Single-mode fiber adaptive coupling method in turbulent environment

Abstract

The invention discloses a single-mode optical fiber adaptive coupling method in a turbulent environment, and relates to the fields of single-mode optical fiber coupling, automatic control, optical fiber lasers, optical fiber imaging and the like. According to the method, the relation between the coupling efficiency and the offset distance is analyzed, and a relation curve is simplified into a Gaussian model for adaptively updating the subsequent closed-loop gain rate. In addition, a gradient direction random generation mode in a traditional control method is changed into omni-directional disturbance, and the gradient direction which accurately and rapidly approaches the target value is obtained through calculation. Two-dimensional perturbation is projected to a one-dimensional space, a least square method is used, the step length and the gradient direction needed by iteration are automatically calculated, and therefore rapid capturing and tracking of focusing light spots are achieved. Compared with a traditional control method, the method has the advantages that the convergence stability is effectively improved while the convergence speed of the algorithm is guaranteed, so that the system has a wider effective correction area, and meanwhile, the closed-loop control bandwidth range of disturbance correction is widened.

Description

TECHNICAL FIELD

[0001] The present application relates to the fields of single-mode fiber coupling, automatic control, fiber laser, fiber imaging, and the like, and in particular to a single-mode fiber adaptive coupling method in a turbulent environment, which has an important application prospect in free space laser communication. BACKGROUND

[0002] As a new type of communication technology, free space laser communication (FSOC) has the advantages of high transmission rate, unlimited bandwidth, good security performance, and has been widely used in inter-satellite, satellite-ground communication and other occasions. In the process of the development of free space optical communication system, more and more mature optical fiber communication technology is combined, which also makes the coupling of space laser to single-mode fiber a key problem in the field of FSOC. In the process of fiber coupling, the coupling efficiency directly affects the communication bit error rate and determines the key factors such as communication transmission distance. The traditional method of optimizing optical lens design and fiber manufacturing to improve coupling efficiency can only improve the static coupling efficiency, and the actual FSOC coupling also needs to cope with the dynamic error problem caused by atmospheric turbulence and mechanical vibration of the carrier platform. In addition to optimizing the structure of the coupling device to achieve the ability to resist dynamic disturbance, the control method of the system to realize fiber coupling is also worth paying attention to.

[0003] At present, the common control methods of fiber coupling include hill climbing method, nutation algorithm, stochastic parallel gradient descent (SPGD) algorithm, etc. The first two methods take a long time and have insufficient correction and tracking ability for dynamic disturbance. SPGD algorithm is currently the most widely used, especially in adaptive optics systems without wavefront sensors, coherent synthesis and fiber coupling, but this method can only obtain good results when the relative position difference between the fiber and the spot is small, and the compensation ability of the fiber coupling efficiency under the vibration of the receiving platform and the disturbance of atmospheric turbulence is limited. At the same time, the space optical coupling system is very concerned about the influence of the received signal light power on the bit error rate of the communication system, and the bit error rate often depends on the minimum value of the coupling power, so stable coupling is more important than high average coupling to a certain extent. In addition, the fiber coupling algorithm cannot be automated at present, and many parameters need to be adjusted manually by experienced engineers, which greatly limits the communication time and application scenarios, especially in the atmospheric turbulence environment, there is light flicker, even if it is manually adjusted, it is also difficult to find the optimal control parameters. Therefore, it is very important to improve the effective range of the algorithm, while maintaining the convergence speed, convergence range and stability of the algorithm after convergence, and to realize the full automation of the algorithm in the flickering environment. The current control methods cannot take into account these characteristics.

[0004] In summary, there is an urgent need to find a fully automatic coupling algorithm with stable convergence, anti-disturbance, and a large effective range to meet the control requirements of fiber coupling in actual dynamic application scenarios. Summary of the Invention

[0005] The technical problem to be solved by this invention is to overcome the poor closed-loop stability, narrow effective correction bandwidth, and low degree of automation of traditional control algorithms in applications involving spatial light coupling to single-mode optical fibers. Furthermore, the problem of convergence failure caused by the inability to obtain gradient information when there is a large positional offset between the optical fiber end face and the optical spot can be overcome. Based on this, a method for adaptive coupling of single-mode optical fibers in turbulent environments is proposed.

[0006] The technical solution adopted by the present invention to solve its technical problem is:

[0007] To address these pressing technical challenges, a single-mode fiber adaptive coupling method (Least Squares Iterative algorithm, LSI) for turbulent environments was proposed. Compared to the traditional SPGD algorithm, the LSI algorithm adjusts both the gradient direction and the gain rate (step size). By combining it with fiber optic two-dimensional imaging scanning technology, it achieves "blind-breaking" optimization, resulting in higher stability, a wider effective convergence range, and a higher-frequency effective correction bandwidth. The implementation steps are as follows:

[0008] A single-mode fiber adaptive coupling method in a turbulent environment is proposed. Before implementing this method, a spatial optical coupling experimental platform is first established to complete the device selection and parameter design including the laser, laser emitting single-mode fiber, laser emitting collimating lens, fiber coupling device and controller. The method is implemented through the following steps:

[0009] Step S110, calibration scanning, includes: inputting scanning voltage range, scanning points, scanning interval time, the controller outputs corresponding voltage signal to drive the optical fiber end face to complete scanning on the two-dimensional plane, and the scanning obtains a one-to-one correspondence between different voltage positions and performance indicators, that is, , solve the parameters required for the Gaussian model: Gaussian model standard deviation ;in, is the performance index at the i-th scanning position, is the control voltage corresponding to the X-axis direction at the i-th step scanning position, is the control voltage corresponding to the Y-axis direction at the i-th step scanning position;

[0010] Step S120, setting the basic parameters of the system closed-loop control, including: the disturbance rate of small disturbances , closed-loop frequency f, and initial voltage ,in and are the control voltages of the X-axis and Y-axis before the closed loop starts, respectively;

[0011] In step S130, when the closed loop starts, the photoelectric detector converts the optical signal into an electrical signal after detecting the optical signal, and inputs the electrical signal into the AD converter, and the controller reads the coupling light energy value at this time as the initial performance index for control;

[0012] In step S140, gradient acquisition: the controller executes the LSI algorithm, and according to the input disturbance rate , applies omnidirectional disturbance, obtains the corresponding coupling light energy size when the positive and negative disturbance voltages are loaded in different directions, calculates the current optimal gradient direction, and projects the performance index corresponding to the disturbance in the gradient direction;

[0013] In step S150, gain rate, that is, iteration step length update: according to the Gaussian model, a system of over-determined equations is constructed, and the least square method is used to update the gain rate , wherein, is the calculated gain rate, is a positive integer from 1 to 4, is the performance index corresponding to the four omnidirectional disturbances, is the projection of the four omnidirectional disturbances in the accurate gradient direction, is the initial performance index obtained in step S130, is the standard deviation of the Gaussian model calculated in step S110;

[0014] In step S160, at this time, the direction and step length of the gradient descent are determined, and the driving voltage of the position of the fiber end face is updated, and the fiber end face is quickly shifted to the corresponding iteration point;

[0015] In step S170, steps S130-S160 are repeated, so that the coupling efficiency of the system converges to the optimal value after multiple iterations and can resist disturbances and other conditions in the closed loop process to maintain stability.

[0016] The beneficial effects of the present application compared with the prior art are:

[0017] The method based on the Gaussian model is adopted to adjust the gain rate, so that the gain rate can be adaptively updated according to the collected light energy, the convergence speed of the algorithm is ensured, the convergence stability, the effective convergence range and the effective correction bandwidth are improved, the interference in the closed loop process can be quickly corrected, and the algorithm will not fall into a local extreme value or cannot converge with the increase of the iteration number or the disturbance number.

[0018] The application utilizes two-dimensional scanning imaging technology of the optical fiber in solving Gaussian model parameters, visualizes the current coupling state without referring to additional equipment, breaks the traditional SPGD 'blind optimization' feature, additionally adds image information, and improves system usability.

[0019] The application realizes automatic operation of single-mode optical fiber coupling, can adaptively solve iteration steps even in a turbulent environment with light intensity flickering, improves unmanned operation capability of the system, widens application scenarios, and improves use convenience.

[0020] In addition, the application realizes full-automatic coupling without parameter adjustment, does not need to introduce additional monitoring equipment or real-time debugging by technical personnel, greatly increases the usability of the system, and has important application prospects in space optical communication. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 For the application principle diagram of a single-mode optical fiber adaptive coupling method in a turbulent environment, the simulated coupling efficiency curve, and the specific experimental principle diagram of the AFC optical fiber coupling system under the wavelength of incident laser light of 1550nm, wherein (a) is a basic principle diagram of single-mode optical fiber coupling, (b) is a curve of coupling efficiency changing with offset under the condition that the wavelength of laser light is 1550nm, the mode field radius of single-mode optical fiber is 5um, the focal length of the coupling lens is 47.4mm, and the clear aperture is 20mm, and (c) is a schematic diagram of a spatial light coupling experiment platform based on an adaptive optical fiber coupler.

[0022] Figure 2 For the algorithm flowchart of a single-mode optical fiber adaptive coupling method in a turbulent environment, it contains the calibration scanning, accurate gradient direction measurement, and LSI iteration correction process.

[0023] Figure 3 For the application of a single-mode optical fiber adaptive coupling method in a turbulent environment in an initial large disturbance deviation experiment (Example 1), the experimental results of the SPGD algorithm and the LSI algorithm of the application are compared, wherein (a) and (d) are the complete coupling power corresponding performance index value curves, including the data in the open loop stage, the large initial deviation stage, the closed loop stage, and the open loop stage again, (b) and (e) show the detailed data from the large initial deviation stage to the closed loop stage, (c) and (f) are the control voltages of the AFC in the control process using the LSI algorithm and the SPGD algorithm respectively.

[0024] Figure 4This is a comparison chart of the experimental results of the SPGD algorithm and the LSI algorithm of the present invention in a variable frequency dynamic disturbance experiment of a single-mode optical fiber adaptive coupling method in a turbulent environment (Example 2) of the present invention, wherein Figures (a) to (d) are corresponding performance index value curves collected by the photoelectric detector of the LSI algorithm of the present invention, and (e) to (f) are corresponding performance index value curves collected by the photoelectric detector of the SPGD algorithm. DETAILED DESCRIPTION

[0025] The present invention will be further described below with reference to the accompanying drawings and two embodiments, but the scope of protection of the present invention shall not be limited thereto.

[0026] This invention provides a method for adaptively coupling single-mode optical fibers in turbulent environments. Before implementing this method, a spatial optical coupling experimental platform must be constructed to select and design the components, including the laser, laser-emitting single-mode optical fiber, laser-emitting collimating lens, fiber coupling device, and controller. The fiber coupling device also requires designing parameters such as the mode field radius of the coupled-receiving single-mode optical fiber and the focal length and aperture of the coupling lens.

[0027] Lasers, single-mode optical fibers, and collimators are used as the transmitting end in optical coupling experiments. Their parameters can be set according to actual communication requirements. The laser output wavelength can be in common bands such as 1550nm, 1064nm, and 808nm.

[0028] The fiber coupling device is not limited, and only needs to have a coupled receiving single-mode fiber and a two-dimensional adjustable driver, such as an adaptive fiber coupler (AFC), a fast reflector, etc.

[0029] The controller on which the control algorithm relies can be an integration of various high-precision analog-to-digital conversion (AD) and digital-to-analog conversion (DA) chips with real-time signal processing chips such as FPGA (Field Programmable Gate Array) and DSP (Digital Signal Processor). Alternatively, closed-loop control can be achieved using an ordinary PC (Personal Computer) and a supporting data acquisition card.

[0030] like Figure 2 As shown, the method includes the following steps:

[0031] Step S110, calibration scanning, includes: inputting scanning voltage range, scanning points, scanning interval time, the controller outputs corresponding voltage signal to drive the optical fiber end face to complete scanning on the two-dimensional plane, and the scanning obtains a one-to-one correspondence between different voltage positions and performance indicators, that is, , the parameters required by the Gaussian model are solved: Gaussian model standard deviation . Wherein, is the performance index at the i-th scanning position, is the control voltage corresponding to the X-axis direction at the i-th scanning position, is the control voltage corresponding to the Y-axis direction at the i-th scanning position. This step is performed because: when the spatial laser beam enters the coupling lens, an Airy diffraction pattern will be formed on the lens focal plane, and according to the mode field matching principle, the light power distribution obtained by coupling is similar to a Gaussian distribution. Static scanning calibration in the room needs to be completed before the experiment, and subsequent scanning is not required to complete closed-loop control. In the case of static undisturbance, the algorithm calibration is first performed. First, the optical path is aligned so that the outgoing parallel light is focused to the center of the fiber end face through the coupling lens. Then, according to the allowed voltage range of the device, the upper and lower limits of the scanning, the scanning accuracy, etc. are set, so as to complete the two-dimensional scanning calibration. The performance index J obtained by scanning is used for imaging, and the required parameters of the model are calculated.

[0032] The scanning voltage range, scanning accuracy, scanning frequency, etc. can be designed according to the performance of the system. The actual scanning mode is not limited, and in the embodiment of the present application, “bow” scanning is adopted, and “Z” scanning, “back” scanning, etc. can be adopted.

[0033] In step S120, the basic parameters of the system closed-loop control are set, including: the disturbance rate of the small disturbance , the closed-loop frequency f, and the initial voltage . Wherein and are the control voltages of the X-axis and the Y-axis before the closed loop starts, respectively.

[0034] Wherein, the disturbance rate can be fixed according to the static scanning result, such as the disturbance of the focused light spot at the center of the fiber end face, and the coupling efficiency after the disturbance is preferably 90% of the optimal coupling efficiency.

[0035] The closed-loop frequency f is set, which can be selected according to the lower one of the tolerable closed-loop frequency of the fiber coupling device (such as the resonance frequency of the fast mirror) and the executable closed-loop frequency of the controller (such as the calculation frequency of the PC).

[0036] In step S130, when the closed loop starts, the photodetector converts the optical signal into an electrical signal after detecting the optical signal, and inputs the electrical signal into the AD converter. The controller reads the coupling light energy value at this time as the initial performance index of the control. The purpose of the closed-loop process is to maximize the coupling light energy, and at this time the corresponding coupling efficiency reaches the highest.

[0037] Among them, the incident laser wavelength should meet the response band requirements of the photodetector; the coupled signal needs to be split and processed, and the specific splitting ratio (such as 95:5) can be designed based on actual communication needs and the rated power of the detector and other requirements.

[0038] The function of a photodetector is to linearly convert optical power to a voltage signal. Its operating wavelength range should cover the wavelength of the signal laser beam. Photodiodes, phototransistors, avalanche diodes, photomultiplier tubes, or other photodetectors can be used, as long as they can achieve a linear conversion between optical and voltage signals. The controller type is not limited. For applications with low bandwidth requirements, a standard PC can be used as the control device. For applications with high bandwidth requirements, real-time signal processing chips such as FPGAs and DSPs can be used.

[0039] S140, Gradient Acquisition: The controller executes the LSI algorithm and calculates the gradient according to the input disturbance rate. Size, apply omnidirectional disturbance, obtain the corresponding coupling light energy size when positive disturbance voltage and negative disturbance voltage are loaded in different directions, calculate the current optimal gradient direction, and project the performance index corresponding to the disturbance on the gradient direction.

[0040] The rules for applying omnidirectional disturbance voltage are as follows: First, positive and negative disturbance voltages are applied only in the X-axis direction, and no disturbance voltage is applied in the Y-axis direction. The magnitude of the disturbance voltage is determined by the disturbance rate. Decision; the detector obtains the corresponding performance index at this time ,in, Represents the performance index obtained by applying a positive disturbance voltage in the X-axis direction, Represents the performance indicator acquisition function, Represents the control voltage in the X-axis direction in the t-1 iteration, t is the current iteration number, Represents the magnitude of the disturbance voltage in the X-axis direction, Representatives in The control voltage in the Y-axis direction in the iteration is, Represents the performance index obtained by applying a negative disturbance voltage in the X-axis direction; secondly, the detector obtains the corresponding performance index when positive and negative disturbance voltages are applied only in the Y-axis direction and no disturbance voltage is applied in the X-axis direction. ,in, and Respectively represent the performance indicators obtained by applying positive disturbance voltage and negative disturbance voltage in the Y-axis direction, Represents the magnitude of the disturbance voltage in the Y-axis direction. The gradient direction is calculated based on the magnitude of the coupled light energy corresponding to the positive disturbance voltage and the negative disturbance voltage in different directions. The calculation rules are as follows: The X-axis gradient is , the Y-axis gradient is wherein, and respectively represent the performance index obtained by applying positive and negative perturbation voltage in the X-axis direction, and respectively represent the performance index obtained by applying positive and negative perturbation voltage in the Y-axis direction, is a very small quantity to prevent the denominator from being zero. The voltage corresponding to the small perturbation is projected in the direction of the accurate gradient to obtain the scalar of the fourth perturbation voltage in the projection direction: , , , wherein, is the perturbation rate, and are the calculated X-axis gradient and Y-axis gradient respectively.

[0041] S150, gain rate, i.e. iteration step length update: according to the Gaussian model, an overdetermined equation set is constructed, so that the gain rate is updated using the least square method wherein, is the calculated gain rate, is a positive integer from 1 to 4, is the performance index corresponding to the fourth omnidirectional perturbation, is the projection of the fourth omnidirectional perturbation in the direction of the accurate gradient, is the initial performance index obtained in step S130, is the standard deviation of the Gaussian model calculated in step S110.

[0042] Step S160, at this point, the direction and step length of the gradient descent are determined, the driving voltage of the position of the fiber end face is updated, and the fiber end face is quickly shifted to the corresponding iteration point.

[0043] wherein, the voltage update rule is: wherein, t is the current iteration number, represents the control voltage in the X-axis direction of the tthiteration, represents the control voltage in the X-axis direction of the (t-1)thiteration, represents the gain rate, and are the X-axis gradient and Y-axis gradient respectively, represents the control voltage in the Y-axis direction of the tthiteration, represents the control voltage in the Y-axis direction of the (t-1)thiteration.

[0044] Step S170, repeat steps S130-S160, so that the coupling efficiency of the system converges to the optimal value after multiple iterations and can resist disturbances in the closed loop process to maintain stability.

[0045] The basic principle of the present application is that as shown in (a) of FIG. 1, the coupling efficiency is defined as the ratio of the power coupled into the fiber to the power of the incident light at the pupil plane of the coupling lens. According to the Parseval theorem, the power of the focused spot at the fiber end face S0 is equivalent to the power of the incident light at the incident pupil plane S1, so the coupling efficiency formula can be expressed as: Figure 1

[0046]

[0047] wherein, represents the coupling efficiency, represents the power of the light coupled into the single-mode fiber, represents the power of the focused spot, is the incident light field at the focal plane S0 of the coupling lens, is the complex conjugate of is the light field transmitted in the single-mode fiber, is the coordinate of the X-axis and the Y-axis in the rectangular coordinate system, and in the polar coordinate can be expressed as:

[0048]

[0049] wherein is the circular constant, represents the imaginary unit, is the focal length of the coupling lens, and D is the aperture of the coupling lens, is the wavelength, is the distance from any point on the focal plane to the center of the Airy disk, is the wave number is the zero-order first kind Bessel function.

[0050] The alignment error between the fiber end face and the focused spot will cause a decrease in the coupling efficiency. This positional mismatch can be caused by the incident light field with tilt aberration, or by the inaccurate positioning of the fiber end face. Because the incident angle deviation is equivalent to the shift of the fiber end face, the beam transmission in the single-mode fiber follows the Nakagami-Rice distribution rule, which can be expressed as:

[0051]

[0052] wherein, is the shift distance of the fiber end face from the center of the focused spot, and k is the wave number is the zero-order modified Bessel function, is the mode field radius of the single-mode fiber. ​​​​​​​

[0053] Combining the above formula, the coupling efficiency can be simplified as:

[0054] .

[0055] The present invention is based on an adaptive fiber coupler (AFC) to illustrate this. , single-mode fiber mode field radius , the focal length and clear aperture of the coupling lens are and The simulation results show that the coupling efficiency changes with the offset. Figure 1 As shown in (b), the curve is approximately Gaussian, and each offset distance corresponds to a unique coupling efficiency value. The coupling efficiency model is approximated by the Gaussian model:

[0056] ,

[0057] in, is the coupling efficiency approximately fitted by the Gaussian model, is the offset distance between the fiber end face and the center of the focused light spot, are the Gaussian model parameters.

[0058] The calculation of Gaussian model parameters can be done by obtaining the real coupling efficiency curve through static scanning, and taking the middle value for solution:

[0059] ,

[0060] in, It is the offset distance between the fiber end face and the center of the focused spot corresponding to half of the maximum coupling efficiency obtained by scanning. is half of the maximum coupling efficiency obtained by scanning, Represents the logarithm operation with a natural constant as the base.

[0061] The present invention proposes to obtain accurate gradient direction through omnidirectional perturbation and realize the solution conversion from two-dimensional to one-dimensional through perturbation projection. for:

[0062] ,

[0063] in, is the gradient direction without adding a very small amount, represents the 2-norm, It is a very small quantity that prevents the denominator from being 0.

[0064] The one-dimensional projection of a two-dimensional small perturbation in the exact gradient direction can be expressed as:

[0065] ,

[0066] ,

[0067] ,

[0068] ,

[0069] wherein, , , and represent the projection of the positive X-axis disturbance, the negative X-axis disturbance, the positive Y-axis disturbance, and the negative Y-axis disturbance in the accurate gradient direction, respectively, is the disturbance rate, is the calculated accurate gradient direction.

[0070] Then, based on the Gaussian model, an overdetermined equation set is constructed:

[0071] ,

[0072] wherein i = 0, 1, 2, 3, 4, 5; JM is the coupled optical power under the optimal coupling efficiency in the current iteration period, and is an unknown number that needs to be offset when solving the equation.

[0073] Taking the logarithm operation on both sides of the above equation, we obtain:

[0074] ,

[0075] wherein j = 1, 2, 3, 4.

[0076] Subtracting the above two equations, JM is eliminated, and we obtain:

[0077] ,

[0078] Among them, only is the unknown number to be solved, so it is an overdetermined equation set, and a function H can be constructed to solve it by least squares:

[0079] ,

[0080] The derivative of H needs to be 0:

[0081] .

[0082] Therefore, using least squares iteration, the distance between the current fiber position and the ideal fiber position (i.e., the position corresponding to the maximum coupling efficiency) can be quickly calculated as the gain rate , combined with the calculated accurate gradient direction, the driving voltage of the optical fiber end face is updated, and the adaptive capture and tracking of the spatial laser focus spot by the single-mode optical fiber end face can be achieved.

[0083] ,

[0084] Where t is the current iteration number, Represents the control voltage vector in the X-axis and Y-axis directions of the t-th iteration, Represents the control voltage vector in the X-axis and Y-axis directions at the t-1th iteration.

[0085] The following two examples illustrate how to implement the present invention. Both examples are based on a spatial optical coupling platform with an adaptive fiber coupler, such as Figure 1 (c) is shown. The platform consists of a 1550nm laser, a collimator, a fast-reflecting mirror for interference generation, a coupling lens, a single-mode fiber, an optical power meter, a controller, a photodetector, a high-voltage amplifier, etc. The fiber coupling device that performs the overall coupling function includes a coupling lens (focal length 47.4mm, aperture 20mm), an AFC with a single-mode fiber (mode field radius 5 ), a 1×2 fiber beam splitter, a photodetector (InGaAs photodiode, model C12485-210), a dual-channel high-voltage amplifier, a PC controller, and an NI PCI-6221 A / D data acquisition card. It's worth noting that this platform uses an AFC as a dynamic coupling device, but in practice, a dynamic coupling device can be a fast mirror or other dynamic coupling device. As long as it has two-dimensional control capabilities, other requirements are not critical.

[0086] The laser output passes through a collimator and reaches a fast-reflecting mirror. It is then reflected by the coupling lens of the fiber coupling device and converged onto the single-mode fiber at the focal plane. The single-mode fiber is connected to a 1×2 fiber beam splitter. The light coupled into the single-mode fiber is split into two parts, 5% of which enters a photodetector. During the algorithm optimization process, the voltage of the photodetector will be used as a closed-loop performance metric. The remaining 95% of the energy enters an optical power meter, which displays the optical power coupled into the fiber in real time.

[0087] Example 1:

[0088] Example 1 is an initial large perturbation deviation experiment, which compares the effective convergence range of the two algorithms.

[0089] Before the experiment begins, the LSI algorithm requires static scanning calibration. This allows for direct closed-loop processing without repeated scans. After setting up the optical path platform, the coupling fiber position is adjusted to maximize coupling efficiency and ensure a single scan is completed without disturbance. The resulting scan image serves as the initial calibration data, from which the Gaussian model parameters are calculated.

[0090] Step 1) A signal generator generates a sinusoidal signal with an amplitude of 100mV and a frequency of 1Hz, which is continuously applied to the fast mirror to simulate dynamic disturbances such as turbulence or platform vibration. The controller applies a set of large bias voltages (-4V, -4V) to the AFC to simulate the large initial deviations that may occur in actual application scenarios.

[0091] Step 2) The controller executes the LSI algorithm of the present invention and updates the gain rate using the performance indicator J value of the current coupling power . (When executing the SPGD algorithm, the disturbance rate and gain ratio It is always the same as the initial setting value and does not need to be updated.)

[0092] Step 3) Based on the gain ratio and disturbance rate Calculate the required voltage The controller outputs the control voltage to the high-voltage amplifier, driving the fiber end face to quickly shift to the corresponding iteration point.

[0093] Step 4) Repeat steps 2) to 3) until the coupling efficiency of the system converges to the optimal value after multiple iterations or ends when the required number of iterations is reached.

[0094] Step 5) Switch to the SPGD algorithm and repeat steps 1) to 4).

[0095] The experimental results are as follows Figure 3 shown. Figure 3 (a) and (d) show the complete coupled power versus performance indicator curves, including data from the open-loop phase, the large initial deviation phase, the closed-loop phase, and the open-loop phase again. (b) and (e) show detailed data from the large initial deviation phase to the closed-loop phase. (c) and (f) show the AFC control voltage during the LSI and SPGD control processes, respectively.

[0096] from Figure 3 It can be clearly seen from (a) and (d) that when faced with such a large bias voltage as (-4V, -4V), it means that the fiber end face deviates far from the focus of the coupled light spot. At this time, the detectable optical power is very low, and the performance index data is close to 0. The extremely low coupling energy makes it impossible for the SPGD algorithm to complete effective control. (f) shows that the control voltage calculated by SPGD almost floats around the initial bias value, making it completely impossible for the fiber end face to locate the light spot. (b) to (c) show that the control voltage calculated by the LSI algorithm of the present invention can be explored quickly over a large range after the closed loop is established, so that the fiber end face can quickly locate the light spot and move it to its center. In addition, the present invention can also effectively correct the sinusoidal disturbance introduced by the fast reflection mirror after convergence. From Figure 3The control voltage of (c) and the closed-loop performance index curve of (a) in the figure can be observed, the application has good ability to track the sinusoidal disturbance: in the closed-loop, the performance index curve is basically stable, and the fluctuation of the coupling light energy is obviously smaller than that in the open-loop period; after re-opening the loop, the static large offset is effectively corrected, but the sinusoidal disturbance still exists, so the performance index curve also fluctuates near the peak. In summary, the effective range of the application is very wide, and it has good ability to offset dynamic disturbances.

[0097] Example 2:

[0098] Example 2 is a variable frequency dynamic disturbance experiment, which compares the effective correction bandwidth of two algorithms. The optical path is built the same as in Example 1, and the experimental parameters set by the control algorithm are also consistent with Example 1. The main difference between the two is that the frequency of the sinusoidal disturbance applied in the optical path is changed, as follows:

[0099] Step 1) The signal generator generates a sinusoidal signal with an amplitude of 300mV and a frequency of 5Hz, which continuously acts on the fast mirror to simulate dynamic disturbances such as turbulence or platform vibration.

[0100] Step 2) The controller executes the LSI algorithm of the application, which updates the gain rate based on the Gaussian model using the current coupling energy performance index J value (The disturbance rate and the gain rate always remain consistent with the initial set value and do not need to be updated).

[0101] Step 3) Calculate the required voltage based on the gain rate and the disturbance rate The controller outputs this control voltage to the high-voltage amplifier to drive the fiber end face to quickly shift to the corresponding iteration point.

[0102] Step 4) Repeat steps 2)~3) to make the coupling efficiency of the system converge to the optimal value after multiple iterations or reach the iteration requirement to end the closed loop and return to the open loop stage again.

[0103] Step 5) Use the SPGD algorithm, repeat steps 2)~4).

[0104] Step 6) Change the frequency of the sinusoidal signal generated by the signal generator, increase the frequency by 5Hz each time, and increase the frequency to 20Hz at the highest, to simulate dynamic disturbances such as turbulence or platform vibration at different frequencies. Repeat steps 2)~5).

[0105] Figure 4 The experimental results are shown in the following figures:The curves of (a)-(d) in the figure are the corresponding performance index value curves collected by the photodetector of the LSI algorithm of the application, including the data in the open loop stage, the closed loop stage and the second open loop stage. It can be observed that the closed loop correction effect gradually weakens as the disturbance frequency increases. This is due to the limitation of the closed loop iteration rate. In actual application scenarios, the closed loop delay can be reduced to increase the closed loop iteration frequency. The upper limit of this frequency is generally determined by the dynamic characteristics of the controlled device (such as the first-order resonant frequency of the AFC / fast mirror). In the four experiments of the LSI, the average value of the performance index J from the photodetector in the closed loop stage is obviously higher than that in the open loop stage, and the fluctuation range is smaller than that in the open loop stage. The minimum value of J is also higher than the minimum value in the open loop stage. After the second open loop, the algorithm corrects a certain amount of initial deviation, and the average value of the performance index J is higher than that in the first open loop. However, the sinusoidal disturbance still exists, and the fluctuation is still more violent than in the closed loop stage.

[0106] Figure 4 The curves of (e)-(f) in the figure are the corresponding performance index value curves collected by the photodetector of the SPGD algorithm, also including the data in the open loop stage, the closed loop stage and the second open loop stage. When the SPGD algorithm is controlled, effective correction can only be completed when the sinusoidal disturbance frequency is 5 Hz. The average value of the performance index J in the closed loop stage is higher than that in the open loop and the second open loop stage. Figure 4 When the sinusoidal disturbance frequency is 10 Hz, compared with (b), the performance index J value from the photodetector fluctuates violently, and its minimum value is lower than that in the second open loop stage, which means that good dynamic tracking is not achieved. When the sinusoidal disturbance frequency is higher than 10 Hz, the SPGD algorithm control has failed, the performance index J value curve diverges, and it cannot be stabilized within a certain range. At this time, the violent shaking caused by the failure of the SPGD algorithm control will cause damage to the device, and the test needs to be stopped. It can be considered that, under the same delay and the same disturbance, the effective correction bandwidth of the SPGD algorithm in the embodiment is close to 10 Hz, while the effective correction bandwidth of the LSI algorithm is greater than 20 Hz, which is twice that of the SPGD algorithm. This shows that the LSI algorithm of the application has a higher effective correction bandwidth.

[0107] At this point, the application has completed the detailed description of a single-mode fiber adaptive coupling method in a turbulent environment. In some examples, well-known methods, structures and techniques are not described in detail in order not to obscure the understanding of the present specification. Although the application is described in terms of a limited number of embodiments, those skilled in the art, with the benefit of the above description, will appreciate that other embodiments can be conceived within the scope of the application described herein. In addition, it should be noted that the language used in the specification is mainly selected for readability and teaching purposes, rather than for the purpose of interpreting or limiting the subject matter of the application.

Claims

1. A single-mode optical fiber adaptive coupling method in a turbulent environment, characterized in that: Before implementing this method, a spatial optical coupling experimental platform is first built to complete the device selection and parameter design including the laser, laser emission single-mode fiber, laser emission collimating lens, fiber coupling device and controller. This is achieved through the following steps: Step S110, calibration scanning, includes: inputting scanning voltage range, scanning points, scanning interval time, the controller outputs corresponding voltage signal to drive the optical fiber end face to complete scanning on the two-dimensional plane, and the scanning obtains a one-to-one correspondence between different voltage positions and performance indicators, that is, , solve the parameters required for the Gaussian model: Gaussian model standard deviation ;in, is the performance index at the i-th scanning position, is the control voltage corresponding to the X-axis direction at the i-th step scanning position, is the control voltage corresponding to the Y-axis direction at the i-th step scanning position; Step S120, setting the basic parameters of the system closed-loop control, including: the disturbance rate of small disturbances , closed-loop frequency f, and initial voltage ,in, and are the control voltages of the X-axis and Y-axis before the closed loop starts; Step S130: When the closed loop starts, the photodetector detects the light signal and converts it into an electrical signal which is input to the AD converter. The controller reads the coupled light energy value at this time as the initial performance indicator of the control. Step S140, gradient acquisition: The controller executes the LSI algorithm and calculates the gradient according to the input disturbance rate. Size, apply omnidirectional perturbation, obtain the corresponding coupling light energy size when loading positive perturbation voltage and negative perturbation voltage in different directions, calculate the current optimal gradient direction, and project the performance index corresponding to the perturbation on the gradient direction; Step S150, the gain rate, i.e., the iterative step size is updated: an overdetermined set of equations is constructed based on the Gaussian model, and the gain rate is updated using the least squares method: ,in, is the calculated gain rate, is a positive integer from 1 to 4, is the performance index corresponding to the fourth omnidirectional perturbation, is the projection of each of the four omnidirectional perturbations in the exact gradient direction, is the initial performance index obtained in step S130, is the Gaussian model standard deviation calculated in step S110; Step S160: At this point, the direction and step size of the gradient descent are determined, and the driving voltage at the current position of the fiber end face is updated to drive the fiber end face to quickly shift to the corresponding iteration point; In step S170 , steps S130 - S160 are repeated, so that the coupling efficiency of the system converges to an optimal value after multiple iterations and can resist disturbances in the closed-loop process to maintain stability.

2. The single-mode optical fiber adaptive coupling method in a turbulent environment according to claim 1, characterized in that: The laser, single-mode optical fiber and collimating lens are used as the emission end in the optical coupling experiment. The laser emission wavelengths include 1550nm, 1064nm and 808nm.

3. The single-mode optical fiber adaptive coupling method in a turbulent environment according to claim 1, characterized in that: The optical fiber coupling device comprises an adaptive optical fiber coupler and a fast reflection mirror.

4. The single-mode optical fiber adaptive coupling method in a turbulent environment according to claim 1, characterized in that: The controller is an integration of high-precision analog-to-digital conversion chip, digital-to-analog conversion chip, FPGA, DSP, or uses an ordinary PC and supporting data acquisition card to achieve closed-loop control.

5. The single-mode optical fiber adaptive coupling method in a turbulent environment according to claim 1, characterized in that: In step S110, the scanning mode includes "bow" shape scanning, "Z" shape scanning, and "circular" shape scanning.

6. The single-mode optical fiber adaptive coupling method in a turbulent environment according to claim 1, characterized in that: Perturbation rate The setting is fixed according to the static scanning result, and the focus spot is disturbed at the center of the fiber end face. The coupling efficiency after disturbance is preferably 90% of the optimal coupling efficiency; the closed-loop frequency f is set based on the lower of the tolerable closed-loop frequency of the optical fiber coupling device and the closed-loop frequency executable by the controller.

7. The single-mode optical fiber adaptive coupling method in a turbulent environment according to claim 1, characterized in that: The photodetector is used to realize linear conversion of optical power to voltage signal. Its working wavelength range covers the wavelength of the signal laser beam, including any one of photodiode, phototransistor, avalanche photodiode and photomultiplier tube.

8. The single-mode optical fiber adaptive coupling method in a turbulent environment according to claim 1, characterized in that: Step S140 includes: the omnidirectional disturbance voltage application process is as follows: first, only apply positive and negative disturbance voltages in the X-axis direction, and no disturbance voltage is applied in the Y-axis direction. The detector obtains the corresponding performance index at this time. ,in, Represents the performance index obtained by applying a positive disturbance voltage in the X-axis direction, Represents the performance indicator acquisition function, Represents the control voltage in the X-axis direction in the t-1 iteration, t is the current iteration number, Represents the magnitude of the disturbance voltage in the X-axis direction, Representatives in The control voltage in the Y-axis direction in the iteration is, Represents the performance index obtained by applying a negative disturbance voltage in the X-axis direction; secondly, the detector obtains the corresponding performance index when positive and negative disturbance voltages are applied only in the Y-axis direction and no disturbance voltage is applied in the X-axis direction. ,in, and Respectively represent the performance indicators obtained by applying positive disturbance voltage and negative disturbance voltage in the Y-axis direction, Represents the magnitude of the disturbance voltage in the Y-axis direction.

9. The single-mode optical fiber adaptive coupling method in a turbulent environment according to claim 8, characterized in that: Step S140 includes: calculating the accurate gradient direction: based on 、 、 、 Calculate the gradient size on the X / Y axis, the X-axis gradient is , the Y-axis gradient is ,in, Prevent the denominator from being 0; Project the voltage corresponding to the small disturbance in the direction of the accurate gradient to obtain the scalar of the four-times disturbance voltage in the projection direction: 、 、 、 .

10. The single-mode optical fiber adaptive coupling method in a turbulent environment according to claim 9, characterized in that: In step S160, the voltage update rule is: , where t is the current iteration number, represents the control voltage in the X-axis direction of the t-th iteration, represents the control voltage in the X-axis direction at the t-1th iteration, represents the gain ratio, and They are the X-axis gradient and the Y-axis gradient, represents the control voltage in the Y-axis direction of the t-th iteration, Represents the control voltage in the Y-axis direction at the t-1th iteration.

Images