CN115021826A - Intelligent coding and decoding computing system and method for optical computing communication - Google Patents
Intelligent coding and decoding computing system and method for optical computing communication Download PDFInfo
- Publication number
- CN115021826A CN115021826A CN202210475945.3A CN202210475945A CN115021826A CN 115021826 A CN115021826 A CN 115021826A CN 202210475945 A CN202210475945 A CN 202210475945A CN 115021826 A CN115021826 A CN 115021826A
- Authority
- CN
- China
- Prior art keywords
- optical
- signal
- communication
- decoding
- computing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000003287 optical effect Effects 0.000 title claims abstract description 235
- 230000006854 communication Effects 0.000 title claims abstract description 166
- 238000004891 communication Methods 0.000 title claims abstract description 165
- 238000000034 method Methods 0.000 title claims abstract description 24
- 239000013307 optical fiber Substances 0.000 claims abstract description 52
- 238000004364 calculation method Methods 0.000 claims abstract description 33
- 238000013528 artificial neural network Methods 0.000 claims abstract description 20
- 238000009826 distribution Methods 0.000 claims description 24
- 230000001427 coherent effect Effects 0.000 claims description 11
- 230000008878 coupling Effects 0.000 claims description 11
- 238000010168 coupling process Methods 0.000 claims description 11
- 238000005859 coupling reaction Methods 0.000 claims description 11
- 239000000835 fiber Substances 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 11
- 238000010801 machine learning Methods 0.000 claims description 10
- 238000005457 optimization Methods 0.000 claims description 9
- 230000006870 function Effects 0.000 abstract description 24
- 238000012545 processing Methods 0.000 abstract description 9
- 230000005540 biological transmission Effects 0.000 description 11
- 238000012549 training Methods 0.000 description 10
- 238000006243 chemical reaction Methods 0.000 description 7
- 238000004088 simulation Methods 0.000 description 6
- 230000006835 compression Effects 0.000 description 4
- 238000007906 compression Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000004907 flux Effects 0.000 description 3
- 238000001459 lithography Methods 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 230000006872 improvement Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 230000008054 signal transmission Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000000206 photolithography Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
- H04B10/50—Transmitters
- H04B10/516—Details of coding or modulation
- H04B10/548—Phase or frequency modulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
- G06N3/067—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using optical means
- G06N3/0675—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using optical means using electro-optical, acousto-optical or opto-electronic means
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/088—Non-supervised learning, e.g. competitive learning
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
- H04B10/25—Arrangements specific to fibre transmission
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Neurology (AREA)
- Electromagnetism (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Optical Communication System (AREA)
Abstract
The application relates to the technical field of electric digital data processing, in particular to an intelligent coding and decoding computing system and method for optical computing communication, wherein the system comprises: the optical calculation coding module is used for coding and encrypting the input communication information in an unsupervised manner to generate an optical communication signal; an optical fiber communication module for transmitting an optical communication signal based on an optical fiber; and the optical calculation decoding module is used for receiving the optical communication signal, decoding and reconstructing the optical communication signal by utilizing optical modulation, and restoring the communication information. Therefore, the technical problem that the optical computing neural network in the related art is not suitable for the coding and decoding functions of optical communication, so that the requirements of speed, quality and confidentiality of the optical communication cannot be met simultaneously is solved.
Description
Technical Field
The application relates to the technical field of electric digital data processing, in particular to an intelligent coding and decoding computing system and method for optical computing communication.
Background
Communication is an important component of modern information society. Fiber-optic communications have assumed over 95% of the capacity of modern information communications with its advantages of low loss, high speed, high throughput, etc. At present, a large amount of coding, decoding and digital signal processing related to optical fiber communication are to convert a preprocessed electrical signal into an optical signal and transmit the optical signal in an optical fiber, or convert an analog optical signal into an electrical signal after receiving the electrical signal and process the electrical signal, so that the optical fiber communication depends heavily on the precision and the speed of photoelectric conversion.
Therefore, if the complex processing in the electrical signal domain can be completed in the optical signal domain through optical computation, not only can the power consumption and time cost of a large number of photoelectric conversion devices be saved, but also the extra error brought by photoelectric or electro-optical conversion can be effectively reduced, and the signal-to-noise ratio of the communication signal can be improved.
In the related technology, the complex intelligent calculation of an optical signal domain is realized through an optical diffraction deep neural network, the complex intelligent calculation comprises a full-connection network, a convolution neural network, a reservoir calculation, a pulse neural network and the like, and various tasks such as image classification, audio recognition, significance detection and the like can be realized. Meanwhile, the optical diffraction deep neural network can achieve higher processing speed and energy consumption reduced by orders of magnitude compared with a homofunctional electronic network.
However, the optical computing neural network that can be implemented in the related art is not suitable for the encoding and decoding functions of optical communication. The optical communication has high requirements for encoding and decoding, not only needs to meet the speed of communication frequency, but also needs to meet high flux and have certain encryption and anti-noise performance, prevents eavesdropping and error codes in the communication process, and needs to be improved urgently.
Disclosure of Invention
The application provides an intelligent coding and decoding computing system and method for optical computing communication, which aim to solve the technical problem that the optical computing neural network in the related technology is not suitable for the coding and decoding functions of optical communication, so that the requirements of speed, quality and confidentiality of optical communication cannot be met simultaneously.
An embodiment of a first aspect of the present application provides an intelligent encoding and decoding computing system for optical computing communication, including: the optical calculation coding module is used for coding and encrypting the input communication information in an unsupervised manner to generate an optical communication signal; an optical fiber communication module for transmitting the optical communication signal based on an optical fiber; and the optical calculation decoding module is used for receiving the optical communication signal, decoding and reconstructing the optical communication signal by utilizing optical modulation, and restoring the communication information.
Optionally, in an embodiment of the present application, the optical calculation coding module includes: the light source is used for converting the digital electric signal formed by the communication information into a coherent light signal; and an optical diffraction phase modulation layer compressing and modulating the coherent optical signal into the optical communication signal.
Optionally, in an embodiment of the present application, the optical computational coding module is further configured to compress and code the coherent optical signal to a preset low-dimensional space by using a pre-established unsupervised learning variational self-coder neural network based on the optical diffraction phase modulation layer, and obtain a compressed optical communication signal by constraining distribution of the signal in the preset low-dimensional space through normal distribution.
Optionally, in an embodiment of the present application, the light calculation decoding module includes: an optical phase diffraction modulation layer for reconstructing the optical communication signal to restore the digital electrical signal.
Optionally, in an embodiment of the present application, the optical phase diffraction modulation layer is programmably controlled by a lithography process or a spatial light modulator, and parameters of the optical phase diffraction modulation layer are optimized by a machine learning method until an optimization condition is satisfied.
Optionally, in an embodiment of the present application, the optical phase diffraction modulation layer is obtained by stacking a plurality of diffraction layers, wherein each diffraction layer is simulated by phase modulation and fresnel propagation of a preset spatial distance.
Optionally, in an embodiment of the present application, the fiber optic communication module includes: an array of coupling mirrors; the optical fiber bundle is used for performing space division multiplexing on the optical communication signals coupled by the coupling mirror array; and the collimating mirror array is used for recovering the signal at the fiber outlet end of the optical fiber bundle into space light to obtain the optical communication signal.
The embodiment of the second aspect of the present application provides an intelligent encoding and decoding computing method for optical computing communication, including: encoding and encrypting input communication information in an unsupervised manner to generate an optical communication signal; transmitting the optical communication signal based on an optical fiber; and receiving the optical communication signal, decoding and reconstructing the optical communication signal by utilizing optical modulation, and restoring the communication information.
The embodiment of the application can execute the encoding and decoding function based on the machine learning unsupervised data through the optical computing and encoding module, creates the optical fiber communication signal encoding and decoding which can be realized without extra photoelectric conversion and digital signal processing in an optical signal domain, combines the optical computing and decoding module, utilizes optical modulation to decode and reconstruct the optical communication signal, can obviously improve the speed and the energy efficiency of the optical fiber communication encoding and decoding, can realize the encryption function while encoding, and has the potential of further improvement by being fused with various existing encoding modes at present. Therefore, the technical problem that the optical computing neural network in the related art is not suitable for the coding and decoding functions of optical communication, so that the requirements on speed, quality and confidentiality of the optical communication cannot be met simultaneously is solved.
Additional aspects and advantages of the present application will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the present application.
Drawings
The foregoing and/or additional aspects and advantages of the present application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
fig. 1 is a schematic structural diagram of an intelligent encoding and decoding computing system for optical computing communication according to an embodiment of the present application;
FIG. 2 is a schematic diagram of a light computing communication intelligent codec computing system according to one embodiment of the present application;
FIG. 3 is a flow diagram of a light computing communication intelligent codec computing system according to one embodiment of the present application;
fig. 4 is a flowchart of a computing method of an intelligent codec for optical computing communication according to an embodiment of the present application.
Detailed Description
Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are exemplary and intended to be used for explaining the present application and should not be construed as limiting the present application.
The following describes an optical computing communication intelligent encoding and decoding computing system and method according to an embodiment of the present application with reference to the drawings. Aiming at the technical problems that the optical computing neural network in the related art is not suitable for the coding and decoding functions of optical communication and cannot simultaneously meet the requirements of speed, quality and confidentiality of the optical communication, which are mentioned in the background technology center, the application provides an intelligent coding and decoding computing system for the optical computing communication, in the system, the encoding and decoding function based on the machine learning unsupervised data can be executed through the optical computing encoding module, the optical fiber communication signal encoding and decoding can be realized without additional photoelectric conversion and digital signal processing in the optical signal domain, and combines with the optical calculation decoding module, utilizes optical modulation to decode and reconstruct the optical communication signals, can obviously improve the speed and the energy efficiency of optical fiber communication coding and decoding, the encryption function can be realized while encoding, and the method has the potential of further improvement by being fused with various existing encoding modes. Therefore, the technical problem that the optical computing neural network in the related art is not suitable for the coding and decoding functions of optical communication, so that the requirements of speed, quality and confidentiality of the optical communication cannot be met simultaneously is solved.
Specifically, fig. 1 is a schematic structural diagram of an optical computing communication intelligent encoding and decoding computing system provided in the embodiment of the present application.
As shown in fig. 1, the optical computing communication intelligent codec computing system 10 includes: the optical computing and encoding module 100, the optical fiber communication module 200 and the optical computing and decoding module 300.
Specifically, the optical computing and encoding module 100 is configured to encode and encrypt the input communication information without supervision, and generate an optical communication signal.
In an actual implementation process, the light calculation coding module 100 of the embodiment of the present application may be configured to perform coding of light velocity or near light velocity on an input original signal through a pre-trained unsupervised neural network, perform feature representation in a low-dimensional space, and generate an optical communication signal, so as to compress the input signal, thereby improving optical communication throughput.
Optionally, in an embodiment of the present application, the light calculation coding module 100 includes: a light source and an optically diffractive phase modulation layer.
The light source is used for converting a digital electric signal formed by communication information into a coherent light signal.
And an optical diffraction phase modulation layer for compressing and modulating the coherent optical signal into an optical communication signal.
It will be appreciated that the optical source may convert the signal to be transmitted, i.e. the digital electrical signal consisting of the communication information, into a coherent optical signal. The optically diffractive phase modulation layer can compress the original optical signal into a signal for transmission band encryption, i.e., an optical communication signal.
Optionally, in an embodiment of the present application, the light calculation coding module 100 is further configured to perform compression coding on the coherent light signal to a preset low-dimensional space by using a pre-established non-supervised learning variational self-encoder neural network based on the optical diffraction phase modulation layer, and obtain a compressed optical communication signal by constraining distribution of the signal in the preset low-dimensional space through normal distribution.
Specifically, the optical diffraction phase modulation layer can update parameters in a process of co-training with a reconstruction module through back propagation, so that the purpose of encoding and compressing input image information into a low-dimensional space unsupervised is achieved, an optical communication signal with low-dimensional characteristics is output, and noise in the optical fiber propagation process can be used as the prior of a neural network and added into training.
Further, the light calculation coding network model can be established as follows:
according to the embodiment of the application, an unsupervised learning variational self-encoder neural network can be established according to the communication data complexity and an optical fiber link model, an input signal is compressed and encoded to a low-dimensional space, the distribution of the input signal in the low-dimensional space is constrained through normal distribution, and the following KL divergence is used as a part of a loss function:
D[Q(z)||P(z|X)]＝E z～Q [logQ(z)-logP(z|X)]，
wherein z is a parameter in the coding space, X is a parameter in the original space, Q is a probability distribution in the coding space, P (z | X) is a data distribution to be estimated by learning, Q (z) is an estimated normal distribution, D is a divergence, D [ Q (z) | | P (z | X)]Is the distance between the two distributions of P (z | X) and Q (z), E z～Q To the expectation of the distribution Q with respect to the variable z.
Due to the particularity of the light calculation and encoding module 100, there is a certain constraint on the value of the low-dimensional space. Therefore, certain constraint can be given to the learned mean value of the Gaussian distribution during training, and meanwhile, the variance in the learned Gaussian distribution can be understood as modeling of the noise amplitude in the optical fiber, so that the variance training can be specified according to the physical actual link.
The optical fiber communication module 200 is used for transmitting optical communication signals based on optical fibers.
As a possible implementation manner, in the embodiment of the present application, the optical fiber communication module 200 may transmit the compressed optical communication signal through the optical fiber bundle in a space division multiplexing manner, so as to implement transmission of the optical communication signal.
Optionally, in an embodiment of the present application, the fiber optic communication module 200 includes: the device comprises a coupling mirror array, a fiber bundle and a collimating mirror array.
The optical fiber bundle is used for performing space division multiplexing on the optical communication signals coupled by the coupling mirror array.
And the collimating mirror array is used for recovering the signal at the fiber outlet end of the optical fiber bundle into space light to obtain an optical communication signal.
In an actual implementation process, the coupling mirror array can be used for coupling encrypted signals for transmission into the optical fiber bundle to realize space division multiplexing, further completing long-distance and low-loss optical signal transmission through the optical fiber bundle, and converting signals transmitted from the optical fiber bundle into space light through the micro lens array to obtain optical communication signals.
The optical computing and decoding module 300 is configured to receive the optical communication signal, decode and reconstruct the optical communication signal by using optical modulation, and restore the communication information.
As a possible implementation manner, the optical computing and decoding module 300 in this embodiment of the application may decode the spatial optical signal output by the optical fiber communication module 200, recover and reconstruct the spatial optical signal into an original signal, and restore original data for detection or next-stage transmission.
Optionally, in an embodiment of the present application, the light calculation decoding module 300 includes: an optical phase diffraction modulation layer.
The optical phase diffraction modulation layer is used for reconstructing the optical communication signal to restore the optical communication signal into a digital electric signal.
Specifically, the optical phase diffraction modulation layer according to the embodiment of the present application may reconstruct and restore the received compressed optical signal into an original signal, and then the restored original signal, that is, a digital electrical signal, may be detected by a detector or transmitted to a lower-level communication link, so as to facilitate normal use of optical communication.
Optionally, in an embodiment of the present application, the optical phase diffraction modulation layer is obtained by stacking a plurality of diffraction layers, wherein each diffraction layer is simulated by phase modulation and fresnel propagation of a preset spatial distance.
Further, the optical phase diffraction modulation layer and the optical diffraction phase modulation layer may each be formed by stacking a plurality of diffraction layers, wherein each diffraction layer of the optical phase diffraction modulation layer may be simulated by phase modulation and fresnel propagation over a spatial distance.
It should be noted that, in the embodiments of the present application, the propagation of light in all free space and homogeneous media can be simulated by fresnel propagation, and in a single-mode fiber, the propagation can be simulated by fundamental mode propagation.
In addition, the preset spatial distance may be set by a person skilled in the art according to practical situations, and is not particularly limited herein.
Alternatively, in one embodiment of the present application, the optical phase diffraction modulation layer is programmably controlled by a lithography process or a spatial light modulator, and the parameters of the optical phase diffraction modulation layer are optimized by a machine learning method until the optimization condition is satisfied.
Specifically, in the embodiment of the present application, a machine learning network may be established according to the simulation model by using inverse gradient propagation optimization model parameters, an image to be processed is used as input, an image to be input is used as a reference for loss function calculation, a suitable training set, a suitable verification set, and a suitable test set are constructed, a minimum mean square error is used as a loss function, an error inverse propagation algorithm is used to iteratively adjust parameters of an optical diffraction phase modulation layer in the optical calculation coding module 100, and an optimal optimization result is obtained by debugging hyper-parameters such as a compression ratio, a noise modeling amplitude, a diffraction layer scale of the optical diffraction phase modulation layer and/or the optical phase diffraction modulation layer.
Further, according to various parameters obtained by simulation optimization, the embodiment of the application can utilize a lithography technology to manufacture or a spatial light modulator to programmably control the optical diffraction phase modulation layer and/or the optical phase diffraction modulation layer, and optimize the parameters of the optical diffraction phase modulation layer and/or the diffraction layer of the optical phase diffraction modulation layer in a machine learning manner.
In summary, the embodiment of the present application can implement signal encoding and decoding of an optical signal domain through an optical speed or a near optical speed device based on an existing general optical fiber communication link, and the optical computation encoding module 100 and the optical computation decoding module 300 support high speed and large flux reconstruction of new data under an unsupervised condition, so that channel flux can be significantly improved, and encryption is implemented while compression and dimension reduction are performed, so that the communication link obtains a better transmission effect.
The working principle of the intelligent codec computing system 10 for optical computing communication according to the present application is described in detail with reference to fig. 2 and fig. 3.
As shown in fig. 2, the light computing communication smart codec computing system 10 may include: the optical computing and encoding module 100, the light source 110, the optical diffraction phase modulation layer 120, the diffraction layer 121, the optical fiber communication module 200, the coupling mirror array 210, the optical fiber bundle 220, the collimating mirror array 230, the optical computation decoding module 300, the optical phase diffraction modulation layer 310, the diffraction layer 311, the detector 320, and the lower transmission system 330.
The light calculation coding module 100 may be configured to perform coding of light speed or near light speed on an input original signal through a pre-trained unsupervised neural network, perform feature representation in a low-dimensional space, and generate an optical communication signal, so as to compress the input signal, thereby improving optical communication throughput.
The optical computing code module 100 may include: a light source 110 and an optically diffractive phase modulation layer 120.
The light source 110 may convert a signal to be transmitted, i.e., a digital electrical signal composed of communication information, into a coherent optical signal.
The optical diffraction phase modulation layer 120 is composed of a plurality of diffraction layers 121, and can compress an original optical signal into an optical communication signal used for transmitting an encrypted signal, and the optical diffraction phase modulation layer 120 can update parameters in a process of training with a reconstruction module through back propagation, so that the purpose of encoding and compressing input image information into a low-dimensional space unsupervised is achieved, and an optical signal with low-dimensional characteristics is output.
The optical fiber communication module 200 may transmit the compressed optical communication signal through the optical fiber bundle 220 in a space division multiplexing manner, so as to realize transmission of the optical communication signal.
The fiber optic communications module 200 may include: a coupling mirror array 210, a fiber bundle 220, and a collimating mirror array 230. In an actual implementation process, the coupling mirror array 210 may be configured to couple encrypted signals for transmission into the optical fiber bundle 220 to implement space division multiplexing, so as to complete long-distance and low-loss optical signal transmission through the optical fiber bundle 220, and convert signals transmitted in the optical fiber bundle 220 into spatial light through the microlens array 230, so as to obtain an optical communication signal.
The optical calculation decoding module 300 may decode the spatial optical signal output by the optical fiber communication module 200, recover and reconstruct the spatial optical signal into an original signal, and restore the original data for detection or transmission at the next stage.
The optical calculation decoding module 300 may include: an optical phase diffraction modulation layer 310, a detector 320 and a lower stage transmission system 330. Specifically, the optical phase diffraction modulation layer 310 of the embodiment of the present application is composed of a plurality of diffraction layers 311, and can reconstruct the received compressed optical signal into the original signal, i.e. the digital electrical signal, and then the recovered original signal can be detected by the detector 320 or transmitted to the lower communication link by the lower transmission system 330, so as to facilitate the normal use of optical communication.
As shown in fig. 3, the embodiment of the present application may include the following steps:
step S301: and establishing a mathematical model of the optical computing coding and decoding network. The optical computing and encoding module 100 implements encoding and encrypting of input communication information unsupervised mainly by establishing a mathematical model of an optical computing codec network, and generates an optical communication signal. Specifically, according to the embodiment of the application, an unsupervised learning variational self-encoder neural network can be established according to the communication data complexity and an optical fiber link model, the input signal is compressed and encoded to a low-dimensional space, the distribution of the input signal in the low-dimensional space is constrained through normal distribution, and the following KL divergence is used as a part of a loss function:
D[Q(z)||P(z|X)]＝E z～Q [logQ(z)-logP(z|X)]，
wherein z is a parameter in the coding space, X is a parameter in the original space, Q is a probability distribution in the coding space, P (z | X) is a data distribution to be estimated by learning, Q (z) is an estimated normal distribution, D is a divergence, D [ Q (z) | | P (z | X)]Is the distance between the two distributions of P (z | X) and Q (z), E z～Q To the expectation of the distribution Q with respect to the variable z.
Step S302: and establishing a physical simulation model of the optical computing coding and decoding network. In the actual implementation process, due to the particularity of the optical computing and encoding module 100, there is a certain constraint on the value of the low-dimensional space. Therefore, the embodiment of the application can provide certain constraint for the learned mean value of the gaussian distribution during training, and meanwhile, the variance in the learned gaussian distribution can be understood as modeling of the noise amplitude in the optical fiber, so that the embodiment of the application can specify the variance training according to the physical actual link.
In the light calculation decoding module 300, the optical phase diffraction modulation layer 310 is a stack of a plurality of diffraction layers 311, each diffraction layer 311 being simulated with phase modulation and fresnel propagation over a spatial distance.
It should be noted that in the embodiments of the present application, the propagation of light in all free space and homogeneous media is simulated by fresnel propagation, and in single-mode fibers, the propagation is simulated by fundamental mode.
Step S303: and constructing and selecting a proper training set, a proper verification set and a proper test set. Specifically, in the embodiment of the present application, a machine learning network may be established according to the simulation model, the image to be processed is used as an input, the image to be input is used as a reference for calculation of a loss function, a suitable training set, a suitable verification set, and a suitable test set are constructed, the minimum mean square error is used as the loss function, the error back propagation algorithm is used to iteratively adjust the parameters of the optical diffraction phase modulation layer 120 in the light calculation coding module 100, and the best optimization result is obtained by debugging the compression ratio, the noise modeling amplitude, the scale of the diffraction layer 121 of the optical diffraction phase modulation layer 120 and/or the scale of the diffraction layer 311 of the optical phase diffraction modulation layer 310, and other hyper-parameters.
Step S304: physical parameters of the optically diffractive phase modulation layer 120 and/or the optically diffractive phase modulation layer 310 are determined using an error back propagation algorithm optimization model. In the embodiment of the present application, the diffraction layer 121 of the optical diffraction phase modulation layer 120 and/or the diffraction layer 311 of the optical phase diffraction modulation layer 310 may be manufactured by using a photolithography technique or may be controlled by a spatial light modulator according to various parameters obtained by simulation optimization.
Step S305: manufacturing a system model, installing a hardware system and testing the communication quality. According to the embodiment of the application, the light source 110, the micro-lens array 230, the optical fiber bundle 220 and the like can be selected as appropriate, and the hardware system can be correctly installed according to the simulation model, so that the functions of optical computing communication intelligent coding and decoding can be realized.
According to the optical computation communication intelligent coding and decoding computing system provided by the embodiment of the application, the coding and decoding function based on machine learning unsupervised data can be executed through the optical computation coding module, the optical fiber communication signal coding and decoding which can be realized without extra photoelectric conversion and digital signal processing in an optical signal domain is created, the optical fiber communication signal is decoded and reconstructed by utilizing optical modulation in combination with the optical computation decoding module, the speed and the energy efficiency of the optical fiber communication coding and decoding can be obviously improved, the encryption function can be realized while coding is carried out, and the optical computation communication intelligent coding and decoding computing system has the potential of being further improved by being fused with various existing coding modes. Therefore, the technical problem that the optical computing neural network in the related art is not suitable for the coding and decoding functions of optical communication, so that the requirements of speed, quality and confidentiality of the optical communication cannot be met simultaneously is solved.
The method for calculating the intelligent coding and decoding of the optical computing communication proposed by the embodiment of the application is described next with reference to the attached drawings.
Fig. 4 is a flowchart of a method for calculating an intelligent codec for optical computing communication according to an embodiment of the present application.
As shown in fig. 4, the method for calculating the optical computing communication intelligent codec using the above-mentioned system for calculating the optical computing communication intelligent codec includes the following steps:
in step S401, the input communication information is encoded and encrypted unsupervised, and an optical communication signal is generated.
In step S402, an optical communication signal is transmitted based on an optical fiber.
In step S403, the optical communication signal is received, and the optical communication signal is decoded and reconstructed by using optical modulation, so as to recover the communication information.
It should be noted that the foregoing explanation of the embodiment of the optical computing communication intelligent codec calculation system is also applicable to the optical computing communication intelligent codec calculation method of the embodiment, and is not repeated here.
According to the optical computation communication intelligent coding and decoding calculation method provided by the embodiment of the application, the coding and decoding function based on machine learning unsupervised data can be executed through the optical computation coding module, the optical fiber communication signal coding and decoding which can be realized without extra photoelectric conversion and digital signal processing in an optical signal domain is created, the optical fiber communication signal is decoded and reconstructed by utilizing optical modulation in combination with the optical computation decoding module, the speed and the energy efficiency of the optical fiber communication coding and decoding can be obviously improved, the encryption function can be realized while coding is carried out, and the method has the potential of being further improved by being fused with various existing coding modes at present. Therefore, the technical problem that the optical computing neural network in the related art is not suitable for the coding and decoding functions of optical communication, so that the requirements of speed, quality and confidentiality of the optical communication cannot be met simultaneously is solved.
In the description herein, reference to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or N embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the feature. In the description of the present application, "N" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
Claims (8)
1. An intelligent codec computing system for optical computing communications, comprising:
the optical calculation coding module is used for coding and encrypting the input communication information in an unsupervised manner to generate an optical communication signal;
an optical fiber communication module for transmitting the optical communication signal based on an optical fiber; and
and the optical calculation decoding module is used for receiving the optical communication signal, decoding and reconstructing the optical communication signal by utilizing optical modulation, and restoring the communication information.
2. The system of claim 1, wherein the light computing encoding module comprises:
the light source is used for converting the digital electric signal formed by the communication information into a coherent light signal;
and an optical diffraction phase modulation layer compressing and modulating the coherent optical signal into the optical communication signal.
3. The system of claim 2, wherein the optical computational encoding module is further configured to, based on the optical diffraction phase modulation layer, utilize a pre-established unsupervised learning variational self-encoder neural network to compressively encode the coherent optical signal into a preset low-dimensional space, and constrain distribution of the signal in the preset low-dimensional space by normal distribution to obtain a compressed optical communication signal.
4. The system of claim 2 or 3, wherein the optical computing decoding module comprises:
an optical phase diffraction modulation layer for reconstructing the optical communication signal to restore the digital electrical signal.
5. The system of claim 4, wherein the optical phase diffraction modulation layer is programmably controlled by a lithographic process or a spatial light modulator, and parameters of the optical phase diffraction modulation layer are optimized by a machine learning method until an optimization condition is satisfied.
6. The system according to claim 4 or 5, wherein the optical phase diffraction modulation layer is obtained by stacking a plurality of diffraction layers, wherein each diffraction layer is simulated by phase modulation and Fresnel propagation at a preset spatial distance.
7. The system of claim 1, wherein the fiber optic telecommunications module comprises:
an array of coupling mirrors;
the optical fiber bundle is used for performing space division multiplexing on the optical communication signals coupled by the coupling mirror array;
and the collimating mirror array is used for recovering the signal at the fiber outlet end of the optical fiber bundle into space light to obtain the optical communication signal.
8. A light computing communication smart codec calculation method, characterized in that it uses a light computing communication smart codec calculation system according to claims 1-7, wherein the method comprises the following steps:
encoding and encrypting input communication information in an unsupervised manner to generate an optical communication signal;
transmitting the optical communication signal based on an optical fiber;
and receiving the optical communication signal, decoding and reconstructing the optical communication signal by utilizing optical modulation, and restoring the communication information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210475945.3A CN115021826B (en) | 2022-04-29 | 2022-04-29 | Intelligent coding and decoding computing system and method for optical computing communication |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210475945.3A CN115021826B (en) | 2022-04-29 | 2022-04-29 | Intelligent coding and decoding computing system and method for optical computing communication |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115021826A true CN115021826A (en) | 2022-09-06 |
| CN115021826B CN115021826B (en) | 2024-04-16 |
Family
ID=83067691
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210475945.3A Active CN115021826B (en) | 2022-04-29 | 2022-04-29 | Intelligent coding and decoding computing system and method for optical computing communication |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115021826B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105572810A (en) * | 2014-10-31 | 2016-05-11 | 安华高科技通用Ip(新加坡)公司 | Conditioned launch of a single mode light source into a multimode optical fiber |
| CN106506138A (en) * | 2016-11-11 | 2017-03-15 | 深圳大学 | The cross-layer encryption method of space optical communication system and system |
| CN110334804A (en) * | 2019-06-20 | 2019-10-15 | 清华大学 | Full optical depth diffraction nerve network system and method based on space segment coherent light |
| CN111582468A (en) * | 2020-04-02 | 2020-08-25 | 清华大学 | Photoelectric hybrid intelligent data generation and calculation system and method |
| CN113411136A (en) * | 2021-06-16 | 2021-09-17 | 广东工业大学 | Quadrature modulation secret optical communication device and method |
| US20220014276A1 (en) * | 2020-07-10 | 2022-01-13 | Raytheon Company | Receiver and system for transporting and demodulating complex optical signals |
-
2022
- 2022-04-29 CN CN202210475945.3A patent/CN115021826B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105572810A (en) * | 2014-10-31 | 2016-05-11 | 安华高科技通用Ip(新加坡)公司 | Conditioned launch of a single mode light source into a multimode optical fiber |
| CN106506138A (en) * | 2016-11-11 | 2017-03-15 | 深圳大学 | The cross-layer encryption method of space optical communication system and system |
| CN110334804A (en) * | 2019-06-20 | 2019-10-15 | 清华大学 | Full optical depth diffraction nerve network system and method based on space segment coherent light |
| CN111582468A (en) * | 2020-04-02 | 2020-08-25 | 清华大学 | Photoelectric hybrid intelligent data generation and calculation system and method |
| US20220014276A1 (en) * | 2020-07-10 | 2022-01-13 | Raytheon Company | Receiver and system for transporting and demodulating complex optical signals |
| CN113411136A (en) * | 2021-06-16 | 2021-09-17 | 广东工业大学 | Quadrature modulation secret optical communication device and method |
Non-Patent Citations (1)
| Title |
|---|
| 徐琳: "光学加密技术与信息安全的融合与应用", 网络安全技术与应用, 15 December 2015 (2015-12-15) * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115021826B (en) | 2024-04-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110334804B (en) | All-optical depth diffraction neural network system and method based on spatial partially coherent light | |
| Erdemir et al. | Generative joint source-channel coding for semantic image transmission | |
| Dür et al. | Quantum repeaters based on entanglement purification | |
| WO2021203243A1 (en) | Artificial intelligence-based mimo multi-antenna signal transmission and detection technique | |
| CN113169943B (en) | Demodulation method, demodulator and modulation system | |
| CN102594492B (en) | A kind of code block dividing method and device | |
| CN111582468B (en) | Photoelectric hybrid intelligent data generation and calculation system and method | |
| Li et al. | Quantization-aware deep optics for diffractive snapshot hyperspectral imaging | |
| CN115021826B (en) | Intelligent coding and decoding computing system and method for optical computing communication | |
| Alawad et al. | Innovative variational autoencoder for an end-to-end communication system | |
| Feng et al. | Deep Learning-Based Joint Channel Estimation and CSI Feedback for RIS-Assisted Communications | |
| CN113228535B (en) | Apparatus and method for adaptive decoding and core scrambling in a multi-core optical fiber transmission system | |
| CN113890801B (en) | Compressed sensing signal acquisition system and method adopting frequency domain coding and mixing | |
| Randles et al. | Success probabilities in time-reversal-based hybrid quantum state transfer | |
| Al-Shamayleh et al. | Performance Evaluation of Optical Transmission Based on Link Estimation by Using Deep Learning Techniques | |
| CN113468554A (en) | High-fidelity holographic orbital angular momentum communication method based on deep learning | |
| Park et al. | Federated Codebook for Multi-User Deep Source Coding | |
| McKinstrie et al. | Information efficiencies of parametric devices | |
| Li et al. | FedBAT: Communication-Efficient Federated Learning via Learnable Binarization | |
| CN102263558A (en) | Signal processing method and system | |
| Blard et al. | Overfitted image coding at reduced complexity | |
| Murthy et al. | High-rate vector quantization for noisy channels with applications to wideband speech spectrum compression | |
| WO1997012330A1 (en) | Method and apparatus for information processing using cellular automata transform | |
| CN111988136B (en) | Multi-precision decoding device and method for continuous variable quantum key distribution | |
| CN116980291A (en) | System optimization method, device, equipment and computer readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |