A mode group demultiplexing method and device based on cascade diffraction network

Abstract

The present invention relates to the technical field of multiplexing and demultiplexing devices in space division multiplexing, and more specifically, to a mode group demultiplexing method and device based on a cascade diffraction network. The mode group demultiplexing method adopts a strategy of non-deterministic mode conversion to design and optimize the cascade diffraction network, which improves the freedom of the output mode field, so that the phase plate design can be carried out with greater freedom, thereby significantly improving the efficiency of beam conversion and greatly reducing the number of phase plates required. The present invention overcomes the shortcomings of low energy efficiency and stability in mode group demultiplexing, and reduces device loss by improving and optimizing the design.

Description

Technical Field

[0001] The present invention relates to the technical field of multiplexing and demultiplexing devices in space division multiplexing, and more specifically, to a mode group demultiplexing method and device based on a cascaded diffraction network. Background Art

[0002] With the rapid development of technologies such as the Internet, the Internet of Things, and 5G communications, the global demand for data transmission has exploded. With the maturity of technologies such as time division multiplexing, wavelength division multiplexing, and high-order modulation, the capacity of single-mode optical fiber has gradually approached its theoretical limit. One upgrade solution is to develop the spatial dimension of optical fiber and use the orthogonality of spatial modes to expand the capacity of a single optical fiber, also known as mode division multiplexing. The disadvantage of mode division multiplexing is that it is greatly affected by mode coupling and heavily relies on partial or complete multiple-input and multiple-output digital signal processing. Therefore, the high cost of short-distance optical interconnection in this way is usually unacceptable. To solve this problem, a mode group multiplexing technology based on weakly coupled optical fibers has been proposed, which enables intensity modulation and direct detection technology to be used in short-distance scenarios. Since the group delays of multiple modes in the same mode group are almost equal, these modes can be detected simultaneously and regarded as a single spatial channel. Although this method sacrifices some mode degrees of freedom, it can greatly reduce the complexity of the system.

[0003] Compared with module multiplexing technology, research on module demultiplexing devices is relatively lacking. Since multiple modes in the same module are strongly coupled, module multiplexing only requires exciting one of the modes. In this way, a mapping of multiple single-mode fibers to multiple mode groups can be established. However, the module multiplexer cannot be simply used in reverse for module demultiplexing, because the module contains multiple mutually coupled modes, and conversion to a single-mode fiber will inevitably cause energy loss in some modes.

[0004] For module demultiplexing, a common approach is to receive a mode within the module and concentrate the module energy on this mode by controlling the crosstalk of the fiber mode. Another solution is to demodulate all modes of the module to multiple single-mode optical fibers, and then combine the multiple single-mode optical fibers into the same optical fiber. The disadvantage of the above-mentioned fiber module demultiplexing solution is that it is impossible to achieve efficient and stable separation of fiber modules. The energy of a single mode in the receiving module will fluctuate with the mode coupling in the module, which will cause instability of the signal at the receiving end. The step-by-step approach of demodulating all modes of the module and then combining them is redundant and prone to additional losses and crosstalk. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of low energy efficiency and low stability in the mode multiplexing of the prior art, and to provide a mode group demultiplexing method and device based on a cascaded diffraction network, which effectively improves the energy efficiency and stability in the mode multiplexing process.

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

[0007] A mode group demultiplexing method based on a cascade diffraction network is provided, comprising the following steps:

[0008] S1. Initialize phase plate parameters;

[0009] S2. Select input mode E 0,m : According to the input weakly coupled optical fiber, the mode field distribution of the selected input mode is set as the input light field;

[0010] S3. Calculate the output mode field of the cascade diffraction: According to the selected input mode E 0,m Calculation of the output mode field E of the cascade diffraction network N+1,m ;

[0011] S4. Calculate the output light field E N+1,m Projection on the substrate: Output mode field E N+1,m Projecting onto the mode basis of the corresponding channel of the output fiber array;

[0012] S5. Calculate module transfer matrix T i,j :Calculate the energy of the output mode field coupled to each channel of the output fiber array. The coupling efficiency is the sum of the corresponding mode base components and is calculated through the transmission matrix T i,j express; T i,j Represents the input terminal i The module sends, and the output end is j The power received by the module channel;

[0013] S6. According to the transmission matrix T i,j Calculate the evaluation function;

[0014] S7. Error gradient back propagation: Determine the phase plate thickness based on the error back propagation method H k The gradient of

[0015] S8. Gradient descent parameter correction: further iteratively updating a new set of phase plate parameters according to the gradient information in step S7 combined with the gradient descent method;

[0016] S9. Calculate the average loss and crosstalk of the mode group demultiplexing device;

[0017] S10. Determine whether the loss crosstalk meets the standard. If so, the optimization ends and the phase plate distribution is output; if not, return to step S2.

[0018] The present invention provides a mode group demultiplexing method based on a cascade diffraction network, which adopts a non-deterministic mode conversion strategy to design and optimize the cascade diffraction network, improves the degree of freedom of the output mode field, and enables the phase plate design to be performed with greater degrees of freedom, thereby significantly improving the efficiency of beam conversion and greatly reducing the number of phase plates required. The present invention overcomes the shortcomings of low energy efficiency and stability in mode group demultiplexing, and reduces device loss by improving and optimizing the design.

[0019] Preferably, in step S3, the output mode field of the cascade diffraction network E N+1,m Calculated by the following formula:

[0020]

[0021] In the formula, N represents the number of phase plates; E 0,m represents the mth input mode of the input weakly coupled fiber; E N+1,m represents the corresponding final output mode field; M k represents the kth phase plate P k The control of the light field, k=1,2,3,...,N, is a diagonal matrix with exp(jPk) as the diagonal element; F k represents the transmission matrix corresponding to the k-th diffraction.

[0022] Preferably, in step S4, the output mode field is calculated by the following formula: E N+1,m Projection onto the substrate:

[0023]

[0024] In the formula, c m,n,r represents the projection coefficient, B n,r The output fiber array n The rth mode basis of the receiving channel; S Indicates the number of mode bases supported by each fiber of the fiber array.

[0025] Preferably, in step S5, the transmission matrix T i,j Calculated by the following formula:

[0026]

[0027] Where, MG i Indicates the first i The set of all patterns in a pattern group, |MG i | is the input if the coupling fiber i The number of patterns contained in a pattern group.

[0028] Preferably, in step S6, according to the transmission matrix T, a scalar evaluation function is defined: L , to inversely optimize the phase plates of the cascade diffraction network P k , the evaluation function L is expressed as follows:

[0029]

[0030] In the formula, L It reflects the loss and crosstalk conditions. a and b are used to control the weights of loss and crosstalk respectively. Q Indicates the number of channels of the fiber array; L Determined by the phase distribution of the phase plate, which is based on the phase plate parameter P k is a multivariate function of the variables.

[0031] Preferably, in step S7, in order to facilitate actual processing, the phase of the phase plate is replaced by the actual thickness distribution H k , taking the reflective phase plate as an example, the thickness of the phase plate is H k With Phase Plate P k Relationship satisfaction P k =4π H k / λ, λ is the working wavelength, and the thickness of the phase plate is determined based on the error back propagation method. H k The gradient of is expressed as follows:

[0032]

[0033] In the formula, Im means taking the imaginary part of the operation result, and diag means taking the diagonal elements of the matrix.

[0034] Preferably, in step S8, the updating of the phase plate parameters is achieved by the following formula:

[0035]

[0036] In the formula, α represents the learning rate, which is used to control the amplitude of parameter update. α can be set to a fixed value or can be adaptively determined by a specific algorithm during the iterative optimization process of parameters.

[0037] Preferably, in step S9, the average loss IL of the device ave and Crosstalk XT ave Calculated by the following formula:

[0038] .

[0039] The above parameter updating process can be terminated until the design loss and crosstalk reach the predetermined index, and the corresponding phase plate distribution is output.

[0040] Preferably, the phase plate can be further processed as an actual diffractive optical element through grayscale exposure or binary photolithography, and a reflection-enhancing film can be evaporated.

[0041] Preferably, the phase plate is processed by binary processing, firstly the phase plate distribution is discretized into 2 V steps, and then binarized as follows:

[0042]

[0043] Where b 0 ,b 1 ,…,b V There are V binary lithography patterns, and the etching depth corresponding to each binary pattern is λ / 2 V+1 ,λ / 2 V ,λ / 2 2 , to achieve 2π phase control at wavelength λ.

[0044] The present invention also provides a mode group demultiplexing device based on a cascaded diffraction network, comprising an input weakly coupled optical fiber, a cascaded diffraction network and an output optical fiber array, wherein the input weakly coupled optical fiber comprises Q modules to be demultiplexed, each module comprising a number of modes; the output optical fiber array comprises Q channels, each optical fiber of the output optical fiber array supports S mode bases, and the number of mode bases is not less than the number of modes in the input weakly coupled optical fiber module; the cascaded diffraction network comprises a number of spatially separated phase plates, wherein the working spaces of the phase plates are between the input weakly coupled optical fiber and the output optical fiber array, and are used to achieve spatial separation of the modules in the input weakly coupled optical fiber, ensuring that the module energy can be maximally coupled to the corresponding output optical fiber array channel; each channel of the output optical fiber array can receive the output of the cascaded diffraction layer with any one of the corresponding mode bases or any linear superposition, and the target output of the cascaded diffraction network is any linear combination of the S mode bases; wherein the mode group demultiplexing method based on the cascaded diffraction network described above is used to achieve the optimal design of the cascaded diffraction network.

[0045] Preferably, the input weakly coupled optical fiber is a few-mode optical fiber or a multi-mode optical fiber, comprising Q modules to be demultiplexed, each module being composed of a number of modes, including but not limited to orbital angular momentum modes, linear polarization modes, Laguerre-Gaussian modes, and Hermitian-Gaussian modes.

[0046] Preferably, the output fiber array comprises Q channels, corresponding to the Q modules to be demultiplexed in the input weakly coupled fiber. The output fiber array is a few-mode fiber array or a multimode fiber array, each fiber supports S mode bases, and multiple fibers can remain consistent and satisfy that the number of mode bases S is not less than the number of modes in the input weakly coupled fiber module. Each channel of the output fiber array can receive the output of the cascade diffraction network with any one of its corresponding mode bases or any linear superposition thereof.

[0047] Preferably, the cascade diffraction network is composed of a number of spatially separated phase plates, which can be transmissive or reflective, and whose working space is between the input weakly coupled fiber and the output fiber array. Its function is to achieve spatial separation of modules in the input weakly coupled fiber and ensure that the module energy can be maximally coupled to the corresponding output fiber array channel; the target output of the cascade diffraction network is not a certain mode, but an arbitrary linear combination of the corresponding channel mode basis of the output fiber array.

[0048] Preferably, the input weakly coupled optical fiber, the cascade diffraction network, and the output optical fiber array can be fixed to the same substrate by ultraviolet glue after preparation to form a stably packaged module demultiplexer.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] The present invention discloses a mode group demultiplexing method and device based on a cascade diffraction network, which adopts a non-deterministic mode conversion strategy to design and optimize the cascade diffraction network, thereby improving the degree of freedom of the output mode field, allowing the phase plate design to be performed with greater degrees of freedom, thereby significantly improving the efficiency of beam conversion and greatly reducing the number of phase plates required. The present invention overcomes the shortcomings of low energy efficiency and stability in mode group demultiplexing, and reduces device loss by improving and optimizing the design. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A schematic diagram of a flow chart of a mode group demultiplexing method based on a cascade diffraction network;

[0052] Figure 2 It is a schematic diagram of the framework structure of a mode group demultiplexing device based on a cascade diffraction network;

[0053] Figure 3 A schematic diagram of a phase diagram design for orbital angular momentum mode reuse in Example 3;

[0054] Figure 4 Schematic diagram of intensity distribution of each mode of the few-mode optical fiber in Example 3;

[0055] Figure 5 This is the corresponding output mode field result after the orbital angular momentum mode in Example 3 passes through the cascade diffraction network.

[0056] In the attached figure: 1. Input weakly coupled optical fiber; 2. Cascade diffraction network; 3. Optical fiber array. DETAILED DESCRIPTION

[0057] The present invention is further described below in conjunction with specific implementation methods. The accompanying drawings are only for illustrative purposes and are only schematic diagrams, not actual drawings, and should not be construed as limiting the present invention. In order to better illustrate the embodiments of the present invention, some parts of the accompanying drawings may be omitted, enlarged, or reduced, and do not represent the size of the actual product. It is understandable to those skilled in the art that some well-known structures and their descriptions in the accompanying drawings may be omitted.

[0058] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right" and the like indicate directions or positional relationships based on the directions or positional relationships shown in the drawings, it is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes and cannot be understood as limitations on the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0059] Embodiment 1

[0060] This embodiment is the first embodiment of the mode group demultiplexing method based on the cascade diffraction network 2. Figure 1 As shown, the following steps are included:

[0061] Step 1. First, a cascade diffraction transmission model needs to be constructed to optimize the multiphase plate of the cascade diffraction network 2. Given the input mode of the weakly coupled fiber 1, the output of the cascade diffraction network 2 is as follows by Rayleigh-Sommerfeld diffraction:

[0062]

[0063] In the formula, N represents the number of phase plates; E 0,m represents the mth input mode of the input weakly coupled fiber 1; E N+1,mrepresents the corresponding final output mode field; M k represents the kth phase plate P k The control of the light field, k=1,2,3,...,N, is a diagonal matrix with exp(jP,k) as the diagonal element; F k represents the transmission matrix corresponding to the k-th diffraction, which depends on the starting position of the k-th diffraction space.

[0064] Step 2. After obtaining the output of cascade diffraction network 2, there is no need to compare it with the specific target mode field. E N+1,m Projected onto the mode basis of the corresponding channel of fiber array 3:

[0065]

[0066] In the formula, c m,n,r represents the projection coefficient, B n,r The output fiber array 3 n The receiving channel r A pattern base; S Indicates the number of mode bases supported by each optical fiber of the optical fiber array 3. The receiving optical fibers of different modules are different, so the corresponding mode bases are different.

[0067] Step 3. Calculate the energy of the output mode field coupled to each channel of the output fiber array 3. The coupling efficiency is the sum of the corresponding mode base components and can be described by the following transmission matrix T:

[0068]

[0069] Where, MG i Indicates the input weakly coupled fiber 1 i The set of all patterns in a pattern group, |MG i | is the first i The number of patterns contained in each pattern group; T i,j Represents the input terminal i The module sends, and the output end is j The power received by the module channel.

[0070] Step 4. To inversely optimize the phase of each phase plate of the cascade diffraction network 2 P k , according to the transfer matrix T, the scalar evaluation function is defined as follows:

[0071]

[0072] In the formula,L It reflects the loss and crosstalk conditions. a and b are used to control the weights of loss and crosstalk respectively. Q represents the number of channels of the optical fiber array 3; L Determined by the phase distribution of the phase plate, which is based on the phase plate parameter P k is a multivariate function of the variables.

[0073] Step 5. To facilitate actual processing, replace the phase plate phase with the actual thickness distribution H k , taking the reflective phase plate as an example, the thickness of the phase plate is H k With Phase Plate P k Relationship satisfaction P k =4π H k / λ, λ is the working wavelength, and the thickness of the phase plate is determined based on the error back propagation method. H k The gradient of is expressed as follows:

[0074]

[0075] In the formula, Im means taking the imaginary part of the operation result, and diag means taking the diagonal elements of the matrix.

[0076] Step 6. The above gradient information indicates the optimization direction, and then the gradient descent method is combined to further iteratively update a new set of phase plate parameters as follows:

[0077]

[0078] In the formula, α represents the learning rate, which is used to control the amplitude of parameter update. α can be set to a fixed value or can be adaptively determined by a specific algorithm during the iterative optimization process of parameters.

[0079] Step 7. Calculate the average loss IL of the mode group demultiplexing device ave and Crosstalk XT ave as follows:

[0080]

[0081] The above parameter updating process can be terminated until the design loss and crosstalk reach the predetermined index, and the corresponding phase plate distribution is output.

[0082] In this embodiment, the phase plate parameters of the cascade diffraction network 2 can be further processed into an actual diffractive optical element through grayscale exposure or binary lithography, and a reflection enhancement film can be evaporated.

[0083] Binary processing of the phase plate can discretize the phase plate distribution into 2 V steps, and then binarized as follows:

[0084]

[0085] Where b 0 ,b 1 ,…,b V There are V binary lithography patterns, and the etching depth corresponding to each binary pattern is λ / 2 V+1 ,λ / 2 V ,λ / 2 2 , to achieve 2π phase control at wavelength λ.

[0086] Although the cascade diffraction network 2 has been previously proven to be useful for multiplexing and demultiplexing of various modes, the original solution cannot be directly applied to module demultiplexing. A key difference is that module demultiplexing must demultiplex multiple modes of the same module to the same spatial position, and these modes can then be received by the same detector. Traditional single-mode optical fibers cannot collect the mode fields of multiple modes at the same time, otherwise it will cause energy fluctuations. For this reason, multimode optical fibers can be used for reception. In this embodiment, taking into account the multi-mode freedom characteristics of multi-mode optical fibers, a non-deterministic mode conversion strategy is adopted to design and optimize the cascade diffraction network 2. The benefit of adopting this strategy is that it increases the freedom of the output mode field, allowing the phase plate design to be carried out with greater freedom. This can significantly improve the efficiency of beam conversion and greatly reduce the number of phase plates required.

[0087] Embodiment 2

[0088] The present embodiment is an embodiment of a mode group demultiplexing device based on a cascaded diffraction network 2. In the present embodiment, the mode group demultiplexing device includes an input weakly coupled fiber, a cascaded diffraction network 2 and an output fiber array 3. The input weakly coupled fiber includes Q modules to be demultiplexed, and each module includes a number of modes; the output fiber array 3 includes Q channels, and each fiber of the output fiber array 3 supports S mode bases, and the number of mode bases is not less than the number of modes in the input weakly coupled fiber module; the cascaded diffraction network 2 includes a number of spatially separated phase plates, and the working spaces of the phase plates are between the input weakly coupled fiber 1 and the output fiber array 3, which are used to achieve spatial separation of the modules in the input weakly coupled fiber 1, ensuring that the module energy can be maximally coupled to the corresponding output fiber array 3 channel; each channel of the output fiber array 3 can receive the output of the cascaded diffraction layer with any one of the corresponding mode bases or any linear superposition, and the target output of the cascaded diffraction network 2 is any linear combination of the S mode bases.

[0089] In this embodiment, the input weakly coupled optical fiber 1 is used as the light source input, including a few-mode optical fiber or a multi-mode optical fiber, and contains Q modules to be demultiplexed, each of which is composed of a number of modes, including but not limited to orbital angular momentum mode, linear polarization mode, Laguerre Gaussian mode, and Hermitian Gaussian mode.

[0090] In this embodiment, the output fiber array 3 includes Q channels, corresponding to the Q modules to be demultiplexed in the input weakly coupled fiber 1. The output fiber array 3 is a few-mode fiber array 3 or a multi-mode fiber array 3, each fiber supports S mode bases, and multiple fibers can remain consistent and satisfy that the number of mode bases S is not less than the number of modes in the module of the input weakly coupled fiber 1. Each channel of the output fiber array 3 can receive the output of the cascade diffraction network 2 with any one of its corresponding mode bases or any linear superposition thereof.

[0091] In this embodiment, the cascade diffraction network 2 is composed of a number of spatially separated phase plates, which can be transmissive or reflective, and its working space is between the input weakly coupled optical fiber 1 and the output optical fiber array 3. Its function is to achieve spatial separation of the modules in the input weakly coupled optical fiber 1, ensuring that the module energy can be maximally coupled to the corresponding output optical fiber array 3 channel; the target output of the cascade diffraction network 2 is not a certain mode, but an arbitrary linear combination of the corresponding channel mode basis of the output optical fiber array 3.

[0092] After constructing the above-mentioned mode group demultiplexing framework, it is necessary to further optimize the design of the cascade diffraction network 2 to ensure that the input weakly coupled optical fiber 1 is reasonably converted to the optical fiber array 3; the optimization process of the cascade diffraction network 2 is implemented using the method in Example 1.

[0093] In this embodiment, the input weakly coupled optical fiber 1, the cascade diffraction network 2, and the output optical fiber array 3 can be fixed to the same substrate by ultraviolet glue after preparation to form a stably packaged module demultiplexer.

[0094] Embodiment 3

[0095] This embodiment provides a specific embodiment of a demultiplexer with 6 mode groups based on the first and second embodiments.

[0096] like Figure 2 As shown, the input weakly coupled fiber 1 adopts a few-mode fiber, which includes Q modules 1-1 to 1-Q to be demultiplexed, where Q is set to 6. Each module consists of 1 (0th order module) or 2 (higher order modules) orbital angular momentum modes, and the 6 modules include -5th to +5th order orbital angular momentum modes.

[0097] As attached Figure 2As shown, the output fiber array 3 uses multimode fiber, including 6 channels, denoted as 3-1 to 3-Q, which correspond to the 6 modules to be demultiplexed in the input weakly coupled fiber 1 to receive the energy of the corresponding modules. Each multimode fiber of the output fiber array 3 only uses the first S mode bases, denoted as 4-1 to 4-S, where S=6, and multiple fibers remain consistent. Using a larger S can increase the mode conversion efficiency, but the larger divergence angle of the higher-order mode will also increase additional coupling loss.

[0098] The reflective cascade diffraction network 2 is composed of N spatially separated phase plates, where N=5, recorded as 2-1 to 2-N. Its working space is between the input weakly coupled fiber 1 and the output fiber array 3. Its function is to achieve spatial separation of the modules in the input weakly coupled fiber 1 and ensure that the module energy can be maximally coupled to the corresponding fiber array 3.

[0099] In this embodiment, the optical fiber array 3 receives the output of the cascade diffraction network 2 in an arbitrary superposition of mode bases 4-1 to 4-S. Accordingly, the target output of the cascade diffraction network 2 does not adopt a certain mode, but is an arbitrary linear combination of mode bases 4-1 to 4-S.

[0100] Based on the above framework, the cascade diffraction network 2 is further optimized to ensure reasonable conversion from the input weakly coupled fiber 1 to the fiber array 3. The process is as follows:

[0101] Step 1: Given the input weakly coupled fiber 1 mode, diffracted by Rayleigh-Sommerfeld, the cascade diffraction output is as follows:

[0102]

[0103] In the formula, N represents the number of phase plates; E 0,m represents the mth input mode of the input weakly coupled fiber 1; E N+1,m represents the corresponding final output mode field; M k represents the kth phase plate P k The control of the light field, k=1,2,3,...,N, is a diagonal matrix with exp(jP,k) as the diagonal element; F k represents the transmission matrix corresponding to the k-th diffraction, which depends on the starting position of the k-th diffraction space.

[0104] Step 2. After obtaining the output of cascade diffraction network 2, there is no need to compare it with the specific target mode field. E N+1,m Projected onto the mode basis of the corresponding channel of fiber array 3:

[0105]

[0106] In the formula, c m,n,r represents the projection coefficient, B n,r The output fiber array 3 n The receiving channel r A pattern base; S Indicates the number of mode bases supported by each optical fiber of the optical fiber array 3. The receiving optical fibers of different modules are different, so the corresponding mode bases are different.

[0107] Step 3. Calculate the energy of the output mode field coupled to each channel of the output fiber array 3. The coupling efficiency is the sum of the corresponding mode base components and can be described by the following transmission matrix T:

[0108]

[0109] Where, MG i Indicates the input weakly coupled fiber 1 i The set of all patterns in a pattern group, |MG i | is the first i The number of patterns contained in a pattern group.

[0110] Step 4. To inversely optimize the phase of each phase plate of the cascade diffraction network 2 P k , according to the transfer matrix T, the scalar evaluation function is defined as follows:

[0111]

[0112] In the formula, L It reflects the loss and crosstalk. a and b are used to control the weights of loss and crosstalk respectively. In this embodiment, a=1 and b=10. Q represents the number of channels of the optical fiber array 3; L Determined by the phase distribution of the phase plate, which is based on the phase plate parameter P k is a multivariate function of the variables.

[0113] Step 5. To facilitate actual processing, replace the phase plate phase with the actual thickness distribution H k , taking the reflective phase plate as an example, the thickness of the phase plate is H k With Phase Plate P k Relationship satisfaction P k =4π H k / λ, λ is the working wavelength, and the thickness of the phase plate is determined based on the error back propagation method. H k The gradient of is expressed as follows:

[0114]

[0115] In the formula, Im means taking the imaginary part of the operation result, and diag means taking the diagonal elements of the matrix.

[0116] Step 6. The above gradient information indicates the optimization direction, and then the gradient descent method is combined to further iteratively update a new set of phase plate parameters as follows:

[0117]

[0118] In the formula, α represents the learning rate, which is used to control the amplitude of parameter update. α can be set to a fixed value or can be adaptively determined by a specific algorithm during the iterative optimization process of parameters.

[0119] Step 7. Calculate the average loss IL of the mode group demultiplexing device ave and Crosstalk XT ave as follows:

[0120]

[0121] When the design loss and crosstalk reach the predetermined index, the above parameter updating process is terminated and the corresponding phase plate distribution is output.

[0122] As attached Figure 3 As shown, a phase diagram design for orbital angular momentum mode multiplexing is shown, which includes N=5 phase blocks and can be used for demultiplexing of 0~5 order orbital angular momentum mode groups.

[0123] In this embodiment, the phase plate design determined by the above process can be processed into an actual diffractive optical element by photolithography. In this embodiment, it can be processed by three binary overlays. To this end, the phase plate distribution needs to be discretized into 8 steps, and then binarized as follows:

[0124]

[0125] Where b 0 ,b 1 ,b 2 There are V binary lithography patterns, and the etching depth corresponding to each binary pattern is λ / 2 4 ,λ / 2 3 ,λ / 2 2 , to achieve 2π phase control at wavelength λ.

[0126] The processed phase plate element can be coated with a reflection enhancement film and combined with a few-mode optical fiber and a multi-mode optical array to construct a module demultiplexer.

[0127] The orbital angular momentum of the few-mode fiber input contains 11 modes, and the intensity distribution is shown in the attached figure. Figure 4 After the orbital angular momentum mode passes through the cascade diffraction network 2, the corresponding output light field result is as shown in the attached figure. Figure 5 As shown. It can be seen that the orbital angular momentum modes of different modules are output to different spatial positions, while the orbital angular momentum modes belonging to the same module are output to the same spatial position after passing through the cascade diffraction network 2, and can then be received through the same multimode optical fiber.

[0128] In the specific contents of the above-mentioned specific implementation methods, the various technical features can be combined in any non-contradictory manner. In order to make the description concise, not all possible combinations of the above-mentioned technical features 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.

[0129] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A mode group demultiplexing method based on a cascade diffraction network, characterized in that: The following steps are involved: S1. Initialize phase plate parameters; S2. Select input mode E 0,m : According to the input weakly coupled optical fiber, the mode field distribution of the selected input mode is set as the input light field; S3. Calculate the output mode field of the cascade diffraction: According to the selected input mode E 0,m Calculation of the output mode field E of the cascade diffraction network N+1,m ; Output mode field of cascaded diffraction network E N+1,m Calculated by the following formula: In the formula, N represents the number of phase plates; E 0,m represents the mth input mode of the input weakly coupled fiber; E N+1,m represents the corresponding final output mode field; M k represents the kth phase plate P k The control of the light field, k=1,2,3,...,N, is a diagonal matrix with exp(jPk) as the diagonal element; F k represents the transmission matrix corresponding to the k-th diffraction; S4. Calculate the output mode field E N+1,m Projection on the substrate: Output mode field E N+1,m Projected onto the mode basis of the corresponding channel of the output fiber array; the output mode field is calculated by the following formula E N+1,m Projection onto the substrate: In the formula, c m,n,r represents the projection coefficient, B n,r The output fiber array n The rth mode basis of the receiving channel; S Indicates the number of mode bases supported by each fiber of the fiber array; S5. Calculate module transfer matrix T i,j :Calculate the energy of the output mode field coupled to each channel of the output fiber array. The coupling efficiency is the sum of the corresponding mode base components and is calculated through the transmission matrix T i,j express; T i,j Represents the input terminal i The module sends, and the output end is j Power received by the module channel; transmission matrix T i,j Calculated by the following formula: Where, MG i Indicates the first i The set of all patterns in a pattern group, |MG i | is the first i The number of patterns contained in each pattern group; S6. According to the transmission matrix T i,j Calculate the evaluation function; S7. Error gradient back propagation: Determine the phase plate thickness based on the error back propagation method H k The gradient of S8. Gradient descent parameter correction: further iteratively updating a new set of phase plate parameters according to the gradient information in step S7 combined with the gradient descent method; S9. Calculate the average loss and crosstalk of the mode group demultiplexing device; and determine whether the loss and crosstalk meet the standards. If so, the optimization ends; if not, return to step S2.

2. The mode group demultiplexing method based on the cascade diffraction network according to claim 1, characterized in that: In step S6, according to the transmission matrix T, a scalar evaluation function is defined. L , to inversely optimize the phase plates of the cascade diffraction network P k , the evaluation function L is expressed as follows: In the formula, L It reflects the loss and crosstalk conditions. a and b are used to control the weights of loss and crosstalk respectively. Q Indicates the number of channels in the fiber array.

3. The mode group demultiplexing method based on the cascaded diffraction network according to claim 2, characterized in that: In step S7, the thickness of the phase plate H k With Phase Plate P k Relationship satisfaction P k =4π H k / λ, λ is the working wavelength, and the thickness of the phase plate is determined based on the error back propagation method. H k The gradient of is expressed as follows: In the formula, Im means taking the imaginary part of the operation result, and diag means taking the diagonal elements of the matrix.

4. The mode group demultiplexing method based on the cascade diffraction network according to claim 3, characterized in that: In step S8, the phase plate parameters are updated by the following formula: In the formula, α represents the learning rate, which is used to control the amplitude of parameter update.

5. The mode group demultiplexing method based on a cascaded diffraction network according to any one of claims 1 to 4, characterized in that: In step S9, the average loss IL of the device ave and Crosstalk XT ave Calculated by the following formula: In the formula, Q Indicates the number of channels in the fiber array.

6. The mode group demultiplexing method based on the cascade diffraction network according to claim 5, characterized in that: The phase plate is processed by binary processing, firstly the phase plate distribution is discretized into 2 V steps, and then binarized as follows: Where b0, b1, …, b V There are V binary lithography patterns, and the etching depth corresponding to each binary pattern is λ / 2 V +1 ,λ / 2 V ,λ / 2 2 , to achieve 2π phase control at wavelength λ.

7. A mode group demultiplexing device based on a cascaded diffraction network, characterized in that: It includes an input weakly coupled fiber, a cascade diffraction network and an output fiber array, wherein the input weakly coupled fiber includes Q modules to be demultiplexed, each module includes a certain number of modes; the output fiber array includes Q channels, each fiber of the output fiber array supports S mode bases, and the number of mode bases is not less than the number of modes in the input weakly coupled fiber module; the cascade diffraction network includes a number of spatially separated phase plates, the working spaces of the phase plates are between the input weakly coupled fiber and the output fiber array, and are used to achieve spatial separation of the modules in the input weakly coupled fiber, ensuring that the module energy can be maximally coupled to the corresponding output fiber array channel; Each channel of the output optical fiber array can receive the output of the cascaded diffraction network as any one of the corresponding mode bases or any linear superposition, and the target output of the cascaded diffraction network is any linear combination of S mode bases; wherein, the mode group demultiplexing method based on the cascaded diffraction network as described in any one of claims 1 to 6 is used to achieve the optimal design of the cascaded diffraction network.

Images