Heterogeneous computing chip-level wavefront sensing-based adaptive optical control method and system

Abstract

The invention discloses an adaptive optical control method and system based on heterogeneous computing chip-level wavefront sensing, and relates to the technical field of heterogeneous computing and adaptive optical control. According to the method, an FPGA high-speed parallel calculation mode is utilized, real-time image correction can be carried out more quickly and efficiently, meanwhile, the whole process is completed on a chip, assistance of a universal computer is removed, and real-time performance is better facilitated. The MPSoC combines heterogeneous computing power of PL + PS, becomes an ideal choice of an AO system, and has the advantages that the MPSoC has high-performance parallel computing power, and FPGA accelerates wavefront reconstruction; the device has the advantages of simple structure, flexible programmability, support of C / C + + (PS) and HDL / HLS (PL), convenience in algorithm upgrading and high-efficiency and low-power calculation control, provision of real-time conditions for wavefront distortion correction and image post-processing of an on-line control system, and improvement of the imaging quality of an imaging system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of heterogeneous computing and adaptive optical control, in particular to a chip-level wavefront sensing adaptive optical control method and system based on heterogeneous computing. BACKGROUND

[0002] An adaptive optics (AO) system is a technology that can correct optical wavefront distortion in real time, and is widely used in astronomical observation, laser communication, retinal imaging, free space optical communication and military optical systems. With the increasing demand for computing, traditional DSP or FPGA solutions gradually cannot meet the real-time requirements, so the adaptive optical system based on MPSoC has become a research hotspot. A chip-level wavefront sensing adaptive optical control method based on MPSoC represents the cross-innovation of high-performance computing + real-time control + optical engineering, and has the advantages of low delay, high energy efficiency and strong scalability compared with the traditional open-loop scheme, which is promoting the revolution in the fields of astronomical observation, biomedical imaging and laser communication. Early AO systems mainly rely on DSP or FPGA for real-time computing, but for DSP, the parallel computing capability is weak, and it is difficult to process high frame rate (>1kHz) wavefront correction, and for FPGA, the development is complex, the flexibility is low, and it is difficult to adapt to dynamically changing control algorithms, which restricts the realization of improving the resolution of the imaging system by wavefront aberration correction.

[0003] MPSoC combines the heterogeneous computing capabilities of PL+PS, and becomes an ideal choice for AO systems, and its advantages include: high-performance parallel computing capability, FPGA accelerated wavefront reconstruction; flexible programmability, supporting C / C++ (PS) and HDL / HLS (PL), facilitating algorithm upgrade; and low power consumption and miniaturization.

[0004] Traditional adaptive optical systems are complex to build, have low integration, and are difficult to ensure high-speed real-time correction, which leads to a decline in image restoration quality.

[0005] Therefore, a new solution is needed to solve the above problems. SUMMARY

[0006] The present application aims to provide a chip-level wavefront sensing adaptive optical control method and system based on heterogeneous computing, which achieves more efficient real-time correction through the high-speed parallel working mode of PL and PS, and the entire process is completed on the chip, which is independent of general-purpose computer assistance and is more beneficial to real-time, to solve the technical problems proposed in the background art.

[0007] In order to achieve the above object, the present application provides the following technical scheme: a system based on a heterogeneous computing chip level wavefront sensing adaptive optical control, comprising a point light source, a deformable mirror (1), a mirror (2), a microlens array (3), an IMX990 indium gallium arsenide infrared sensor (4), and a deformable mirror drive (5);

[0008] The point light source at infinity is equivalent to parallel light, and wavefront distortion is generated in the transmission process, reaches the deformable mirror (1), and then reaches the mirror (2) after being reflected by the deformable mirror (1);

[0009] After being reflected by the mirror (2), the parallel light enters the microlens array (3) to be focused into a multifocal plane array, and then enters the IMX990 indium gallium arsenide infrared plane array sensor (4) to be imaged;

[0010] The light signal collected by the IMX990 indium gallium arsenide infrared plane array sensor (4) is input into a system on chip to be converted into an electrical signal, and finally the electrical signal is transmitted to the deformable mirror drive (5) position for wavefront correction.

[0011] Further, a control chip is further included, and an image restoration processing unit is pre-set in the control chip, and the image restoration processing unit is used for collecting and deserializing frame, row and image data signals;

[0012] Then, through serial-parallel conversion and four-channel data integration, each frame of data is subjected to algorithm threshold removal and current centroid calculation, the current slope is calculated by using a direct slope method, then the slope matrix is transmitted to a PS end to calculate a restoration matrix by using a pseudo-inverse algorithm, and finally the restoration matrix is transmitted to a PL end to calculate an error voltage;

[0013] Finally, the error voltage is transmitted back to the PS end to calculate a control voltage by using a PI control algorithm, and the control is transmitted through Ethernet to realize image restoration of the imaging system;

[0014] The PL and the PS end work cooperatively, the PL end mainly performs data deserialization, threshold removal, slope calculation and error voltage calculation, and the PS end mainly performs IIC communication, pseudo-inverse algorithm restoration matrix calculation and PI algorithm control voltage calculation, so as to complete wavefront restoration calculation and control calculation.

[0015] A heterogeneous computing chip level wavefront sensing adaptive optical control method is used in the system based on the heterogeneous computing chip level wavefront sensing adaptive optical control, and at least includes the following steps:

[0016] S1: switching the light source into a point light source, and acquiring image data of microlens array imaging through the IMX990 indium gallium arsenide infrared plane array sensor (4);

[0017] S2: the collected image data is output in a synchronization code format, and after serial-parallel conversion, the synchronization code is added before and after "virtual signal + OB signal + valid pixel data" and output, according to the IMX990 indium gallium arsenide infrared array sensor (4) synchronization code output timing, the FSM is set, and the valid pixel data is distinguished;

[0018] S3: the valid data collected by the four channels is merged and arranged, and bit splicing is integrated into each frame of data, that is, the four channels are spliced into one frame of data in the order of 1, 2, 0, and 3;

[0019] S4: after obtaining the complete data of each frame, the data is synchronized through a FIFO timing and enters the next processor for calculation, parallel light is focused on the Shack-Hartmann detector focal plane through the microlens array (3), the calculation of the sub-aperture spot centroid is performed, and finally the wavefront slope g x and g y are obtained;

[0020] S5: after obtaining the wavefront slopes g x and g y , the recovery matrix D + is derived and obtained;

[0021] S6: the recovery matrix D + calculated in S5 is transmitted back to the PL end, and the error voltage v is calculated;

[0022] S7: the calculated error voltage v is transmitted back to the PS end, and the control voltage is calculated by using a PI control algorithm;

[0023] S8: output to the deformable mirror (1) control unit through Ethernet to obtain the fast recovery of the image, and the entire wavefront sensing adaptive optical control system is integrated on the control chip to complete, and is independent of the general-purpose computer auxiliary.

[0024] Further, the formula for calculating the sub-aperture spot centroid is:

[0025]

[0026] In the formula, x i and y i are the coordinates of each sub-aperture position of the detector; I i,j is the gray value of the (i, j)th detector pixel; and L and M are the length and width of the detector, respectively.

[0027] The calculation of the spot centroid in each sub-aperture at the PL end can obtain the reference wavefront centroid (x c0 , y c0 ) and the distorted wavefront centroid (x c1 , y c1After removing a range threshold, the wavefront slope g is then calculated. x and g y .

[0028] Furthermore, the real-time calculation process of S5, also known as wavefront restoration, employs the direct slope method. The slopes of each sub-aperture are then transmitted to the PS terminal via the AXI bus to obtain a slope matrix relationship:

[0029] g = Dv(3) where v = [v1, v2, ..., v2, ..., v3, ..., v4, ..., v5, ..., v6, ..., v7, ..., v8, ..., v9, ..., v1, ..., v1, ..., v2 ... n T is the driver voltage vector, n is the number of drivers; D is the slope response matrix of the deformable mirror (1);

[0030] D + It is the pseudo-inverse of the D matrix, also known as the wavefront reconstruction matrix of the direct slope method;

[0031] When an adaptive optics system is working, the required control voltage can be obtained based on the obtained wavefront slope.

[0032] By utilizing the method of finding the pseudo-inverse of a matrix through singular value decomposition, the pseudo-inverse matrix can be obtained more quickly and accurately. This is then applied to the decomposition of the calculated slope matrix D:

[0033] D=UΣK T (4)

[0034] Where U is an m×m orthogonal matrix, and Σ is an m×n diagonal matrix with diagonal elements being singular values ​​σ1≥σ2≥...≥σ r >0, K is an n×n orthogonal matrix;

[0035] Construct the pseudo-inverse matrix Σ. + Transpose Σ and replace each non-zero singular value σi with its reciprocal 1 / σ. i Calculate the pseudo-inverse matrix D of D. + :

[0036] D + =KΣ + U T (5)

[0037] A pseudo-inverse matrix is ​​a generalization of the concept of matrix inverse, applicable to non-square or singular matrices.

[0038] Furthermore, the formula for calculating the error voltage v in S6 is as follows:

[0039] v=D + g(6)

[0040] Where g is the current real-time slope vector.

[0041] Furthermore, the formula for calculating the control voltage is as follows:

[0042]

[0043] where v(t) = r(t) - y(t) is the error signal, r(t) is the set value, and y(t) is the process variable; K p is the proportional gain, which determines the response strength to the current error; K i is the integral gain, which is used to eliminate the steady-state error (such as static wavefront aberration); dτ is the integral in the time domain.

[0044] Compared with the prior art, the present application has the beneficial effects that:

[0045] The present application utilizes the high-speed parallel computing mode of FPGA, can more quickly and efficiently perform the real-time correction image method, and the entire process is completed on the chip, so that the general-purpose computer assistance is avoided, and the real-time performance is improved. The MPSoC combines the heterogeneous computing capabilities of PL+PS, and becomes an ideal choice for the AO system, and has the following advantages: high-performance parallel computing capability, FPGA accelerated wavefront reconstruction; flexible programmability, support for C / C++ (PS) and HDL / HLS (PL), convenient algorithm upgrade, efficient and low-power computing control, which provides real-time conditions for online control system wavefront distortion correction and image post-processing, and improves the imaging quality of the imaging system. BRIEF DESCRIPTION OF DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0047] Figure 1 It is a schematic diagram of the whole system of the present application;

[0048] Figure 2 It is an architecture block diagram of the multi-core system on chip of the present application;

[0049] Figure 3 It is a signal flow block diagram of the typical adaptive optical system of the present application;

[0050] Figure 4 It is an image data deserializing synchronization code sequence diagram of the infrared sensor of the present application;

[0051] Figure 5 It is a bit splicing sequence diagram of the four-channel merging of the image data of the present application;

[0052] Figure 6 It is a structure diagram of the Shack-Hartmann wavefront sensor of the present application;

[0053] Figure 7 Workflow diagram for the multi-processor system-on-a-chip of the present application.

[0054] In the figure: 1, deformable mirror; 2, reflecting mirror; 3, microlens array; 4, IMX990 indium gallium arsenide infrared sensor; 5, deformable mirror drive. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments.

[0056] Embodiment one:

[0057] The system based on the wavefront sensing adaptive optical control of the heterogeneous computing chip includes a point light source, a deformable mirror (1), a reflecting mirror (2), a microlens array (3), an IMX990 indium gallium arsenide infrared sensor (4), and a deformable mirror drive (5).

[0058] The point light source at infinity is equivalent to parallel light, and wavefront distortion is generated in the transmission process, reaches the deformable mirror (1), and then reaches the reflecting mirror (2) after being reflected by the deformable mirror (1).

[0059] After being reflected by the reflecting mirror (2), the parallel light enters the microlens array (3) to be focused into a multi-focal plane array, and then enters the IMX990 indium gallium arsenide infrared plane array sensor (4) to be imaged.

[0060] The light signal collected by the IMX990 indium gallium arsenide infrared plane array sensor (4) is input to the system-on-a-chip to be converted into an electrical signal, and finally the electrical signal is transmitted to the deformable mirror drive (5) position for wavefront correction.

[0061] It also includes a control chip, and an image restoration processing unit is pre-set in the control chip, which is used for collecting and deserializing frame, line, and image data signals.

[0062] Then, through serial-parallel conversion and four-channel data integration, each frame of data is subjected to algorithm thresholding and current centroid calculation, the current slope is calculated by using the direct slope method, the slope matrix is then transmitted to the PS end to calculate the restoration matrix by using the pseudo-inverse algorithm, and the restoration matrix is then transmitted to the PL end to calculate the error voltage.

[0063] Finally, the error voltage is transmitted back to the PS end to calculate the control voltage by using the PI control algorithm, and the control is transmitted through Ethernet to realize image restoration of the imaging system.

[0064] PL and PS work together, PL mainly carries out data deserialization, centroid calculation, threshold removal, slope calculation and error voltage calculation, while PS mainly carries out IIC communication, pseudo-inverse algorithm calculation and PI algorithm calculation to complete wavefront recovery calculation and control calculation.

[0065] The whole adaptive optical system process is shown in Figure 3 .

[0066] A wavefront detector (WFS) measures wavefront distortion, and a wavefront recovery calculator (WFC) and a control calculator (CC) are used in a system on chip to obtain a control voltage signal, which is converted into an analog signal by a digital-to-analog converter (DAC) and a high-voltage amplifier (HVA), so that a deformable mirror (DM) generates a required compensation wavefront. The whole adaptive optical system is a digital-analog hybrid control system. The wavefront control operation is to obtain a driver control voltage by a control algorithm from the recovered error voltage.

[0067] Embodiment two

[0068] Referring to Figures 1-7 , a wavefront sensing adaptive optical control method based on a heterogeneous computing chip, for the system based on the wavefront sensing adaptive optical control of the heterogeneous computing chip described in the above embodiment, at least comprising the following steps:

[0069] S1: Switch the light source to a point light source, and acquire image data of micro-lens array imaging by an IMX990 indium gallium arsenide infrared area array sensor (4);

[0070] S2: Output the collected image data in a synchronization code format, add and output a synchronization code before and after "virtual signal + OB signal + valid pixel data" after serial-parallel conversion, and set an FSM according to the synchronization code output timing of the IMX990 indium gallium arsenide infrared area array sensor (4) to distinguish valid pixel data; referring to Figure 4 , the frame header and frame tail synchronization code of the valid / invalid image data of SENSOR deserialization in the figure represent that four channels simultaneously collect synchronization codes of three different bit depths of 8 / 10 / 12 bits, and the sensor output signal is output by an internal circuit with a delay time relative to the horizontal synchronization signal;

[0071] S3: Merge and arrange the valid data collected by the four channels, and splice and integrate each frame of data by bits, i.e. four channels are spliced into one frame of data in the order of 1, 2, 0, and 3 (see Figure 5 );

[0072] S4: After getting the complete data of each frame, the data is sent to the next processor through a FIFO time sequence synchronization. Parallel light is focused on the focal plane of the Shack-Hartmann detector through the microlens array (3) to calculate the centroid of the sub-aperture light spot and finally to obtain the wavefront slope g x and g y ;

[0073] S5: After obtaining the wavefront slope g x and g y , the recovery matrix D + is derived

[0074] S6: The recovery matrix D + calculated in S5 is sent back to the PL end to calculate the error voltage v

[0075] S7: The calculated error voltage v is sent back to the PS end to calculate the control voltage by using the PI control algorithm

[0076] S8: The output is sent to the deformable mirror (1) control unit through the Ethernet to obtain the fast recovery of the image. The entire wavefront sensing adaptive optical control system is integrated on the control chip to complete, which is independent of the general-purpose computer auxiliary, as shown in Figure 2 .

[0077] Referring to Figure 6 , the formula for calculating the centroid of the sub-aperture light spot is:

[0078]

[0079] In the formula, x i and y i are the coordinates of the position of each sub-aperture of the detector; I i,j is the gray value of the (i, j) detector pixel; and L and M are the length and width of the detector, respectively.

[0080] The centroid of the reference wavefront (x c0 , y c0 ) and the centroid of the distorted wavefront (x c1 , y c1 ) can be obtained by calculating the centroid of the light spot in each sub-aperture of the PL end. After removing a range threshold, the wavefront slope g x and g y are calculated.

[0081] The real-time calculation process of S5 is also called the recovered wavefront. The direct slope method is adopted. The slope of each sub-aperture is transmitted to the PS end through the AXI bus to obtain a slope matrix relationship:

[0082] g = Dv (3) where v = [v1, v2, …, v nT is the driver voltage vector, n is the number of drivers; D is the slope response matrix of the deformable mirror (1);

[0083] D + is the pseudo-inverse matrix of D matrix, also called direct slope method wavefront recovery matrix;

[0084] In the adaptive optics system, according to the obtained wavefront slope, the required control voltage can be obtained;

[0085] Using the singular value decomposition method to calculate the pseudo-inverse matrix, the pseudo-inverse matrix is calculated more quickly and accurately, and the slope matrix D is decomposed:

[0086] D = UΣK T (4)

[0087] Where U is an m x m orthogonal matrix, Σ is an m x n diagonal matrix, the diagonal elements are singular values σ1≥σ2≥...≥σ r >0, K is an n x n orthogonal matrix;

[0088] The pseudo-inverse matrix of Σ is constructed + , transpose Σ and replace each non-zero singular value σi with its reciprocal 1 / σi, and calculate the pseudo-inverse matrix D + of D:

[0089] D + = KΣ + U T (5)

[0090] The pseudo-inverse matrix is a generalization of the inverse concept of matrix, which is suitable for non-square matrix or singular matrix. Because the slope matrix is a non-square matrix, the pseudo-inverse matrix can be calculated effectively by using singular value decomposition.

[0091] The formula for calculating the error voltage v is as follows:

[0092] v = D + g (6)

[0093] Where g is the current real-time slope vector.

[0094] The formula for calculating the control voltage is as follows:

[0095]

[0096] Where v(t) = r(t) - y(t) is the error signal, r(t) is the set value, and y(t) is the process variable; K p is the proportional gain, which determines the response strength to the current error; K i is the integral gain, which is used to eliminate steady-state error (such as static wavefront aberration); dτ is the integral in the time domain.

[0097] It will be apparent to those skilled in the art that the application is not limited to the details of the above-exemplified embodiments and that the present application can be implemented in other particular forms without departing from the spirit or essential characteristics of the present application. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. No reference signs in the claims shall be construed as limiting the scope of the claims.

Claims

1. A system for adaptive optics control based on heterogeneous computing chip-level wavefront sensing, characterized in that: It includes a point light source, a deformable mirror (1), a reflector (2), a microlens array (3), an IMX990 indium gallium arsenide infrared sensor (4), and a deformable mirror driver (5); The point light source at infinity is equivalent to parallel light. During the transmission process, wavefront distortion occurs, and the light reaches the deformable mirror (1). After being reflected by the deformable mirror (1), the light reaches the reflecting mirror (2). After being reflected by the mirror (2), the parallel light enters the microlens array (3) and is focused into a multifocal array, and then enters the IMX990 indium gallium arsenide infrared array sensor (4) for imaging; The optical signal collected by the IMX990 indium gallium arsenide infrared array sensor (4) is input into the on-chip system and converted into an electrical signal. Finally, the electrical signal is transmitted to the position of the deformable mirror driver (5) for wavefront correction.

2. The adaptive optics control system based on heterogeneous computing chip-level wavefront sensing according to claim 1, characterized in that: It also includes a control chip, which has a preset image restoration processing unit for acquiring and deserializing frame, line and image data signals; Then, through serial-to-parallel conversion and four-channel data integration, the data of each frame is subjected to threshold removal and current centroid calculation using an algorithm. The current slope is calculated using the direct slope method. The slope matrix is ​​then passed to the PS end and the pseudo-inverse algorithm is used to calculate the restoration matrix. The restoration matrix is ​​then passed to the PL end and the algorithm calculates the error voltage. Finally, the error voltage is transmitted back to the PS terminal, and the PI control algorithm is used to calculate the control voltage. The control is then transmitted via Ethernet to realize image restoration of the imaging system. The PL and PS terminals work together. The PL terminal mainly performs data deserialization, calculates the centroid and removes the threshold, calculates the slope and calculates the error voltage, while the PS terminal mainly performs IIC communication, calculates the restoration matrix using the pseudo-inverse algorithm and calculates the control voltage using the PI algorithm, thus completing the wavefront restoration calculation and control calculation.

3. A heterogeneous computing chip-level wavefront sensing adaptive optics control method, used in the system of heterogeneous computing chip-level wavefront sensing adaptive optics control as described in any one of claims 1-2, characterized in that: At least the following steps are included: S1: Switch the light source to a point light source and acquire the image data of the microlens array imaging through the IMX990 indium gallium arsenide infrared array sensor (4); S2: Output the acquired image data in the synchronization code format. After serial-to-parallel conversion, the synchronization code is added and output immediately before and after "virtual signal + OB signal + effective pixel data". According to the synchronization code output timing of the IMX990 indium gallium arsenide infrared array sensor (4), set the FSM and identify the effective pixel data. S3: Merge and organize the valid data collected simultaneously from the four channels, and then stitch them together to form each frame of data. That is, the four channels are stitched together in the order of 1, 2, 0, 3 to form a frame of data. S4: After obtaining complete data for each frame, the data is synchronized through a FIFO and then processed by the next processor. Parallel light is focused onto the focal plane of the Shaker-Hartmann detector through a microlens array (3) to calculate the centroid of the light spot within the sub-aperture, and finally the wavefront slope g is obtained. x and g y ; S5: After obtaining the wavefront slope g x With g y Then, the restored matrix D is derived. + ; S6: Restore the matrix D calculated in S5 + The error voltage V is calculated by sending the data back to the PL terminal. S7: Transmit the calculated error voltage v back to the PS terminal, and use the PI control algorithm to calculate the control voltage; S8: By outputting the image to the deformable mirror (1) control unit via Ethernet, the image can be quickly restored. The entire wavefront sensing adaptive optics control system is integrated on the control chip, eliminating the need for general computer assistance.

4. The adaptive optics control method based on heterogeneous computing chip-level wavefront sensing according to claim 3, characterized in that: The formula for calculating the centroid of the light spot within the sub-aperture is as follows: In the formula, x i and y i These are the coordinates of the positions of each sub-aperture of the detector; I i，j Let be the gray value of the (i, j)th detector pixel; L and M are the length and width of the detector, respectively. By calculating the centroid of the light spot within each sub-aperture at the PL end, the reference wavefront centroid (x) can be obtained. c0 y c0 ) and the centroid of the distorted wavefront (x c1 y c1 After removing a range threshold, the wavefront slope g is then calculated. x and g y .

5. The adaptive optics control method based on heterogeneous computing chip-level wavefront sensing according to claim 4, characterized in that: The real-time calculation process of S5, also known as wavefront restoration, employs the direct slope method. Furthermore, the slopes of each sub-aperture are transmitted to the PS terminal via the AXI bus to obtain a slope matrix relationship: g = Dv(3) Where, v = [v1, v2, ..., v] n T is the driver voltage vector, n is the number of drivers; D is the slope response matrix of the deformable mirror (1); D + It is the pseudo-inverse matrix of the D matrix, also known as the wavefront reconstruction matrix of the direct slope method; When an adaptive optics system is working, the required control voltage can be obtained based on the obtained wavefront slope. By utilizing the method of finding the pseudo-inverse of a matrix through singular value decomposition, the pseudo-inverse matrix can be obtained more quickly and accurately. This is then applied to the decomposition of the calculated slope matrix D: D=UΣK T (4) Where U is an m×m orthogonal matrix, Σ is an m×n diagonal matrix, and the diagonal elements are singular values ​​σ1≥σ2≥...≥σ r >0, K is an n×n orthogonal matrix; Construct the pseudo-inverse matrix Σ. + Transpose Σ and assign each non-zero singular value σ i Replace it with its reciprocal 1 / σ i Calculate the pseudo-inverse matrix D of D. + : D + =KΣ + U T (5) A pseudo-inverse matrix is ​​a generalization of the concept of matrix inverse and is applicable to non-square or singular matrices.

6. The adaptive optics control method based on heterogeneous computing chip-level wavefront sensing according to claim 5, characterized in that: The formula for calculating the error voltage v in S6 is as follows: v=D + g(6) Where g is the current real-time slope vector.

7. The adaptive optics control method based on heterogeneous computing chip-level wavefront sensing according to claim 6, characterized in that: The formula for calculating the control voltage is as follows: Where v(t) = r(t) - y(t) is the error signal, r(t) is the setpoint, and y(t) is the process variable; K p It is the proportional gain, which determines the strength of the response to the current error; K i dτ is the integral gain, used to eliminate steady-state error; dτ is the integral in the time domain.

Images