Layered grating collimation system based on automatic calibration

Abstract

The invention discloses a layered grating collimation system based on automatic calibration, which combines transmission-reflection grating collaborative error equalization, self-adaptive optical path phase adjustment and multivariable weight optimization control to realize high-precision collimation and dynamic error correction of light beams. Through a layered grating error equalization regulation and control mechanism, cooperative compensation of transmission and reflection beam errors is realized, collimation precision reduction caused by error accumulation is avoided, the propagation stability of the beams in a complex medium is enhanced, and based on an optical path non-uniform phase dynamic compensation strategy, the optical path non-uniform phase dynamic compensation is realized. The optical path nonlinear deviation can be effectively corrected, the propagation precision of long-distance light beams is improved, the light beam drift prediction capability is improved, dynamic control optimization is achieved, a multivariable adaptive weight optimization mechanism is adopted, short-period error correction and long-period trend correction strategies are combined, it is ensured that the light beams keep high stability under different environment interferences, and the dynamic control optimization is achieved. And the automatic calibration capability of the system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of measurement, and particularly to a hierarchical grating collimation system based on automatic calibration. Background Art

[0002] The field of optical collimation and calibration technology includes various optical systems and methods for beam alignment and direction adjustment, which are mainly applied to fields such as precision measurement, optical detection, optical communication, and laser processing. The core content of this technical field includes the arrangement of optical elements, the optimization of the beam path, the detection and correction of alignment errors, and the application of an automatic calibration mechanism. Among them, a hierarchical grating collimation system based on automatic calibration refers to a system that uses a multi-level optical grating structure combined with an automatic calibration mechanism to achieve precise beam collimation. For the detection and compensation of beam alignment errors, this system adopts a hierarchical grating design, and the beam is collimated in stages through multiple optical gratings to reduce the cumulative error of beam offset. At the same time, the automatic calibration mechanism is based on a photodetector and a feedback control unit to achieve real-time monitoring of the beam offset, and adjusts the position of the optical element through an electric adjustment component to ensure the accurate transmission of the beam. The beam collimation technology plays a key role in high-precision optical measurement, long-distance communication, lithography, and laser processing.

[0003] Existing beam collimation systems mainly rely on a single transmissive or reflective grating, and error correction is performed through mechanical adjustment, phase compensation, or feedback control. However, these methods have significant limitations. For example, the mechanical adjustment method has a slow response speed and is greatly affected by environmental factors; a single phase compensation strategy is difficult to adapt to dynamic error changes; the traditional feedback control scheme is easily affected by noise interference, resulting in an error accumulation effect. Therefore, there is still a large room for optimization in the beam collimation accuracy, error control ability, and dynamic response performance of the existing technology. To address this problem, this paper proposes a hierarchical grating collimation system based on automatic calibration, which realizes the beam collimation ability with high precision and low error accumulation through a multi-level error compensation mechanism, an adaptive control algorithm based on error parameters, and a dynamic adjustment strategy of the grating structure. Summary of the Invention

[0004] The technical problem solved by the present invention is that existing beam collimation systems mainly rely on a single transmissive or reflective grating, and error correction is performed through mechanical adjustment, phase compensation, or feedback control. However, these methods have significant limitations and cannot dynamically compensate for the error accumulation caused by non-uniform phase distortion of the optical path and environmental disturbances. A single-layer grating is difficult to simultaneously suppress the polarization / angle deviation of the transmission and reflection paths, and the existing control algorithms have insufficient prediction accuracy for multi-variable coupling drift, resulting in poor long-term collimation stability.

[0005] To solve the above technical problems, the present invention provides the following technical solutions: A hierarchical grating collimation system based on automatic calibration, comprising a hierarchical grating error equalization control module, an optical path non-uniform phase dynamic compensation module, and a beam drift prediction control module with adaptive weight adjustment: The hierarchical grating error equalization control module is used to perform dynamic error equalization compensation on beams with different incident angles and polarization states through a cooperative error compensation strategy of a transmission grating TG and a reflection grating RG; The optical path non-uniform phase dynamic compensation module is used to generate a compensation map, characterize the phase error of the beam in different propagation regions, calculate the optimal phase adjustment compensation amount, and optimize the compensation weight coefficient; The beam drift prediction control module with adaptive weight adjustment is used to establish a beam drift model, perform predictive training of a support vector machine on the offset trend graph based on an objective function to obtain a beam drift model, calculate the future drift path and offset correction amount according to the optimized weight, and adopt a control strategy combining short-period error correction and long-period trend correction.

[0006] Preferably, at the grating interface, independent sampling is respectively performed on the incident light angle deviation, wavefront aberration, and energy distribution non-uniformity, the phase change gradient of the transmitted beam and the direction offset trend of the reflected beam are extracted, and the transmission error and the reflection error are calculated; When the beam is incident on the grating surface, the angle deviation between the incident beam and the ideal incident angle is measured by an optical sensor, and the angle deviation is the difference between the angle of the beam and the preset ideal collimation direction; Multiple incident light angle data are collected at different positions and different time points, and the angle deviation between each sampling point and the ideal incident direction is calculated based on the incident light angle data; Wavefront sampling is performed at the grating interface by a wavefront sensor, and the beam wavefront shape is measured. The beam wavefront shape includes distortions caused by medium non-uniformity, phase changes of the beam itself, or system errors. The phase values of each sampling point are obtained through wavefront sampling, and the change of the beam wavefront shape over time is recorded. The phase change gradient of the beam wavefront shape is calculated based on the beam wavefront shape and the phase values of each point. The phase change of the beam wavefront shape is obtained by calculating the phase difference between adjacent sampling points.

[0007] Preferably, the key phase change features of the beam wavefront shape are extracted. The key phase change features include wavefront bending, distortion amplitude, and phase change gradient. The energy distribution data of the beam is collected at the grating interface by a beam profile analyzer. The energy distribution data is a three-dimensional energy distribution map. The light intensity difference between the light intensity value in the energy distribution data and the ideal uniform light intensity is calculated, and the non-uniformity of the energy distribution is evaluated according to the light intensity difference. The evaluation logic includes: Calculate the standard deviation of the intensity difference corresponding to all phase points, and preset a standard deviation threshold. When the standard deviation of the intensity difference of the phase points corresponding to the energy distribution points is greater than the standard deviation threshold, the energy distribution is non-uniform; Use a precise direction sensor to sample in the direction of the reflected beam, measure the direction offset of the reflected beam, record the cause label of the direction offset, where the cause label includes the microstructure change of the grating, surface unevenness, and external disturbance. Record the direction change of the reflected beam in time series, calculate the angle offset trend between the reflected beam and the ideal reflection direction, and obtain the reflection error after weighted average calculation of the angle offset trend and the phase change of the beam wavefront morphology. Add the phase error and the intensity error after weighting to obtain the transmission error.

[0008] Preferably, according to the transmission error and the reflection error, dynamically adjust the microstructure and period parameters of the compensation grating, and dynamically adjust the period gradient of the transmission grating and the microstructure inclination angle of the reflection grating through a local period adjustment algorithm based on gradient optimization to compensate and control the error. The compensation control logic includes: Dynamically calculate the local error weights of the transmission error and the reflection error. The mathematical expressions for the local error weights of the transmission error and the reflection error are: ; ; where, is the angle deviation, is the maximum value of the angle deviation, is the phase change amount, is the theoretical maximum value of the phase change amount, is the polarization offset amount, is the theoretical maximum value of the polarization offset amount, is the wavelength drift, is the theoretical maximum value of the wavelength drift, is the local error weight of the transmission error, is the local error weight of the reflection error, 、 、 and are weight coefficients; Preferably, the optical path non-uniform phase dynamic compensation module includes an optical path deviation measurement unit, a dynamic compensation mapping generation unit, and an adaptive phase correction unit: The optical path deviation measurement unit is used to extract the optical path deviation characteristics, independently measure the medium refractive index gradient, the beam path perturbation offset, and the optical path cumulative error during the beam propagation process according to the optical path deviation characteristics, and calculate the influence degree of the optical path of each region on the phase distortion; Extracting the optical path deviation feature includes performing phase error correction after constructing the optical path deviation matrix, which is obtained by arranging the optical path deviations of all sampling points. The mathematical expression for calculating the optical path deviation is: ; ; Among them, is the total length of the transmission medium, is the ideal refractive index of the medium, is the optical path deviation, is the phase value of the sampling point, is the i-th equally divided medium segment, is the refractive index of the i-th equally divided medium, is the length of the i-th equally divided medium, is the ideal optical path; When in the visible light wavelength range, decompose the diffraction wavelength of the calculated optical path deviation to obtain diffraction waves of different colors in a local area. Record the components and intensities of the enhanced color light corresponding to each observation angle, extract the diffraction wave in the superposition state and the color corresponding to the superposition state diffraction wave, use big data to collect the color superposition database, and respectively query the superposition color of the diffraction wave in the superposition state and the number of superposed diffraction waves corresponding to each observation angle according to the collected color superposition database; By performing optical delay line movement on one of the decomposed diffraction waves of the superposed diffraction wave, introducing optical path perturbation detection, performing real-time compensation on the optical delay line of the other diffraction wave according to the optical path deviation value, and calculating the current optical path deviation twice, calculate the absolute value of the difference between the two optical path deviations as the target residual.

[0009] Preferably, the dynamic compensation mapping generation unit is used to generate a compensation mapping, characterize the phase error of the light beam in different propagation regions, and calculate the optimal phase adjustment compensation amount. The mathematical expression for loading the reverse phase distribution in the optical path compensator is: ; Among them, is the reverse phase, is the wave number, is the target residual.

[0010] Preferably, the adaptive phase correction unit includes: Calculate the time derivative of the optical path deviation and the phase change to obtain the optical path change rate and the phase change rate per unit time. Preset the change rate threshold. When the optical path change rate and the phase change rate reach the change rate threshold, automatically adjust the compensation weight coefficient based on the mathematical expression of the local error weight. The optimization objective function for optimizing the compensation weight coefficient is: ; Among them, is the reversed phase after current compensation, is the optimization objective function; By minimizing the value of the optimization objective function, the optimal compensation weight coefficient is obtained. Based on the feedback phase change data and optical path deviation, the phase of the optical path compensator is adjusted according to the optimal compensation weight coefficient calculated by the optimization algorithm.

[0011] Preferably, the beam drift prediction control module with adaptive weight adjustment includes a drift state quantity dynamic acquisition unit, a prediction weight optimization unit, and a real-time control adjustment unit: The drift state quantity dynamic acquisition unit is used to collect key parameters such as the offset of the beam center position, the change rate of the beam divergence angle, and the drift acceleration in real time through a multi-point monitoring mechanism, compare the offset trends of the beam main axes at different positions, and establish a beam drift model; The state information of the beam is collected through a CCD array. The state information includes the coordinate change of the beam center and the beam divergence angle. The offset is calculated according to the coordinate change of the beam center. The mathematical expression for calculating the offset is: ; ; Among them, and are the offsets of the coordinate change at time t, and are the measured beam center positions, and are the ideal reference positions of the beam; The unit rate of the divergence angle is calculated by taking the derivative. The drift acceleration is obtained by taking the second derivative of the offset. Through the multi-point monitoring mechanism, the offset of the beam center position, the change rate of the divergence angle, and the drift acceleration are calculated and collected at different array nodes of the CCD array. Based on the two-dimensional position coordinate axes of the beam measured by the CCD array, an offset trend graph is drawn. The horizontal axis coordinate of the offset trend graph is the position parameter, and the vertical axis coordinate is the offset of the beam center position, the change rate of the divergence angle, and the drift acceleration; Establishing the beam drift model includes: The mathematical expression of the objective function of the beam drift model is: ; Among them, is the offset vector of the beam, A is the coefficient of the beam velocity, which is used to describe the linear velocity of the beam drift, B is the coefficient of the beam acceleration, which is used to describe the acceleration of the beam drift, and C is the constant term, which is used to represent the initial position; Perform support vector machine prediction training on the offset trend graph based on the objective function to obtain a beam drift model, and output the prediction of the beam drift behavior.

[0012] Preferably, the prediction weight optimization unit is used to calculate the offset increment, profile change rate, and drift stability coefficient of the beam within different time windows through a multivariate offset trend analysis algorithm. The offset increment is the displacement change amount of the beam within different time windows, the profile change rate is the change speed of the beam shape, and the profile change rate is obtained by measuring the difference in the spot width within different time windows and taking the derivative. The drift stability coefficient is used to measure the stability of the beam drift and is obtained by calculating the variance of the offset increment. Collect historical drift states through big data, perform weighted regression calculation on all historical drift states to obtain the influence of historical data on the current drift trend. The prediction weight in the weighted regression calculation is the influence degree of each historical drift state data point on the current corresponding drift state, and the influence degree is obtained by calculating the probability of the t-distribution based on the time axis.

[0013] After the prediction weight is updated, it is used for real-time drift compensation of the beam, and the optical path compensator and laser source positioning device are adjusted according to the current drift prediction weight.

[0014] Preferably, the real-time control adjustment unit is used to calculate the future drift path and offset correction amount according to the optimized weight, and adopt a control strategy combining short-period error correction and long-period trend correction. The control strategy logic includes: The short-period correction amount is used to immediately adjust the beam offset direction, and the long-period correction amount is used to adjust the overall collimation direction and stability parameters of the system. When the sensor fails, switch to the beam path prediction mode based on historical data and enable the standby grid layer. When the main channel fails, switch to the redundant channel through the polarization beam splitter.

[0015] The beneficial effects of the present invention: improve the beam collimation accuracy, reduce the error accumulation effect, realize the collaborative compensation of the transmission and reflection beam errors through the hierarchical grid error equalization control mechanism, avoid the decline of the collimation accuracy caused by error accumulation, enhance the propagation stability of the beam in complex media, based on the optical path non-uniform phase dynamic compensation strategy, can effectively correct the optical path nonlinear deviation, improve the propagation accuracy of the long-distance beam, improve the beam drift prediction ability, realize dynamic control optimization, adopt a multivariate adaptive weight optimization mechanism, combine short-period error correction and long-period trend correction strategies, ensure the beam maintains high stability under different environmental interferences, improve the system's automatic calibration ability, reduce the need for manual intervention, realize adaptive error compensation and dynamic adjustment through intelligent optimization algorithms, improve the system's self-adaptation ability, and reduce the complexity of traditional optical systems relying on manual adjustment. Description of the Drawings

[0016] Figure 1 This is a schematic diagram of the basic process of a hierarchical grating collimation system based on automatic calibration provided by an embodiment of the present invention. Detailed Embodiments

[0017] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific embodiments of the present invention will be given in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments.

[0018] Referring to Figure 1 , an embodiment of the present invention provides a hierarchical grating collimation system based on automatic calibration, including a hierarchical grating error equalization and regulation module, an optical path non-uniform phase dynamic compensation module, and a beam drift prediction and control module with adaptive weight adjustment: The hierarchical grating error equalization and regulation module is used to perform dynamic error equalization compensation on beams with different incident angles and polarization states through the cooperative error compensation strategy of the transmission grating TG and the reflection grating RG to reduce the error accumulation effect; The optical path non-uniform phase dynamic compensation module is used to generate a compensation map, characterize the phase error of the beam in different propagation regions, calculate the optimal phase adjustment compensation amount, and optimize the compensation weight coefficient; The beam drift prediction and control module with adaptive weight adjustment is used to establish a beam drift model, perform prediction training of the support vector machine on the offset trend graph based on the objective function to obtain the beam drift model, calculate the future drift path and offset correction amount according to the optimized weight, and adopt a control strategy combining short-period error correction and long-period trend correction.

[0019] At the grating interface, independent sampling is respectively performed on the incident light angle deviation, wavefront distortion, and energy distribution non-uniformity, the phase change gradient of the transmitted beam and the direction offset trend of the reflected beam are extracted, and the transmission error and reflection error are calculated according to the existing physical formulas; When the beam is incident on the grating surface, the angle deviation between the incident beam and the ideal incident angle is measured by an optical sensor, and the angle deviation is the difference between the angle of the beam and the preset ideal collimation direction; Multiple incident light angle data are collected at different positions and different time points to detect possible changes and non-uniformities, and the angle deviation between each sampling point and the ideal incident direction is calculated based on the incident light angle data; Wavefront sampling is carried out at the grating interface using a wavefront sensor, and the wavefront shape of the light beam is measured. The wavefront shape of the light beam includes distortions caused by medium inhomogeneity, phase changes of the light beam itself, or systematic errors. Phase values of each sampling point are obtained through wavefront sampling, and the change of the wavefront shape of the light beam over time series is recorded. The phase change gradient of the wavefront shape of the light beam is calculated based on the wavefront shape of the light beam and the phase values of each point. The phase change of the wavefront shape of the light beam is obtained by calculating the phase difference between adjacent sampling points.

[0020] Key features of the phase change of the wavefront shape of the light beam are extracted. The key features of the phase change include wavefront bending, distortion amplitude, and phase change gradient. Energy distribution data of the light beam is collected at the grating interface using a beam profile analyzer. The energy distribution data is a three-dimensional energy distribution map, which reflects the change of the light beam intensity with spatial position. The intensity difference between the measured light intensity value in the energy distribution data and the ideal uniform light intensity is calculated, and the non-uniformity of the energy distribution is evaluated based on the intensity difference. The evaluation logic includes: The standard deviation of the intensity differences corresponding to all phase points is calculated, and a standard deviation threshold is preset. When the standard deviation of the intensity differences of the phase points corresponding to the energy distribution points is greater than the standard deviation threshold, the energy distribution is non-uniform. Sampling is carried out in the direction of the reflected light beam using a precise direction sensor to measure the direction offset of the reflected light beam, and the cause labels of the direction offset are recorded. The cause labels include microstructural changes of the grating, surface unevenness, and external disturbances. The direction change of the reflected light beam is recorded over time series. The angular offset trend between the reflected light beam and the ideal reflection direction is calculated, and the reflection error is obtained by weighted average calculation of the angular offset trend and the phase change of the wavefront shape of the light beam. The transmission error is obtained by weighted summation of the phase error and the intensity error.

[0021] Based on the transmission error and the reflection error, the microstructure and period parameters of the compensation grating are dynamically adjusted. The period gradient of the transmission grating and the microstructure tilt angle of the reflection grating are dynamically adjusted through a local period adjustment algorithm based on gradient optimization to compensate and control the error. The compensation and control logic includes: The local error weights of the transmission error and the reflection error are dynamically calculated. The mathematical expressions of the local error weights of the transmission error and the reflection error are: ; ; where, is the angular deviation, is the maximum value of the angular deviation, is the phase change amount, is the theoretical maximum value of the phase change amount, is the polarization offset amount, is the theoretical maximum value of the polarization offset, is the wavelength drift, is the theoretical maximum value of the wavelength drift, is the local error weight of the transmission error, is the local error weight of the reflection error, 、 、 and are the weight coefficients; The transmission grating TG and the reflection grating RG adopt an asymmetric periodic design. The period gradient of TG is 10 - 100 μm / cm, and the adjustment range of the microstructure inclination angle of RG is ±5°; Optimize the overall error compensation amplitude of the transmitted and reflected beams according to the change rate of the beam incident direction, so that the beam error is maintained within the optimal stable threshold, and avoid the decrease of the collimation accuracy caused by long-term error accumulation.

[0022] Adjust the period of the transmission grating TG and the inclination angle of the reflection grating RG by the conjugate gradient method, so that the total error of the transmission error and the reflection error is less than or equal to 0.1λ, where λ is the working wavelength.

[0023] The optical path non-uniform phase dynamic compensation module includes an optical path deviation measurement unit, a dynamic compensation mapping generation unit, and an adaptive phase correction unit: The optical path deviation measurement unit is used to extract the optical path deviation characteristics, and independently measure the refractive index gradient of the medium, the perturbation offset of the beam path, and the optical path cumulative error during the beam propagation process according to the optical path deviation characteristics, and calculate the influence degree of the optical path of each region on the phase distortion; Extracting the optical path deviation characteristics includes performing phase error correction after constructing the optical path deviation matrix. The optical path deviation matrix is obtained by arranging the optical path deviations of all sampling points. The mathematical expression for calculating the optical path deviation is: ; ; where, is the total length of the transmission medium, is the ideal refractive index of the medium, is the optical path deviation, is the phase value of the sampling point, is the i-th evenly divided medium segment, is the refractive index of the i-th evenly divided medium, is the length of the i-th evenly divided medium, is the ideal optical path; When in the visible light wavelength range, decompose the diffraction wavelength of the calculated optical path deviation to obtain diffraction waves of different colors in a local area, record the components and intensities of the enhanced color light corresponding to each observation angle, where the components and intensities are represented by existing component and intensity levels, extract the diffraction waves in the superposition state and the colors corresponding to the diffraction waves in the superposition state, collect a color superposition database using big data, and respectively query the superposition colors of the diffraction waves in the superposition state and the number of superposed diffraction waves corresponding to each observation angle according to the collected color superposition database; Since the constructed phase difference field has only one degree of freedom, the phase gradient correspondingly is also a set of parallel straight lines with vertical scratches. The direction change of the fringe color follows the direction where the light-transmitting area is blocked, that is, the direction perpendicular to the scratches. Therefore, the extension direction of the colored line is actually the direction of a gradient line of the optical path difference field.

[0024] By moving an optical delay line for one of the decomposed diffraction waves of the superposed diffraction wave, introduce optical path perturbation detection, perform real-time compensation on the optical delay line of another diffraction wave according to the optical path deviation value, calculate the current optical path deviation twice, and calculate the absolute value of the difference between the two optical path deviations as the target residual.

[0025] The dynamic compensation mapping generation unit is used to generate a compensation mapping, characterize the phase error of the light beam in different propagation regions, and calculate the optimal phase adjustment compensation amount to maintain the stability of the phase correction in the light beam propagation direction. The mathematical expression for loading the reverse phase distribution in the optical path compensator is: ; Among them, is the reverse phase, is the wave number, is the target residual.

[0026] In order to calculate the optimal phase adjustment compensation amount, the purpose is to eliminate or compensate the phase error in propagation, so that the propagation direction of the light beam is corrected and maintained stable. Load this reverse phase distribution in the optical path compensator to achieve stable phase correction.

[0027] By adjusting the phase distribution of the optical path compensator in real time, the phase distortion of the light beam can be effectively controlled, and the stable propagation of the light beam can be maintained.

[0028] The adaptive phase correction unit includes: Calculate the time derivatives of the optical path deviation and the phase change to obtain the optical path change rate and the phase change rate per unit time. Preset a change rate threshold. When the optical path change rate and the phase change rate reach the change rate threshold, automatically adjust the compensation weight coefficient based on the mathematical expression of the local error weight to ensure the coherence of the light beam. The optimization objective function for optimizing the compensation weight coefficient is: ; wherein, is the compensated reverse phase at present, is the optimization objective function; By minimizing the value of the optimization objective function, the optimal compensation weight coefficient is obtained. Based on the feedback phase change data and optical path deviation, the phase of the optical path compensator is adjusted according to the optimal compensation weight coefficient calculated by the optimization algorithm to ensure that the light beam is corrected in time during propagation.

[0029] Through the above steps, we can dynamically adjust the optical path compensation weight in a complex propagation medium, and optimize the phase adjustment coefficient according to the real-time changing phase error and optical path change rate. Through adaptive optimization, the coherence of the light beam during propagation is ensured to be stable, and the robustness and accuracy of the hierarchical grating collimation system based on automatic calibration are improved.

[0030] The beam drift prediction control module with adaptive weight adjustment includes a drift state quantity dynamic acquisition unit, a prediction weight optimization unit, and a real-time control adjustment unit: The drift state quantity dynamic acquisition unit is used to establish a beam drift model by collecting key parameters such as the offset of the light beam center position, the change rate of the light beam divergence angle, and the drift acceleration in real time through a multi-point monitoring mechanism, and comparing the offset trends of the main axes of the light beam at different positions; The state information of the light beam is collected through a CCD array. The state information includes the coordinate change of the light beam center and the light beam divergence angle. The offset is calculated according to the coordinate change of the light beam center. The mathematical expression for calculating the offset is: ; ; wherein, and are the offsets of the coordinate change at time t, and are the measured light beam center positions, and are the ideal reference positions of the light beam; The unit rate of the divergence angle is calculated by taking the derivative, the drift acceleration is obtained by taking the second derivative of the offset, and through the multi-point monitoring mechanism, the offset of the light beam center position, the change rate of the divergence angle, and the drift acceleration are calculated and collected at different array nodes of the CCD array. Based on the two-dimensional position coordinate axis of the light beam measured by the CCD array, an offset trend graph is drawn. The horizontal axis coordinate of the offset trend graph is the position parameter, and the vertical axis coordinates are the offset of the light beam center position, the change rate of the divergence angle, and the drift acceleration; For example, by measuring the position offsets of the light beam on a series of sensors and combining with the changes in its divergence angle, the stability of the light beam during propagation and the possible distortion trends can be identified. This information can reveal the offset patterns that may occur in certain regions of the light beam, helping us understand the propagation characteristics of the light beam in complex media.

[0031] Establishing a light beam drift model includes: The mathematical expression of the objective function of the light beam drift model is: ; Among them, is the offset vector of the light beam, A is the coefficient of the light beam velocity, used to describe the linear velocity of the light beam drift, B is the coefficient of the light beam acceleration, used to describe the acceleration of the light beam drift, and C is the constant term, used to represent the initial position; Based on the objective function, perform support vector machine prediction training on the offset trend graph to obtain the light beam drift model and output the prediction of the light beam drift behavior.

[0032] By fitting multi-point monitoring data, the coefficients A and B can be estimated, and a mathematical model suitable for the current light beam drift can be constructed. According to the established drift model, error correction can be performed through a real-time feedback mechanism. For example, if it is detected that the light beam offset is too large, the optical path compensation of the system or the pointing of the laser source can be adjusted to maintain the stability of the light beam. At the same time, according to the prediction results of the drift model, compensation measures can be taken in advance to avoid further expansion of the light beam offset.

[0033] The prediction weight optimization unit is used to calculate the offset increment, profile change rate, and drift stability coefficient of the light beam within different time windows through a multivariate offset trend analysis algorithm. The offset increment is the displacement change amount of the light beam within different time windows. The profile change rate is the change speed of the light beam shape, which is obtained by measuring the difference in the spot width within different time windows and taking the derivative. This change rate reflects the degree of expansion of the light beam and is usually closely related to the perturbation of the light beam during propagation and the medium change. A larger profile change rate may mean that the light beam is undergoing severe distortion or instability. The drift stability coefficient is used to measure the stability of the light beam drift and is obtained by calculating the variance of the offset increment; Collect historical drift states through big data, perform weighted regression calculations on all historical drift states to obtain the influence of historical data on the current drift trend. The prediction weight in the weighted regression calculation is the influence degree of each historical drift state data point on the current corresponding drift state, and the influence degree is obtained by calculating the probability of the t-distribution based on the time axis. Dynamically adjust the prediction weight based on historical data and the current drift state. For example, if the offset trend of the light beam becomes unstable at a certain moment, the prediction weight at the current moment can be increased according to the change rate and the drift stability coefficient to ensure accurate prediction of the upcoming drift.

[0034] After the prediction weight is updated, it is used for real-time drift compensation of the light beam. Adjust the optical path compensator and the laser source positioning device according to the current drift prediction weight to ensure the stability of the light beam in the complex medium.

[0035] Through the multi-variable offset trend analysis algorithm, the offset increment, profile change rate, and drift stability coefficient of the light beam can be accurately analyzed within multiple time windows, so as to extract the influence weight of the historical drift state on the current drift trend, and can effectively predict and optimize the drift behavior of the light beam during propagation. Finally, the system can dynamically adjust the optical path compensation and other adjustment measures through the real-time feedback mechanism to ensure the stability and coherence of the light beam.

[0036] The real-time control adjustment unit is used to calculate the future drift path and offset correction amount according to the optimized weight, and adopt a control strategy that combines short-period error correction and long-period trend correction. The control strategy logic includes: The short-period correction amount is used to immediately adjust the offset direction of the light beam, and the long-period correction amount is used to adjust the overall collimation direction and stability parameters of the system to ensure that the light beam maintains a high-precision collimation state during long-term transmission; When the sensor fails, switch to the light beam path prediction mode based on historical data and enable the standby grid layer; When the main channel fails, switch to the redundant channel through the polarization beam splitter.

[0037] In the specific implementation process of this solution, the deep reinforcement learning DRL is used to optimize the control strategy, combined with the error feedback data, to iteratively train the drift control parameters, so that the system can quickly converge to the optimal collimation state under different environmental disturbances.

[0038] The present invention improves the beam collimation accuracy, reduces the error accumulation effect, and through a hierarchical grid error equalization control mechanism, realizes the collaborative compensation of the errors of transmitted and reflected beams, avoids the decrease of collimation accuracy caused by error accumulation, enhances the propagation stability of the beam in complex media, and based on the optical path non-uniform phase dynamic compensation strategy, can effectively correct the non-linear deviation of the optical path, improve the propagation accuracy of the long-distance beam, improve the beam drift prediction ability, realize dynamic control optimization, adopts a multi-variable adaptive weight optimization mechanism, combines the short-period error correction and long-period trend correction strategies, ensures that the beam maintains high stability under different environmental interferences, improves the automatic calibration ability of the system, reduces the need for manual intervention, realizes adaptive error compensation and dynamic adjustment through intelligent optimization algorithms, improves the self-adaptive ability of the system, and reduces the complexity of traditional optical systems relying on manual adjustment.

[0039] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code therein. Among them, the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, abbreviated as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, abbreviated as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, abbreviated as EPROM), programmable read-only memory (Programmable Red-Only Memory, abbreviated as PROM), read-only memory (Read-Only Memory, abbreviated as ROM), magnetic memory, flash memory, magnetic disk or optical disc. These computer program instructions can also be stored in a computer-readable memory capable of guiding a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured article including an instruction device, and the instruction device realizes the functions specified in one process Figure 1 one process or multiple processes and / or boxes Figure 1 specified in one box or multiple boxes.

[0040] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A layered grid alignment system based on automatic calibration, characterized in that: It includes a layered grid error equalization control module, an optical path non-uniform phase dynamic compensation module, and a beam drift prediction control module with adaptive weight adjustment: The layered grating error equalization control module is used to perform dynamic error equalization compensation for light beams with different incident angles and polarization states through a coordinated error compensation strategy between the transmission grating TG and the reflection grating RG; The optical path non-uniform phase dynamic compensation module is used to generate a compensation map, characterize the phase error of the light beam in different propagation areas, calculate the optimal phase adjustment compensation amount, and optimize the compensation weight coefficient; The beam drift prediction control module with adaptive weight adjustment is used to establish a beam drift model, perform prediction training of a support vector machine on the offset trend graph based on the objective function, obtain the beam drift model, calculate the future drift path and offset correction amount based on the optimized weights, and adopt a control strategy that combines short-cycle error correction with long-cycle trend correction.

2. The automatic calibration based layered grid alignment system according to claim 1, characterized in that: At the grating interface, the incident light angle deviation, wavefront distortion and energy distribution non-uniformity are sampled independently, the phase change gradient of the transmitted light beam and the directional deviation trend of the reflected light beam are extracted, and the transmission error and reflection error are calculated. When the light beam is incident on the grid surface, the angle deviation between the incident light beam and the ideal incident angle is measured by an optical sensor, where the angle deviation is the difference between the angle of the light beam and the preset ideal collimation direction; Collecting a plurality of incident light angle data at different positions and at different time points, and calculating the angle deviation between each sampling point and the ideal incident direction based on the incident light angle data; A wavefront sensor is used to perform wavefront sampling at the grating interface and measure the wavefront morphology of the light beam. The wavefront morphology of the light beam includes distortion caused by medium inhomogeneity, phase change of the light beam itself or system error. The phase value of each sampling point is obtained through wavefront sampling, and the change of the wavefront morphology of the light beam with time series is recorded. The phase change gradient of the wavefront morphology of the light beam is calculated based on the wavefront morphology of the light beam and the phase value of each point. The phase change of the wavefront morphology of the light beam is obtained by calculating the phase difference between adjacent sampling points.

3. The automatic calibration based layered grid alignment system according to claim 2, characterized in that: Extract the key phase change features of the beam wavefront morphology, the key phase change features include wavefront curvature, distortion amplitude and phase change gradient, collect the energy distribution data of the beam at the grid interface through a beam profile analyzer, the energy distribution data is a three-dimensional energy distribution diagram, calculate the light intensity difference between the light intensity value in the energy distribution data and the ideal uniform light intensity, and evaluate the non-uniformity of the energy distribution according to the light intensity difference. The evaluation logic includes: Calculate the standard deviation of the light intensity difference values ​​corresponding to all phase points, preset a standard deviation threshold, and when the standard deviation of the light intensity difference values ​​of the phase points corresponding to the energy distribution points is greater than the standard deviation threshold, the energy distribution is non-uniform; A precise direction sensor is used to sample in the direction of the reflected light beam, measure the directional deviation of the reflected light beam, and record the cause label of the directional deviation, which includes microstructure changes of the grating, surface unevenness and external disturbances. The direction change of the reflected light beam is recorded in time series, and the angular deviation trend between the reflected light beam and the ideal reflection direction is calculated. The reflection error is obtained by weighted averaging the angular deviation trend and the phase change of the light beam wavefront morphology, and the transmission error is obtained by weighted summing the phase error and the intensity error.

4. The automatic calibration based layered grid alignment system according to claim 3, characterized in that: According to the transmission error and the reflection error, the microstructure and period parameters of the compensation grating are dynamically adjusted, and the period gradient of the transmission grating and the microstructure inclination angle of the reflection grating are dynamically adjusted through a local period adjustment algorithm based on gradient optimization to compensate for the error. The compensation control logic includes: The local error weights of the transmission error and the reflection error are dynamically calculated, and the mathematical expressions of the local error weights of the transmission error and the reflection error are: ； ； in, is the angle deviation, is the maximum angle deviation, is the phase change, is the theoretical maximum value of phase change, is the polarization offset, is the theoretical maximum value of polarization offset, is the wavelength drift, is the theoretical maximum value of wavelength drift, is the local error weight of the transmission error, is the local error weight of the reflection error, , , and is the weight coefficient.

5. The automatic calibration based layered grid alignment system according to claim 4, characterized in that: The optical path non-uniform phase dynamic compensation module includes an optical path deviation measurement unit, a dynamic compensation mapping generation unit and an adaptive phase correction unit: The optical path deviation calculation unit is used to extract the optical path deviation characteristics, independently calculate the medium refractive index gradient, the beam path disturbance offset and the optical path cumulative error during the light beam propagation process according to the optical path deviation characteristics, and calculate the influence of the optical path of each area on the phase distortion; Extracting the optical path deviation feature includes constructing an optical path deviation matrix and then performing phase error correction. The optical path deviation matrix is ​​obtained by arranging the optical path deviations of all sampling points. The mathematical expression for calculating the optical path deviation is: ； ； in, is the total length of the transmission medium, is the ideal refractive index of the medium, is the optical path deviation, is the phase value of the sampling point, is the ith equally divided medium segment, is the refractive index of the ith equidistributed medium, is the length of the i-th equally divided medium, is the ideal optical path; When it is in the visible light wavelength range, the calculated diffraction wavelength of the optical path deviation is decomposed to obtain diffraction waves of different colors in the local area, the composition and intensity of the enhanced color light corresponding to each observation angle are recorded, the superposition state of the diffraction wave and the color corresponding to the superposition state of the diffraction wave are extracted, and the color superposition database is collected using big data. According to the collected color superposition database, the superposition color of the diffraction wave corresponding to the superposition state of each observation angle and the number of superimposed diffraction waves are queried respectively; By moving one of the decomposed diffraction waves of the superimposed diffraction waves through an optical delay line, optical path disturbance detection is introduced. The optical delay line of the other diffraction wave performs real-time compensation according to the optical path deviation value, and the current optical path deviation is calculated twice. The absolute value of the difference between the two optical path deviations is calculated as the target residual.

6. The automatic calibration based layered grid alignment system according to claim 5, characterized in that: The dynamic compensation map generation unit is used to generate a compensation map, characterize the phase error of the light beam in different propagation areas, and calculate the optimal phase adjustment compensation amount. The mathematical expression for loading the reverse phase distribution in the optical path compensator is: ； in, For reverse phase, is the wave number, is the target residual.

7. The automatic calibration based layered grid alignment system according to claim 6, characterized in that: The adaptive phase correction unit comprises: The time derivative of the optical path deviation and the phase change is calculated to obtain the optical path change rate and the phase change rate per unit time, and a change rate threshold is preset. When the optical path change rate and the phase change rate reach the change rate threshold, the compensation weight coefficient is automatically adjusted based on the mathematical expression of the local error weight. The optimization objective function for optimizing the compensation weight coefficient is: ； in, is the reverse phase after current compensation, To optimize the objective function; The optimal compensation weight coefficient is obtained by minimizing the value of the optimization objective function. Based on the feedback phase change data and the optical path deviation, the phase of the optical path compensator is adjusted according to the optimal compensation weight coefficient calculated by the optimization algorithm.

8. The automatic calibration based layered grid alignment system according to claim 7, characterized in that: The beam drift prediction control module with adaptive weight adjustment includes a drift state quantity dynamic acquisition unit, a prediction weight optimization unit and a real-time control adjustment unit: The drift state quantity dynamic acquisition unit is used to collect key parameters such as the beam center position offset, beam diffusion angle change rate and drift acceleration in real time through a multi-point monitoring mechanism, compare the beam main axis offset trends at different positions, and establish a beam drift model; The state information of the light beam is collected by the CCD array. The state information includes the coordinate change of the center of the light beam and the beam diffusion angle. The offset is calculated according to the coordinate change of the center of the light beam. The mathematical expression for calculating the offset is: ； ； in, and is the offset of the coordinate change at time t, and is the measured beam center position, and It is the ideal reference position of the light beam; The diffusion angle unit rate is calculated by derivation, the drift acceleration is obtained by performing a second-order derivative on the offset, the center position offset, the diffusion angle change rate and the drift acceleration of the collected light beam are calculated at different array nodes of the CCD array through a multi-point monitoring mechanism, and a deviation trend graph is drawn based on the two-dimensional position coordinate axis of the light beam measured by the CCD array, wherein the horizontal axis coordinate of the deviation trend graph is the position parameter, and the vertical axis coordinate is the center position offset, the diffusion angle change rate and the drift acceleration of the light beam; Modeling beam drift involves: The mathematical expression of the objective function of the beam drift model is: ； in, is the offset vector of the beam, A is the coefficient of the beam velocity, which is used to describe the linear velocity of the beam drift, B is the coefficient of the beam acceleration, which is used to describe the acceleration of the beam drift, and C is a constant term, which is used to represent the initial position; Based on the objective function, the support vector machine prediction training is performed on the deviation trend diagram to obtain the beam drift model and output the beam drift behavior prediction.

9. The automatic calibration based layered grid alignment system according to claim 8, characterized in that: The prediction weight optimization unit is used to calculate the offset increment, profile change rate and drift stability coefficient of the light beam in different time windows through a multivariate offset trend analysis algorithm. The offset increment is the displacement change of the light beam in different time windows. The profile change rate is the change speed of the light beam shape. The profile change rate is obtained by measuring the difference in the spot width in different time windows and deriving it. The drift stability coefficient is used to measure the stability of the light beam drift and is obtained by calculating the variance of the offset increment. The historical drift states are collected through big data, and weighted regression calculation is performed on all historical drift states to obtain the influence of historical data on the current drift trend. The prediction weight in the weighted regression calculation is the influence degree of each historical drift state data point on the current corresponding drift state, and the influence degree is obtained by probability calculation based on the t distribution of the time axis; The prediction weights are updated and used for real-time drift compensation of the light beam, and the optical path compensator and the laser source positioning device are adjusted according to the current drift prediction weights.

10. The automatic calibration based layered grid alignment system according to claim 9, characterized in that: The real-time control adjustment unit is used to calculate the future drift path and offset correction amount according to the optimized weights, and adopt a control strategy combining short-term error correction with long-term trend correction. The control strategy logic includes: The short-period correction is used to instantly adjust the beam deviation direction, and the long-period correction is used to adjust the overall collimation direction and stability parameters of the system; When the sensor fails, it switches to the beam path prediction mode based on historical data and activates the backup grid layer; When the main channel fails, it switches to the redundant channel through a polarization beam splitter.

Images