Intelligent structure optical communication system based on optical fiber laser coherent array

Abstract

The invention provides an intelligent structure optical communication system based on an optical fiber laser coherent array. An information coding input module is used for inputting code element information; the optical fiber laser coherent synthesis array module generates an array light beam with an orbital angular momentum multiplexing mode corresponding to the code element information according to the code element information; the far-field light spot image data acquisition module is used for acquiring far-field light spot image data of array light beams output by the optical fiber laser coherent array; the spatial pattern recognition module inputs the far-field light spot image data into a trained spatial pattern recognition model to perform orbital angular momentum multiplexing pattern recognition; and the information coding output module outputs code element information corresponding to the orbital angular momentum multiplexing mode based on the identified orbital angular momentum multiplexing mode. According to the invention, the problems of difficult composite mode information extraction, small transmission capacity and low transmission efficiency in the prior art are solved, and the possibility of constructing an intelligent optical communication system by a high-power fiber laser coherent synthesis technology is explored.

Description

Technical Field

[0001] The present invention belongs to the technical field of optical fiber laser coherent synthesis, and in particular relates to an intelligent structured light communication system based on an optical fiber laser coherent array. Background Art

[0002] Coherent array is a special form of phased array system, which can achieve high-power laser output through large-scale array structure, has mature technology in atmospheric turbulence suppression and pointing control, and has certain application potential in long-distance deep space communication. At the same time, fiber laser coherent synthesis technology can generate orbital angular momentum (OAM), a new dimension, by virtue of flexible and controllable aperture arrangement array, which is conducive to expanding transmission capacity and further focusing on the field of long-distance free-space structured optical communication.

[0003] Among them, multiplexing and demultiplexing technology is the key link in free-space structured optical communication. After the OAM light beam is multiplexed and transmitted, it is usually demodulated at the receiving end to obtain the corresponding mode information. At this stage, the mainstream solutions include spiral phase plates, diffractive optical elements, mode conversion methods, phase holograms, etc. However, traditional solutions have certain limitations in the number of mode demodulations, and the system is relatively complex. Machine learning, as a new demultiplexing technology, has advantages in demodulation speed and accuracy, low cost, and can be commercialized.

[0004] Introducing intelligent solutions into the field of fiber laser coherent synthesis technology can further solve the problems of difficulty in extracting composite mode information, small transmission capacity and low transmission efficiency in existing technologies, which is conducive to building strong robustness / high-dimensional communication links and expanding its application potential in the field of free-space structured optical communications. Summary of the invention

[0005] In view of the defects and shortcomings in the prior art, the present invention discloses an intelligent structured light communication system based on a fiber laser coherent array.

[0006] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0007] The present invention provides an intelligent structured light communication system based on a fiber laser coherent array, comprising:

[0008] An information coding input module is used to input code element information;

[0009] A fiber laser coherent synthesis array module generates an array beam having an orbital angular momentum multiplexing pattern corresponding to the code element information according to the code element information;

[0010] A far-field spot image data acquisition module is used to acquire far-field spot image data of the array beam output by the fiber laser coherent array;

[0011] The spatial pattern recognition module inputs the far-field spot image data into the trained spatial pattern recognition model to perform orbital angular momentum multiplexing pattern recognition;

[0012] The information encoding output module is based on the identified orbital angular momentum multiplexing mode and outputs codeword information corresponding to the orbital angular momentum multiplexing mode.

[0013] Preferably, the fiber laser coherent synthesis array module includes a fiber laser coherent synthesis optical path, and the fiber laser coherent synthesis optical path includes a seed source, a pre-amplifier, a fiber beam splitter, a phase modulator, a cascade fiber amplifier, and an adaptive fiber collimator array. The fiber beam splitter has multiple output ends. The seed laser output by the seed source is amplified by the pre-amplifier and then divided into multiple unit light beams by the fiber beam splitter. Each output end of the fiber beam splitter is respectively connected to a phase modulator, a cascade fiber amplifier, and an adaptive fiber collimator. Multiple adaptive fiber collimators are arranged in a circular or regular polygonal array to form an adaptive fiber collimator array. The adaptive fiber collimator array includes m layers of adaptive fiber collimator annular sub-arrays from the inside to the outside. The number m of the annular sub-array ring layers determines that the fiber laser coherent synthesis array module can generate 2 m An array beam with an orbital angular momentum multiplexing mode.

[0014] Preferably, the training method of the spatial pattern recognition model comprises:

[0015] Get 2 m Far-field spot image data corresponding to the array beam in the orbital angular momentum multiplexing mode and the corresponding phase information and amplitude information;

[0016] 2 m The pixel information of the far-field spot image data corresponding to the array beam of the orbital angular momentum multiplexing mode is encoded according to a certain encoding rule to obtain the code element information corresponding to the various orbital angular momentum multiplexing modes;

[0017] The far-field spot image data and the corresponding orbital angular momentum multiplexing mode codeword information are used as training data, and the loss function is constructed as follows:

[0018]

[0019] The true label k real ={0,1,2,...,2 m -1}, indicating the symbol information of various orbital angular momentum multiplexing modes, k pred represents the predicted label; J(k real ,kpred ) represents the difference function, which is used to compare the true label k real and predicted label k pred The difference between is the regularization term, γ represents the regularization parameter, is the optimization function related to θ;

[0020] When the loss function is minimized or reaches the set threshold, the model converges, and the model parameters θ that make the model converge are saved, the model training is completed, and a trained spatial pattern recognition model is obtained.

[0021] Preferably, the fiber laser coherent synthesis optical path controls the amplitude and phase of each unit light beam in the fiber laser coherent synthesis optical path according to the code element information, and generates an array light beam in an orbital angular momentum multiplexing mode corresponding to the code element information after coherent superposition.

[0022] Preferably, the fiber laser coherent synthesis array module includes a phase control unit, which performs signal processing by running a phase optimization control algorithm to generate a phase control signal corresponding to each unit light beam, and transmits it to the corresponding phase modulator to achieve phase control of each unit light beam.

[0023] Preferably, the present invention also includes a beam splitter, a focusing lens, and a photoelectric detection module. The array light beam output from the emission plane of the adaptive fiber collimator array is divided into two parts by the beam splitter, one part of which is transmitted to the far-field spot image data acquisition module through the focusing lens, and the other part is received by the photoelectric detection module and converted into an electrical signal and fed back to the phase control unit.

[0024] The phase optimization control algorithm described in the present invention is not limited, including but not limited to the random parallel gradient descent algorithm, simulated annealing algorithm or particle swarm optimization algorithm commonly used in the art.

[0025] Preferably, the array beam of the present invention comprises m layers of annular subarrays from the inside to the outside, each annular subarray is composed of unit beams uniformly distributed along the angular direction, the waist radius of each unit beam on each annular subarray is w0, the wavelength is λ, the beam aperture is d, the emission amplitude is A0, and the distance between the center of each unit beam on the i-th annular subarray and the center of the array beam is R i , i = 1, 2, 3, ..., m, wherein the mth annular subarray is the outermost annular subarray farthest from the center of the array beam. The specific form of the array beam is not limited, and can be a circular array beam or other regular polygonal array beams.

[0026] Based on the above technical solution, the technical effects that can be produced by the present invention are:

[0027] The present invention trains a spatial pattern recognition model and uses the trained spatial pattern recognition model for recognizing the composite pattern of a fiber laser coherent array, thereby exploring the feasibility of high-power fiber laser coherent synthesis technology in intelligent optical communication applications.

[0028] The trained spatial pattern recognition model can be used to accurately identify composite patterns in the presence of system errors and external interference, providing a technical basis for establishing a highly robust communication link based on fiber laser coherent arrays.

[0029] The present invention can recognize high-dimensional coding patterns of large-scale coherent arrays using a trained spatial pattern recognition model, which is beneficial to further improve communication efficiency and communication capacity. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying any creative work.

[0031] Figure 1 It is a schematic structural diagram of an intelligent structured optical communication system based on a fiber laser coherent array according to an embodiment;

[0032] Figure 2 is a schematic diagram of a network structure of a spatial pattern recognition model in one embodiment;

[0033] Figure 3 This is a schematic diagram of a pattern recognition effect based on hexadecimal encoding in one embodiment;

[0034] Figure 4 This is a schematic diagram of image transmission based on hexadecimal encoding in one embodiment;

[0035] Numbers in the figure:

[0036] 1. Information encoding input module; 2. Seed source; 3. Pre-amplifier; 4. Fiber beam splitter; 5. Phase modulator; 6. Cascade fiber amplifier; 7. Adaptive fiber collimator array; 8. Beam splitter; 9. Focusing lens; 10. Far-field spot image data acquisition module; 11. Spatial pattern recognition module; 12. Photoelectric detection module; 13. Phase control unit. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] One embodiment provides an intelligent structured light communication system based on a fiber laser coherent array, comprising:

[0039] Information coding input module 1, used for inputting code element information;

[0040] A fiber laser coherent synthesis array module generates an array beam having an orbital angular momentum multiplexing pattern corresponding to the code element information according to the code element information;

[0041] A far-field spot image data acquisition module is used to acquire far-field spot image data of the array beam output by the fiber laser coherent array;

[0042] The spatial pattern recognition module inputs the far-field spot image data into the trained spatial pattern recognition model to perform orbital angular momentum multiplexing pattern recognition;

[0043] The information encoding output module is based on the identified orbital angular momentum multiplexing mode and outputs codeword information corresponding to the orbital angular momentum multiplexing mode.

[0044] Reference Figure 1 , Figure 1 The structure diagram of an intelligent structured light communication system based on a fiber laser coherent array of an embodiment includes an information coding input module 1, a fiber laser coherent synthesis array module, a far-field spot image data acquisition module 10, a spatial pattern recognition module 11, and an information coding output module. The fiber laser coherent synthesis optical path controls the amplitude and phase of each unit light beam in the fiber laser coherent synthesis optical path according to the code element information of the information coding input module 1, and generates an array light beam of an orbital angular momentum multiplexing mode corresponding to the code element information after coherent superposition.

[0045] The fiber laser coherent synthesis optical path in the fiber laser coherent synthesis array module includes a seed source 2, a pre-amplifier 3, a fiber beam splitter 4, a phase modulator 5, a cascade fiber amplifier 6, and an adaptive fiber collimator array 7. The fiber beam splitter 4 has multiple output ends. The seed laser output by the seed source 2 is amplified by the pre-amplifier 3 and then divided into multiple unit light beams after passing through the fiber beam splitter 4. Each output end of the fiber beam splitter 4 is respectively connected to a phase modulator 5, a cascade fiber amplifier 6, and an adaptive fiber collimator. Multiple adaptive fiber collimators are arranged in a circular or regular polygonal array to form an adaptive fiber collimator array 7. The adaptive fiber collimator array 7 includes m layers of adaptive fiber collimator annular sub-arrays from the inside to the outside. The number m of the annular sub-array ring layers determines that the fiber laser coherent synthesis array module can generate 2 m An array beam with an orbital angular momentum multiplexing mode.

[0046] The array beam includes m layers of annular subarrays from the inside to the outside. Each annular subarray is composed of unit beams uniformly distributed along the angular direction. The waist radius of each unit beam on each annular subarray is w0, the wavelength is λ, the beam diameter is d, the output amplitude is A0, and the distance between the center of each unit beam on the i-th annular subarray and the center of the array beam is R i , i = 1, 2, 3, ..., m, wherein the mth annular subarray is the outermost annular subarray farthest from the center of the array beam. The specific form of the array beam is not limited, and can be a circular array beam or other regular polygonal array beams.

[0047] The complex amplitude distribution of the array beam on the emission surface is expressed as:

[0048]

[0049] Where (x, y) is the coordinate on the emission surface, N i is the number of unit beams contained in the i-th layer of the annular subarray, a total of N i beam unit beam; (a i,j , b i,j ) is the center coordinate of the jth unit beam on the i-th ring subarray, satisfying

[0050] is the piston phase of the j-th unit beam on the i-th ring subarray, satisfying: l i is the orbital angular momentum of the vortex beam generated by the i-th ring subarray, l1,l2,···,l n None of them are equal to each other;

[0051] The complex amplitude distribution of the target plane in the spatial pattern recognition module is: E(x',y') = F{E(x,y)}; then the light intensity image obtained by the far-field spot image data acquisition module is: I = E(x',y') × E(x',y') * .

[0052] The array beam output from the emission plane of the adaptive fiber collimator array 7 is divided into two parts by a beam splitter 8, one part of which is transmitted to the far-field spot image data acquisition module 10 through a focusing lens 9, and the other part is received by the photoelectric detection module 12 and converted into an electrical signal and fed back to the phase control unit 13. The phase control unit 13 performs signal processing by running a phase optimization control algorithm, generates a phase control signal corresponding to each unit beam, and transmits it to the corresponding phase modulator 5 to achieve phase control of each unit beam. The phase optimization control algorithm is not limited, including but not limited to the random parallel gradient descent algorithm, simulated annealing algorithm or particle swarm optimization algorithm commonly used in the art.

[0053] The network structure of the spatial pattern recognition model in the present invention is not limited, and a convolutional neural network, a feedforward neural network or a deep diffraction neural network can be used. Figure 2 This is a schematic diagram of the network structure of a spatial pattern recognition model in an embodiment, and the ResNet-18 deep convolutional neural network structure used includes a series of convolutional layers.

[0054] Furthermore, an embodiment provides a method for training the spatial pattern recognition model, including:

[0055] Get 2 m Far-field spot image data corresponding to the array beam in the orbital angular momentum multiplexing mode and the corresponding phase information and amplitude information;

[0056] 2 m The pixel information of the far-field spot image data corresponding to the array beam of the orbital angular momentum multiplexing mode is encoded according to a certain encoding rule to obtain the code element information corresponding to the various orbital angular momentum multiplexing modes. In this way, the phase information and amplitude information of the array beam of the orbital angular momentum multiplexing mode are determined by the code element information.

[0057] The far-field spot image data and the corresponding orbital angular momentum multiplexing mode codeword information are used as training data, and the loss function is constructed as follows:

[0058]

[0059] The true label k real ={0,1,2,...,2 m -1}, indicating the symbol information of various orbital angular momentum multiplexing modes, k pred represents the predicted label; J(kreal ,k pred ) represents the difference function, which is used to compare the true label k real and predicted label k pred The difference between is the regularization term, γ represents the regularization parameter, is the optimization function related to θ;

[0060] When the loss function is minimized or reaches the set threshold, the model converges, and the model parameters θ that make the model converge are saved, the model training is completed, and a trained spatial pattern recognition model is obtained.

[0061] In a specific application example of the intelligent structured optical communication system based on the fiber laser coherent array, first of all, considering the realization of hexadecimal encoding, a circular ring subarray with 4 ring layers is required to participate in the mode multiplexing process. The circular ring subarrays of the 2nd, 4th, 6th and 8th layers of the array beam are selected as the orbital angular momentum mode information carrying units. The complex amplitude distribution of the circular array beam on the emission surface is expressed as:

[0062]

[0063] Wherein, (x, y) is the coordinate on the emission surface, and the numbers of unit beams included in the 2nd, 4th, 6th and 8th layers of the circular ring subarray are 12, 24, 36 and 48 respectively; (a i,j , b i,j ) is the center coordinate of the jth unit beam on the i-th layer of the circular subarray, satisfying is the piston phase of the j-th unit beam on the i-th ring subarray, satisfying: l i is the orbital angular momentum of the vortex beam generated by the designated i-th layer of the circular ring subarray, l1,l2,···,l n They are not equal to each other; among them, l2=1, l4=3, l6=5, l8=7.

[0064] The complex amplitude distribution of the target plane in the spatial pattern recognition module is: E(x',y') = F{E(x,y)}; then the light intensity image obtained by the far-field spot image data acquisition module is: I = E(x',y') × E(x',y') * .

[0065] refer to Figure 3 In this example, the array light field is used to generate Figure 3 (a) shows a total of 16 modes, including the case of no light intensity. At the same time, the system is used to collect 16,000 spot images in real time, that is, a single code collects 1,000 spot images, and 950 spot images are selected as the training set, and the other 50 spot images are selected as the test set. Figure 3(b) shows the confusion matrix showing the effect of reuse pattern recognition in the test set, which shows that the predicted labels are consistent with the actual labels. Figure 3 (c) shows the accuracy curve of the network after 25 iterations of training. It can be seen that the curve finally reaches a convergence state.

[0066] refer to Figure 4 , is a schematic diagram of image transmission based on hexadecimal encoding in the embodiment. After encoding the image data to be transmitted, the code element information is obtained, and the pattern multiplexing information carrying the code element information is transmitted to the far-field spot image data acquisition module, and input into the trained spatial pattern recognition model obtained above for recognition. Finally, the code element information is obtained through data processing, and the code element information is decoded to output the image data. The results show that the accuracy of demodulating the pattern information using the intelligent algorithm can reach 99.96%, and the bit error rate of the keyed communication link is 1.11×10 -5 .

[0067] Matters not covered by the present invention are known technologies.

[0068] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0069] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

[0070] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. Intelligent structured light communication system based on fiber laser coherent array, characterized in that: include: An information coding input module is used to input code element information; A fiber laser coherent synthesis array module generates an array beam having an orbital angular momentum multiplexing pattern corresponding to the code element information according to the code element information; A far-field spot image data acquisition module is used to acquire far-field spot image data of the array beam output by the fiber laser coherent array; The spatial pattern recognition module inputs the far-field spot image data into the trained spatial pattern recognition model to perform orbital angular momentum multiplexing pattern recognition; The information encoding output module is based on the identified orbital angular momentum multiplexing mode and outputs codeword information corresponding to the orbital angular momentum multiplexing mode.

2. The intelligent structured light communication system based on fiber laser coherent array according to claim 1, characterized in that: The fiber laser coherent synthesis array module includes a fiber laser coherent synthesis optical path, which includes a seed source, a pre-amplifier, a fiber beam splitter, a phase modulator, a cascade fiber amplifier, and an adaptive fiber collimator array. The fiber beam splitter has multiple output ends. The seed laser output by the seed source is amplified by the pre-amplifier and then divided into multiple unit light beams by the fiber beam splitter. Each output end of the fiber beam splitter is respectively connected to a phase modulator, a cascade fiber amplifier, and an adaptive fiber collimator. Multiple adaptive fiber collimators are arranged in a circular or regular polygonal array to form an adaptive fiber collimator array. The adaptive fiber collimator array includes m layers of adaptive fiber collimator annular sub-arrays from the inside to the outside. The number of annular sub-array circles m determines that the fiber laser coherent synthesis array module can generate 2 m An array beam with an orbital angular momentum multiplexing mode.

3. The intelligent structured light communication system based on fiber laser coherent array according to claim 2, characterized in that: The training method of the spatial pattern recognition model comprises: Get 2 m Far-field spot image data corresponding to the array beam in the orbital angular momentum multiplexing mode and the corresponding phase information and amplitude information; 2 m The pixel information of the far-field spot image data corresponding to the array beam of the orbital angular momentum multiplexing mode is encoded according to a certain encoding rule to obtain the code element information corresponding to the various orbital angular momentum multiplexing modes; The far-field spot image data and the corresponding orbital angular momentum multiplexing mode codeword information are used as training data, and the loss function is constructed as follows: The true label k real ={0,1,2,...,2 m -1}, indicating the symbol information of various orbital angular momentum multiplexing modes, k pred represents the predicted label; J(k real ,k pred ) represents the difference function, which is used to compare the true label k real and predicted label k pred The difference between is the regularization term, γ represents the regularization parameter, is the optimization function related to θ; When the loss function is minimized or reaches the set threshold, the model converges, and the model parameters θ that make the model converge are saved, the model training is completed, and a trained spatial pattern recognition model is obtained.

4. The intelligent structured light communication system based on fiber laser coherent array according to claim 3, characterized in that: The fiber laser coherent synthesis optical path controls the amplitude and phase of each unit light beam in the fiber laser coherent synthesis optical path according to the code element information, and generates an array light beam in an orbital angular momentum multiplexing mode corresponding to the code element information after coherent superposition.

5. The intelligent structured light communication system based on fiber laser coherent array according to claim 4, characterized in that: The fiber laser coherent synthesis array module includes a phase control unit, which performs signal processing by running a phase optimization control algorithm to generate a phase control signal corresponding to each unit light beam and transmits it to the corresponding phase modulator to achieve phase control of each unit light beam.

6. The intelligent structured light communication system based on fiber laser coherent array according to claim 5, characterized in that: It also includes a beam splitter, a focusing lens, and a photoelectric detection module. The array light beam output from the emission plane of the adaptive fiber collimator array is divided into two parts by the beam splitter, one part of which is transmitted to the far-field spot image data acquisition module through the focusing lens, and the other part is received by the photoelectric detection module and converted into an electrical signal and fed back to the phase control unit.

7. The intelligent structured light communication system based on fiber laser coherent array according to claim 5, characterized in that: The phase optimization control algorithm is a random parallel gradient descent algorithm, a simulated annealing algorithm or a particle swarm optimization algorithm.

8. The intelligent structured light communication system based on fiber laser coherent array according to claim 2, characterized in that: The array beam contains m layers of annular subarrays from the inside to the outside. Each annular subarray is composed of unit beams uniformly distributed along the angular direction. The waist radius of each unit beam on each annular subarray is w0, the wavelength is λ, the beam diameter is d, the output amplitude is A0, and the distance between the center of each unit beam on the i-th annular subarray and the center of the array beam is R i , i = 1, 2, 3, ..., m, where the mth annular sub-array is the outermost annular sub-array farthest from the center of the array beam.

Images