the technology of this embodiment can improve photoelectric conversion efficiency by compensating the wavefront so that the beam irradiated to the photoelectric conversion unit becomes the desired beam without using a wavefront sensor.

[0071] The following additional notes are provided regarding the above-described embodiments.

[0072] <Additional Notes> (Additional Item 1) A control device that controls a light receiving device having a deformable mirror and a photoelectric conversion unit, comprising: a memory; and at least one processor connected to the memory, wherein the processor monitors a voltage value of each cell in the photoelectric conversion unit that receives a beam reflected by the deformable mirror, predicts wavefront information required to obtain a desired beam pattern in the photoelectric conversion unit based on the voltage value of each cell, and controls the deformable mirror using the predicted wavefront information. (Additional Item 2) The control device according to Additional Item 1, wherein the processor predicts the wavefront information using a machine learning model. (Additional Item 3) The light receiving device according to Additional Item 1, comprising the control device according to Additional Item 1. (Supplementary Item 4) A learning device that trains a machine learning model that predicts wavefront information for obtaining a desired beam pattern in a photoelectric conversion unit of a light receiving device having a photoelectric conversion unit that receives a beam reflected by a deformable mirror, the learning device including: a memory; and at least one processor connected to the memory, wherein the memory holds a voltage value of each cell in the photoelectric conversion unit and the wavefront information for obtaining the desired beam pattern as learning data, and the processor trains the machine learning model so that the machine learning model receives the voltage value of each cell as an input and outputs the wavefront information. (Supplementary Item 5) A control method executed by a control device that controls a light receiving device having a deformable mirror and a photoelectric conversion unit, the control method comprising: a monitoring step of monitoring the voltage value of each cell in the photoelectric conversion unit that receives the beam reflected by the deformable mirror; a prediction step of predicting wavefront information necessary for obtaining a desired beam pattern in the photoelectric conversion unit based on the voltage value of each cell; and a mirror control step of controlling the deformable mirror using the wavefront information predicted by the prediction step. (Supplementary Item 6) A non-transitory storage medium storing a program for causing a computer to function as each unit in the control device according to Supplementary Item 1 or 2.

[0073] Although the present embodiment has been described above, the present invention is not limited to such a specific embodiment, and various modifications and changes are possible within the scope of the gist of the present invention described in the claims.

[0074] REFERENCE SIGNS LIST 1 Laser 2 Photoelectric conversion unit 3 Light emitting surface 4 Wavefront sensor 5 Deformable mirror 6 Beam splitter 10 Light transmitting device 20 Light receiving device 100 Control device 110 Monitor unit 120 Prediction unit 130 Mirror control unit 140 Data storage unit 200 Learning device 210 Input unit 220 Learning unit 230 Output unit 240 Data storage unit 1000 Drive device 1001 Recording medium 1002 Auxiliary storage device 1003 Memory device 1004 CPU 1005 Interface device 1006 Display device 1007 Input device 1008 Output device

Claims

1. A control device that controls a light receiving device having a deformable mirror and a photoelectric conversion unit, comprising: a monitor unit that monitors the voltage values ​​of each cell in the photoelectric conversion unit that receives the beam reflected by the deformable mirror; a prediction unit that predicts wavefront information required to obtain a desired beam pattern in the photoelectric conversion unit based on the voltage values ​​of each cell; and a mirror control unit that controls the deformable mirror using the wavefront information predicted by the prediction unit.

2. The control device according to claim 1, wherein the prediction unit predicts the wavefront information using a machine learning model.

3. The light receiving device according to claim 1, comprising the control device according to claim 1.

4. A learning device that learns a machine learning model that predicts wavefront information for obtaining a desired beam pattern in a photoelectric conversion unit of a light receiving device having a photoelectric conversion unit that receives a beam reflected by a deformable mirror, comprising: a data storage unit that holds the voltage values ​​of each cell in the photoelectric conversion unit and the wavefront information for obtaining the desired beam pattern as learning data; and a learning unit that trains the machine learning model so that the machine learning model receives the voltage values ​​of each cell as input and outputs the wavefront information.

5. A control method executed by a control device that controls a light receiving device having a deformable mirror and a photoelectric conversion unit, comprising: a monitoring step of monitoring the voltage value of each cell in the photoelectric conversion unit that receives a beam reflected by the deformable mirror; a prediction step of predicting wavefront information required to obtain a desired beam pattern in the photoelectric conversion unit based on the voltage value of each cell; and a mirror control step of controlling the deformable mirror using the wavefront information predicted by the prediction step.

6. A program for causing a computer to function as each part of the control device according to claim 1 or 2.

Images