CN113065494B - Vortex electronic mode identification system, method and device and electronic equipment - Google Patents
Vortex electronic mode identification system, method and device and electronic equipment Download PDFInfo
- Publication number
- CN113065494B CN113065494B CN202110394918.9A CN202110394918A CN113065494B CN 113065494 B CN113065494 B CN 113065494B CN 202110394918 A CN202110394918 A CN 202110394918A CN 113065494 B CN113065494 B CN 113065494B
- Authority
- CN
- China
- Prior art keywords
- oam
- vortex
- electron
- intensity distribution
- field intensity
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000009826 distribution Methods 0.000 claims abstract description 67
- 238000010894 electron beam technology Methods 0.000 claims abstract description 51
- 230000003321 amplification Effects 0.000 claims abstract description 22
- 238000003199 nucleic acid amplification method Methods 0.000 claims abstract description 22
- 238000012545 processing Methods 0.000 claims abstract description 21
- 238000012549 training Methods 0.000 claims description 74
- 238000013473 artificial intelligence Methods 0.000 claims description 44
- 238000012795 verification Methods 0.000 claims description 14
- 238000013527 convolutional neural network Methods 0.000 claims description 13
- 238000002360 preparation method Methods 0.000 claims description 10
- 230000009471 action Effects 0.000 claims description 5
- 230000002238 attenuated effect Effects 0.000 claims description 5
- 238000012706 support-vector machine Methods 0.000 claims description 5
- 239000013078 crystal Substances 0.000 claims description 4
- 238000004088 simulation Methods 0.000 claims description 4
- 238000003066 decision tree Methods 0.000 claims description 3
- 239000010931 gold Substances 0.000 claims description 3
- 229910052737 gold Inorganic materials 0.000 claims description 3
- 230000006698 induction Effects 0.000 claims description 3
- 239000013598 vector Substances 0.000 claims description 3
- 230000005428 wave function Effects 0.000 claims description 3
- 230000008878 coupling Effects 0.000 claims description 2
- 238000010168 coupling process Methods 0.000 claims description 2
- 238000005859 coupling reaction Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 16
- 239000010410 layer Substances 0.000 description 12
- 238000004891 communication Methods 0.000 description 9
- 230000006870 function Effects 0.000 description 7
- 238000013528 artificial neural network Methods 0.000 description 6
- 238000011176 pooling Methods 0.000 description 6
- 230000005540 biological transmission Effects 0.000 description 5
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 4
- 238000007796 conventional method Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000002003 electron diffraction Methods 0.000 description 2
- 238000011084 recovery Methods 0.000 description 2
- 238000001228 spectrum Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 238000012850 discrimination method Methods 0.000 description 1
- 230000005520 electrodynamics Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 229910052751 metal Inorganic materials 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 239000013384 organic framework Substances 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000002344 surface layer Substances 0.000 description 1
- 230000002195 synergetic effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
-
- 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
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Analysing Materials By The Use Of Radiation (AREA)
Abstract
The invention discloses a vortex electronic mode recognition system, a method, a device and electronic equipment. The vortex electron generated by the vortex electron generation module is coupled with orbital angular momentum to generate vortex electron beams carrying different modes of OAM; the diffraction amplification module is used for diffracting and amplifying the vortex electron beam; the image receiving and collecting module is used for receiving the diffraction amplified vortex electron beams and collecting the diffraction amplified vortex electron beams to obtain an OAM field intensity distribution image; the data processing and identifying module receives the OAM field intensity distribution image and carries out OAM modal identification, and the data processing and identifying module not only realizes vortex electronic orbital angular momentum modal identification, but also can reduce the influence of nonlinear distortion on an identification result, and has higher identification accuracy and lower hardware complexity and cost.
Description
Technical Field
The invention belongs to the technical field of artificial intelligence and electromagnetic wave Orbital Angular Momentum (OAM) wireless communication, and particularly relates to a vortex electronic mode identification system, method, device and electronic equipment.
Background
Compared with conventional electromagnetic wave communication, the OAM is introduced, so that the transmission capacity and the spectrum efficiency can be effectively increased, the situation of spectrum resource shortage is relieved, and high-capacity high-speed transmission is realized. The free-space optical OAM communication system has the characteristics that the OAM light beam generator is low in integration level, light waves are greatly influenced by environmental factors such as atmospheric turbulence and haze, and the like. The OAM wave beam divergence angle of radio frequency electromagnetic wave (below 300 GHz) is large, long-distance transmission is difficult to realize, and the receiving end OAM antenna has a complex structure, is difficult to integrate and has high cost. Unlike the above statistical OAM beams, the use of quantum OAM for communication can solve the above problems well.
Quantum OAM realizes information transmission by orbital angular momentum carried by Vortex electrons, and documents (Zhang, c., Xu, p., & Jiang, X. (2020.) Vortex electron generated by microwave photon with atomic and modulated metal in a magnetic field. aipaddynamics, 10,105230.) propose that based on Quantum Electrodynamics (QED) theory, electromagnetic waves carrying OAM can be directly radiated by relativistic cyclotron electrons or multiple electrons at a transmitting end, then the Quantum electromagnetic waves are absorbed by the relativistic electrons at a receiving end to become Vortex electrons, and the information recovery is realized by identifying the OAM mode of the Vortex electrons. A corresponding method for identifying the mode of OAM of the vortex electrons is given in the literature (Zhang, c., Xu, P. & Jiang, x.detecting synergistic organic frameworks in the magnetic field by the crystal diffusion. eur. phys. j. plus 136,60 (2021)): the eddy electrons are diffracted by gold foil to amplify the size of a wave beam, different mode OAM electrons are separated to different positions through a mapper and a sorting device, and then information recovery is realized by a position judgment method. In practical application, the magnetic field generated by the gyrotron at the receiving end cannot be directly cut off, but gradually attenuates from the diffraction gold foil to the fluorescent screen. Therefore, eddy electrons are affected by the residual magnetic field and undergo a shift, which causes a serious nonlinear distortion, and the effect of OAM mode separation by a device-based method of ideal magnetic field modeling is not ideal. In addition, an image acquisition Device, such as a CCD (Charge Coupled Device) camera, may be used to collect a diffraction image of the electronic OAM, and then perform OAM mode identification and recover information by means of image identification, but the conventional image identification method can only extract surface layer features such as edges of the image, and cannot cope with the non-linear distortion caused by the residual magnetic field well. Table 1 compares the above common vortex electron OAM mode identification methods, which mainly include a device-based method and an image processing method, and both methods have merits but cannot better resist nonlinear distortion of vortex electron diffraction images. Therefore, it is urgently needed to provide a new OAM electronic mode identification system which can cope with the influence of the nonlinear distortion of the magnetic field and ensure higher identification accuracy and low hardware complexity.
Table 1: vortex electronic OAM modal identification method comparison
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a vortex electronic mode identification system based on artificial intelligence, which introduces a large amount of distorted OAM field intensity distribution images in training data and adopts convolution to extract a neural network to extract deep features. The system can realize vortex electronic mode detection on the premise of low cost and low hardware complexity, can still ensure high identification accuracy and high identification rate when the image is seriously distorted, and can reduce modal crosstalk caused by defects of a receiving end discrimination method when used for an OAM quantum state wireless communication system.
A vortex electronic modality identification system, comprising:
the vortex electron generation module is used for generating cyclotron electrons and coupling the cyclotron electrons with orbital angular momentum of received OAM electromagnetic waves to generate vortex electron beams carrying different modes of OAM;
the diffraction amplification module is used for diffracting and amplifying the size of the vortex electron beam;
the image receiving and collecting module is used for receiving the diffraction amplified vortex electron beam and collecting the vortex electron beam to obtain an OAM field intensity distribution image;
the data processing and identifying module is connected behind the image receiving and collecting module and used for receiving the OAM field intensity distribution image and performing data processing and OAM modal identification, and the data processing and identifying module comprises:
the training data preparation submodule is used for generating an OAM field intensity distribution image which is influenced by a residual magnetic field and contains nonlinear distortion as training data of the artificial intelligent model training submodule, wherein the residual magnetic field is a magnetic field which is generated by the vortex electron generation module and gradually attenuated from the diffraction amplification module to the image receiving and acquiring module;
the artificial intelligence model training submodule is used for constructing an artificial intelligence model and performing artificial intelligence model training by using the training data so as to obtain the artificial intelligence model with the highest verification and identification accuracy;
and the mode identification submodule is used for carrying out real-time OAM mode identification on the OAM field intensity distribution image acquired by the image receiving and acquiring module by using the artificial intelligence model with the highest verification and identification accuracy.
Optionally, the laguerre-gaussian vortex electron beam is used to simulate a vortex electron beam carrying different modes of OAM along the z-axis direction, and the expression is as follows:
wherein
Is the wave function of the eddy electron beam carrying the OAM,
is a constant determined by the number of radial vectors n and the number of OAM modes l,
is a generalized laguerre polynomial,
is the radius of the girdling, wherein
Is a reduced Planck constant, e is the element charge, B is the magnetic induction of the environment outside the electron beam, i is the imaginary unit,
is the three coordinate variables of the cylindrical coordinate system, k is the wave number.
Optionally, the vortex electron beam generated by the vortex electron generation module includes a vortex electron beam carrying pure OAM and/or a plurality of vortex electron beams of mixed OAM, where any mixed OAM is formed by combining a plurality of pure OAM.
Optionally, the vortex electron generation module includes a power supply, a gyrotron, and a superconducting magnet, where the gyrotron includes an electron gun and an electron cyclotron generation module, the superconducting magnet is located outside the electron cyclotron generation module, the power supply supplies power to the electron gun, electrons emitted and accelerated by the electron gun enter the electron cyclotron generation module, and perform cyclotron motion under the action of a magnetic field of the superconducting magnet, and an OAM electromagnetic wave couples its orbital angular momentum to the cyclotron electrons, so as to generate vortex electron beams carrying OAM in different modes;
the diffraction amplification module comprises a poly-gold crystal film which is used for performing diffraction amplification on the vortex electron beam;
the image receiving and collecting module comprises a fluorescent screen and a camera, the fluorescent screen is used for receiving the vortex electron beams after diffraction amplification, and the camera is used for collecting images on the fluorescent screen as OAM field intensity distribution images.
Optionally, the training data includes OAM field intensity distribution images affected by residual magnetic fields of different degrees, where the nonlinear distortion generated by the residual magnetic fields refers to one or more of nonlinear divergence and nonlinear convergence of the OAM field intensity distribution images.
Optionally, the non-linear divergence and the non-linear convergence refer to that distances from two points before and after the distortion of the residual magnetic field to the electron cyclotron center O satisfy the following relationship:
wherein R and R' are distances of the electrons before and after the distortion from the electron cyclotron center O in the cylindrical coordinates, respectively, R is a radius of the electron cyclotron motion, β is a central angle of the electron cyclotron motion, and a deflection angle of the positions of the electrons before and after the distortion with respect to the electron cyclotron center O is obtained by the following equation:
α＝arcsin(2R[sin(β/2)] 2 /r' 2 )。
optionally, the artificial intelligence model includes one or more of a convolutional neural network, a decision tree, a multi-classification Support Vector Machine (SVM), and a K-nearest neighbor algorithm.
The invention also provides a vortex electronic mode identification method, which comprises the following steps:
after receiving an OAM electromagnetic wave transmitted in a free space, a vortex electron generation module couples the orbital angular momentum of the OAM electromagnetic wave to a cyclotron electron to generate vortex electron beams carrying different modes of OAM;
the diffraction amplification module is used for carrying out diffraction amplification on the vortex electron beam;
the image receiving and collecting module comprises a fluorescent screen and a camera, the fluorescent screen receives the vortex electron beam amplified by diffraction to obtain an OAM field intensity distribution image, and the camera collects the OAM field intensity distribution image in real time and sends the OAM field intensity distribution image to the data processing and identifying module;
the data processing and identifying module utilizes the artificial intelligence model with the highest accuracy to identify the received OAM field intensity distribution image in real time to complete the OAM modal identification,
the training data preparation submodule of the data processing and recognition module performs binarization data processing on a plurality of OAM field intensity distribution images which are influenced by a residual magnetic field and contain nonlinear distortion and are received in advance, then the processed data are used as a part of model training data to be sent to the artificial intelligent model training submodule, the training data preparation submodule also generates a part of distorted and/or undistorted simulation OAM field intensity distribution images as another part of model training data through simulation, and the part of model training data and/or the other part of model training data jointly form training data,
and the artificial intelligence model training submodule utilizes the training data to carry out artificial intelligence model training, so that the artificial intelligence model with the highest verification and identification accuracy is obtained.
The present invention also provides an OAM mode identification apparatus, which includes:
the training data preparation submodule is used for generating an OAM field intensity distribution image which is influenced by a residual magnetic field and contains nonlinear distortion as training data of the artificial intelligent model training submodule, wherein the residual magnetic field is a magnetic field which is generated by the vortex electron generation module and gradually attenuated from the diffraction amplification module to the image receiving and acquiring module;
the artificial intelligence model training submodule is used for constructing an artificial intelligence model and performing artificial intelligence model training by using the training data so as to obtain the artificial intelligence model with the highest verification and identification accuracy;
and the mode identification submodule is used for carrying out real-time OAM mode identification on the OAM field intensity distribution image acquired by the image receiving and acquiring module by using the artificial intelligence model with the highest verification and identification accuracy.
The present invention also provides an electronic device, including:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the steps of:
carrying out binarization data processing on a plurality of OAM field intensity distribution images which are influenced by a residual magnetic field and contain nonlinear distortion and are received in advance, and then sending the processed data as a part of model training data to an artificial intelligent model training submodule;
carrying out artificial intelligence model training by using the training data so as to obtain an artificial intelligence model with the highest verification and identification accuracy;
and identifying the received OAM field intensity distribution image in real time by using the artificial intelligence model with the highest accuracy rate to complete OAM mode identification.
Compared with the prior art, the invention has the beneficial effects that: the vortex electronic OAM mode identification system does not need excessive devices, is low in hardware complexity and cost, can reduce the influence of nonlinear distortion caused by a residual magnetic field and the like on an identification result, is high in identification accuracy, and can be used in a plurality of application scenes such as quantum state OAM wireless communication and the like.
Drawings
The foregoing and/or additional aspects and advantages of the present invention 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 diagram of a vortex electronic mode identification system according to an embodiment of the present invention;
FIG. 2 is a diagram of a system hardware architecture according to an embodiment of the present invention;
FIG. 3 is a geometric relationship of the positions before and after an electronic distortion according to an embodiment of the present invention;
fig. 4(a) is a diagram showing the OAM field intensity distribution of a pure state OAM (l ═ 1) before being distorted by the residual magnetic field;
fig. 4(b) is an OAM field intensity distribution diagram of a pure state OAM (l ═ 1) that converges after being affected by a remanent magnetic field;
fig. 4(c) is a diagram showing the OAM field intensity distribution of a pure OAM (l ═ 1) in which divergent distortion occurs after being affected by a residual magnetic field;
fig. 5(a1) - (d1) show field intensity distributions of electron beams in which OAM modes are l {1, {1, -2}, {3, -3}, {1,2,3} } before distortion is affected by the residual magnetic field;
fig. 5(a2) - (d2) are field intensity distribution diagrams showing convergence distortion of electron beams in OAM modes l ═ {1, {1, -2}, {3, -3}, {1,2,3} } under the influence of the residual magnetic field;
fig. 5(a3) - (d3) are field intensity distribution diagrams of divergent distortion of an electron beam in which OAM modes are l {1, {1, -2}, {3, -3}, {1,2,3} } respectively, after being influenced by a residual magnetic field;
fig. 6(a) is an OAM field intensity distribution plot for the mode number l ═ {3, -3} in the absence of noise (signal-to-noise ratio + ∞);
fig. 6(b) - (d) are OAM field intensity distribution diagrams of the mode number l ═ {3, -3} at the SNR of 30, 20 and 15, respectively, after gaussian white noise is added;
FIG. 7 is a convolutional neural network structure, according to an embodiment of the present invention;
FIG. 8 is a graph of loss function values for a neural network, in accordance with an embodiment of the present invention;
FIG. 9 is a graph of the accuracy of a neural network, according to an embodiment of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.
As shown in fig. 1, a vortex electronic modality identification system 10 includes a vortex electronic generation module 100, a diffraction amplification module 200, an image receiving and acquisition module 300, and a data processing and identification module 400.
The eddy current electron generation module 100 includes a high voltage power supply 101, a gyrotron 102, and a superconducting magnet 103, wherein the gyrotron 102 is mainly composed of a high speed electron gun 1021 and a cylindrical electron cyclotron generation module. The high-voltage power supply 101 supplies power to the electron gun 1021, provides high-voltage direct current needed by electron acceleration, and electrons are emitted and accelerated by the electron gun 1021; the superconducting magnet 103 is located outside an electron cyclotron generation module of the gyrotron, electrons perform high-speed cyclotron motion under the action of a magnetic field of the superconducting magnet 103, and after OAM electromagnetic waves propagated in a free space irradiate the high-speed cyclotron, orbital angular momentum of the OAM electromagnetic waves is coupled to the cyclotron, so that vortex electron beams carrying different modes of OAM are generated; the diffraction amplification module 200 is used for carrying out diffraction amplification on the vortex electron beam by using the poly-gold crystal film 201; the image receiving and acquisition module 300 is located behind the diffraction amplification module and comprises a fluorescent screen 301 and a high speed camera 302. The fluorescent screen 301 is used for receiving the diffracted and amplified vortex electron beams; the high-speed camera 302 is positioned behind the fluorescent screen and is used for collecting an OAM field intensity distribution image (namely an OAM field intensity distribution image) on the fluorescent screen; the data processing and recognition module 400 is a computer built with a convolutional neural network, and includes the following sub-modules: a training data preparation submodule 401 for generating training data of the neural network; an artificial intelligence model training module 402, configured to construct an artificial intelligence model, such as a convolutional neural network, and perform network training and verification; and a mode identification module 403, which inputs the acquired OAM field intensity distribution image into a trained convolutional neural network for OAM mode identification. The position relationship of each module in the system is shown in the system hardware structure diagram of fig. 2.
The system can be used as a receiving end of quantum state OAM wireless communication, the transmitting end transmits information by transmitting OAM electromagnetic waves with different modal numbers, the receiving end receives the OAM electromagnetic waves transmitted by the transmitting end, the received OAM modes are coupled to vortex electrons by the system, and the identification of the OAM modes is realized after diffraction amplification and data processing, so that the transmitted original information can be recovered. The values of the relevant parameters involved in the system in this example are shown in table 2.
Table 2: list of relevant parameters
In this embodiment, the vortex electron beam carrying OAM generated by the vortex electron generation module 100 may be simulated by a Laguerre-Gaussian (LG) vortex electron beam, and the LG vortex electron along the z-axis direction (the same direction as the axial direction of the vortex electron cyclotron central axis) may be described as LG vortex electron
Wherein
Is the wave function of the eddy electron beam carrying the OAM,
is a constant determined by the number of radial vectors n and the number of OAM modes l,
is a generalized laguerre polynomial,
is the radius of the girdling, wherein
Is the reduced planck constant, e is the element charge, B is the magnetic induction of the environment outside the electron beam (i.e. the superconducting magnet outside the cyclotron tube), i is the imaginary unit,
is the three coordinate variables of the cylindrical coordinate system, k is the wave number.
In this embodiment, the remanent magnetic field is along the z-axis with a magnitude of B z Decays along the z-axis and is a constant magnetic field. Under the influence of the lorentz force, the electrons make a helical motion, i.e. in addition to a motion along the z-axis, they also make a counterclockwise circular motion along a plane perpendicular to the z-axis. Because the exit angles of the diffracted vortex electrons of different modes are different, the gyration radii in the residual magnetic field are also different, and the gyration radius R can be calculated by the following formula:
where gamma is the Lorentzian factor of the electron, m e Is the electron mass, e is the electron element charge, v is the velocity of the electron before diffraction, and θ is the exit angle of the diffracted electron. Since the exit angle θ is small and the remanent magnetic field is large from the beginning as compared with that before diffraction, the electron cyclotron radius is small, and the remanent magnetic field is attenuated in the z-axis direction, so that the electron cyclotron radius gradually increases.
Under the action of the residual magnetic field, the cyclotron motion of the electrons causes the OAM field intensity distribution image acquired by the image receiving and acquiring module 300 to be distorted under the action of the residual magnetic field, i.e. to be radially dispersed or converged under the cylindrical coordinates, and the geometrical relationship between the positions before and after the electron distortion is shown in fig. 3, where point a and point a' are the position after the electron diffraction and the position on the fluorescent screen after the magnetic field distortion, respectively, and then the distance relationship between the two points and the electron cyclotron central axis O is:
wherein R and R' are the distances from the electron cyclotron center in the cylindrical coordinates before and after the distortion, respectively, R is the radius of the electron cyclotron motion, and β is the central angle of the electron cyclotron motion. Furthermore, the electron position before and after the distortion is deflected in the counterclockwise direction by an angle with respect to the position on the screen, and the deflection angle of the electron position before and after the distortion with respect to the electron cyclotron center axis O can be obtained by the following equation
α＝arcsin(2R[sin(β/2)] 2 /r' 2 ) (4)
For a single-mode pure OAM, the effect of the residual magnetic field on pure OAM vortex electrons (where l is 1 as an example) can be represented by a field intensity distribution diagram of OAM, where fig. 4(a) is the OAM field intensity distribution before being distorted by the residual magnetic field, fig. 4(b) is the OAM field intensity distribution diagram after being distorted by the residual magnetic field, and fig. 4(c) is the OAM field intensity distribution diagram after being affected by the residual magnetic field and then being distorted by divergence. The central hole of the distorted image becomes smaller or larger, so that modal blurring occurs between different pure OAM modes, and the receiving end cannot recognize the distorted image. Vortex electrons generated by the vortex electron generation module can be vortex electrons carrying any pure-state OAM and/or vortex electrons carrying mixed-state OAM, and the mixed-state OAM refers to combination of a plurality of pure-state OAMs. For example, the field intensity distribution diagrams of 1 pure OAM and 3 mixed OAM are shown below, where the OAM mode number l ═ {1, {1, -2}, {3, -3}, {1,2,3} }, where l ═ 1 is pure OAM, and l ═ {1, -2}, {3, -3}, and {1,2,3} are three mixed OAM modes, respectively.
Wherein fig. 5(a1) - (d1) are field intensity distributions of different electron beams before being distorted by the residual magnetic field, fig. 5(a2) - (d2) are field intensity distribution diagrams of different electron beams after being distorted by the residual magnetic field, and fig. 5(a3) - (d3) are field intensity distribution diagrams of different electron beams after being distorted by the residual magnetic field. In this example, as a receiving end of quantum state OAM wireless communication, in order to ensure a lower bit error rate, a transmitting end uses a mixed state OAM with a large difference between pure state and 3 field intensity distribution maps for data transmission, and its LG mode number l ═ 1, {1, -2}, {3, -3}, {1,2,3} } respectively represents a quadruple number 0-3, i.e. a communication system with four-step modulation.
In this embodiment, the training data preparation submodule generates distortion data of a plurality of residual magnetic field levels as training data, introduces position gaussian noise of different signal-to-noise levels, and the noise-containing data sets of the levels have the same ratio, as shown in fig. 6, which is an OAM field intensity distribution diagram of ± 3 modes when the noise-free and signal-to-noise ratios (SNRs) are 30, 20, and 15, respectively. To verify the generalization of the training model, the verification set data was modified with varying degrees of distortion data not present in the training set.
In this embodiment, the AI (artificial intelligence) model specifically uses a convolutional neural network, inputs an OAM field intensity distribution image with a data size of 128 × 128, and outputs a modality category corresponding to the OAM field intensity distribution image, where 4 modalities of this example correspond to 4 categories. Preferably, one-hot encoding is employed for the modality class.
The convolutional neural network structure adopted in this embodiment is shown in fig. 7, and includes an input layer, 5 convolutional pooling layers, 1 fully-connected layer, and an output layer, which are connected in sequence, where each convolutional pooling layer includes a convolutional layer and a pooling layer, which are connected in sequence, where the convolutional pooling layer performs feature learning and extraction on an OAM field intensity distribution image, and besides features such as shape, texture, color, and edge, features of nonlinear distortion introduced by a residual magnetic field, that is, features representing data sets specifically at a deeper layer, can be extracted mainly through massive data training. The hidden features are more efficient and accurate in expression of the data set, and the extracted abstract features are stronger in robustness, so that higher identification accuracy can be obtained. As shown in fig. 7, each convolutional layer contains a plurality of 5 × 5 convolutional kernels, for example, the first convolutional layer contains 20 5 × 5 convolutional kernels. Wherein the largest pooling layer comprises a2 x 2 pooling window.
The method adopts cross entropy to express the unfit degree of the convolutional neural network obtained by current training to the training data, namely a cross entropy loss function, and the expression is
Wherein X is input data of the neural network, namely OAM field intensity distribution image data, n is the number of input data samples, f (·) represents the OAM mode type output by the neural network, and Y represents the real OAM mode type.
Fig. 8 and 9 show the curves of the loss function value and the accuracy of the convolutional neural network according to the variation with the training times, and 1000 OAM field intensity distribution images are trained and verified, and the network parameters are updated to make the loss function obtain the minimum value as much as possible.
And solving the gradient of the loss function in the training process by adopting an adaptive moment estimation (Adam) optimizer, so that the loss function is continuously searched and iterated towards the minimum value by updating the network parameter value through the gradient. The batch processing size of the optimizer is 32, the accuracy of the artificial intelligence model after 30 times of training on the training set is 99.46%, and the accuracy of the verification set can reach 99.99%. For a conventional method, namely a general convolutional neural network, the influence of a residual magnetic field on OAM field intensity distribution is not considered, training parameters such as a network structure, a loss function and an optimization method which are the same as those of the example are adopted, but training data are samples without distortion, an AI model obtained by training and the AI model provided by the example are tested on the same test data set, and the recognition accuracy rates of the method of the example of the invention and the conventional method are respectively 95.75% and 69.75%, so that the recognition method of the example of the invention has obvious advantages in the recognition accuracy rate.
The foregoing is illustrated with a convolutional neural network as the AI model, and in fact, the AI model may also be a combination of one or more of convolutional neural network, decision tree, multi-classification Support Vector Machine (SVM) and K-nearest neighbor algorithm. Specifically, the models are combined and integrated by an ensemble learning technology and then are subjected to image recognition and classification, namely different weak classifiers are trained aiming at the same training set, and then the weak classifiers are integrated to form a stronger strong classifier serving as a final classifier, and common combination methods comprise boosting, adaboost and the like.
Although a specific embodiment of the present invention has been described above, it should be understood that the above-described embodiment is illustrative, and not restrictive, and that various changes, modifications, substitutions, and alterations can be made herein by those skilled in the art without departing from the scope of the invention.
Claims (8)
1. A vortex electronic modality identification system, comprising:
the vortex electron generation module is used for generating cyclotron electrons and coupling the cyclotron electrons with orbital angular momentum of received OAM electromagnetic waves to generate vortex electron beams carrying different modes of OAM;
the diffraction amplification module is used for diffracting and amplifying the size of the vortex electron beam;
the image receiving and collecting module is used for receiving the diffraction amplified vortex electron beam and collecting the vortex electron beam to obtain an OAM field intensity distribution image;
the data processing and identifying module is connected behind the image receiving and collecting module and used for receiving the OAM field intensity distribution image and performing data processing and OAM modal identification, and the data processing and identifying module comprises:
the training data preparation submodule is used for generating an OAM field intensity distribution image which is influenced by a residual magnetic field and contains nonlinear distortion as training data of the artificial intelligent model training submodule, wherein the residual magnetic field is a magnetic field which is generated by the vortex electron generation module and gradually attenuated from the diffraction amplification module to the image receiving and acquiring module;
the artificial intelligence model training submodule is used for constructing an artificial intelligence model and performing artificial intelligence model training by using the training data so as to obtain the artificial intelligence model with the highest verification and identification accuracy;
the mode identification submodule is used for carrying out real-time OAM mode identification on the OAM field intensity distribution image acquired by the image receiving and acquiring module by using the artificial intelligence model with the highest verification and identification accuracy, the training data comprise the OAM field intensity distribution image influenced by residual magnetic fields with different degrees, wherein the nonlinear distortion generated by the influence of the residual magnetic fields refers to one or more conditions of nonlinear divergence and nonlinear convergence of the OAM field intensity distribution image,
the nonlinear divergence and the nonlinear convergence refer to that the distances from the electron cyclotron center O to two points before and after the distortion of the residual magnetic field of the electrons satisfy the following relations:
wherein R and R' are distances of the electrons before and after the distortion from the electron cyclotron center O in the cylindrical coordinates, respectively, R is a radius of the electron cyclotron motion, β is a central angle of the electron cyclotron motion, and a deflection angle of the positions of the electrons before and after the distortion with respect to the electron cyclotron center O is obtained by the following equation:
α＝arcsin(2R[sin(β/2)] 2 /r' 2 )。
2. the vortex electronic mode identification system of claim 1, wherein a Laguerre-Gaussian vortex electron beam is used to simulate a vortex electron beam carrying different modes of OAM along the z-axis direction, and the expression is as follows:
wherein
Is the wave function of the eddy electron beam carrying the OAM,
is a constant determined by the number of radial vectors n and the number of OAM modes l,
is a generalized laguerre polynomial,
is the radius of the girdling, wherein
Is a reduced Planck constant, e is the element charge, B is the magnetic induction of the environment outside the electron beam, i is the imaginary unit,
is the three coordinate variables of the cylindrical coordinate system, k is the wave number.
3. The vortex electronic mode identification system according to claim 1, wherein the vortex electron beam generated by the vortex electron generation module comprises a vortex electron beam carrying pure OAM and/or a plurality of vortex electron beams of mixed-state OAM, wherein any mixed-state OAM is formed by combining a plurality of pure OAM.
4. A vortex electronic modality identification system according to claim 1,
the vortex electron generation module comprises a power supply, a gyrotron and a superconducting magnet, wherein the gyrotron comprises an electron gun and an electron cyclotron generation module, the superconducting magnet is located outside the electron cyclotron generation module, the power supply supplies power to the electron gun, electrons emitted and accelerated by the electron gun enter the electron cyclotron generation module and carry out cyclotron motion under the action of a magnetic field of the superconducting magnet, and OAM electromagnetic waves couple orbital angular momentum of the OAM electromagnetic waves to the cyclotron electrons, so that vortex electron beams carrying different modes of OAM are generated;
the diffraction amplification module comprises a poly-gold crystal film which is used for performing diffraction amplification on the vortex electron beam;
the image receiving and collecting module comprises a fluorescent screen and a camera, the fluorescent screen is used for receiving the vortex electron beams after diffraction amplification, and the camera is used for collecting images on the fluorescent screen as OAM field intensity distribution images.
5. A vortex electronic modality identification system according to claim 1, wherein the artificial intelligence model comprises a combination of one or more of a convolutional neural network, a decision tree, a multi-classification support vector machine and a K-nearest neighbor algorithm.
6. A vortex electronic mode identification method is characterized by comprising the following steps:
the method comprises the steps that a vortex electron generation module couples orbital angular momentum of an OAM electromagnetic wave to cyclotron electrons to generate vortex electron beams carrying different modes of OAM after receiving the OAM electromagnetic wave transmitted in a free space;
the diffraction amplification module is used for carrying out diffraction amplification on the vortex electron beam;
the image receiving and collecting module comprises a fluorescent screen and a camera, the fluorescent screen receives the diffraction amplified vortex electron beam to obtain an OAM field intensity distribution image, and the camera collects the OAM field intensity distribution image in real time and sends the OAM field intensity distribution image to the data processing and identifying module;
the data processing and identifying module utilizes the artificial intelligence model with the highest accuracy to identify the received OAM field intensity distribution image in real time to complete the OAM modal identification,
the training data preparation submodule of the data processing and recognition module performs binarization data processing on a plurality of OAM field intensity distribution images which are influenced by a residual magnetic field and contain nonlinear distortion and are received in advance, then the processed data are used as a part of model training data to be sent to the artificial intelligent model training submodule, the training data preparation submodule also generates a part of distorted and/or undistorted simulation OAM field intensity distribution images as another part of model training data through simulation, and the part of model training data and/or the other part of model training data jointly form training data,
the artificial intelligence model training sub-module utilizes the training data to carry out artificial intelligence model training, thereby obtaining an artificial intelligence model with the highest verification and identification accuracy,
the training data comprises OAM field intensity distribution images affected by different degrees of residual magnetic fields, wherein the nonlinear distortion caused by the residual magnetic fields refers to one or more of nonlinear divergence and nonlinear convergence of the OAM field intensity distribution images,
the nonlinear divergence and the nonlinear convergence refer to that the distances from the electron cyclotron center O to two points before and after the distortion of the residual magnetic field of the electrons satisfy the following relations:
wherein R and R' are distances of the electrons before and after the distortion from the electron cyclotron center O in the cylindrical coordinates, respectively, R is a radius of the electron cyclotron motion, β is a central angle of the electron cyclotron motion, and a deflection angle of the positions of the electrons before and after the distortion with respect to the electron cyclotron center O is obtained by the following equation:
α＝arcsin(2R[sin(β/2)] 2 /r' 2 )。
7. an OAM modality recognition apparatus, comprising:
the training data preparation submodule is used for generating an OAM field intensity distribution image which is influenced by a residual magnetic field and contains nonlinear distortion as training data of the artificial intelligent model training submodule, wherein the residual magnetic field is a magnetic field which is generated by the vortex electron generation module and gradually attenuated from the diffraction amplification module to the image receiving and acquiring module;
the artificial intelligence model training submodule is used for constructing an artificial intelligence model and performing artificial intelligence model training by using the training data so as to obtain the artificial intelligence model with the highest verification and identification accuracy;
the mode identification submodule is used for carrying out real-time OAM mode identification on the OAM field intensity distribution image acquired by the image receiving and acquiring module by using the artificial intelligence model with the highest verification and identification accuracy,
the training data comprises OAM field intensity distribution images affected by different degrees of residual magnetic fields, wherein the nonlinear distortion caused by the residual magnetic fields refers to one or more of nonlinear divergence and nonlinear convergence of the OAM field intensity distribution images,
the non-linear divergence and the non-linear convergence refer to that the distances between two points before and after the distortion of the residual magnetic field and an electron cyclotron center O satisfy the following relation:
wherein R and R' are distances of the electron from the electron cyclotron center O in the cylindrical coordinates before and after the distortion, respectively, R is a radius of the electron cyclotron motion, β is a central angle of the electron cyclotron motion, and a deflection angle of the position of the electron relative to the electron cyclotron center O before and after the distortion is obtained by the following formula:
α＝arcsin(2R[sin(β/2)] 2 /r' 2 )。
8. an electronic device, characterized in that the electronic device comprises:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of vortex electronic modality identification of claim 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110394918.9A CN113065494B (en) | 2021-04-13 | 2021-04-13 | Vortex electronic mode identification system, method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110394918.9A CN113065494B (en) | 2021-04-13 | 2021-04-13 | Vortex electronic mode identification system, method and device and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113065494A CN113065494A (en) | 2021-07-02 |
| CN113065494B true CN113065494B (en) | 2022-09-09 |
Family
ID=76566641
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110394918.9A Active CN113065494B (en) | 2021-04-13 | 2021-04-13 | Vortex electronic mode identification system, method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113065494B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113660762A (en) * | 2021-09-20 | 2021-11-16 | 三兄弟(珠海)科技有限公司 | Quantum state electromagnetic wave generating device for material detection |
| CN115015912B (en) * | 2022-02-24 | 2024-07-19 | 西安电子科技大学 | Vortex electromagnetic wave-based rotation target space angle estimation method |
| CN115276784B (en) * | 2022-07-26 | 2024-01-23 | 西安电子科技大学 | Deep learning-based orbital angular momentum modal identification method |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1149620C (en) * | 1994-01-14 | 2004-05-12 | 株式会社日立制作所 | Cathode ray tube and deflection aberration correcting method of the same |
| ITUB20152792A1 (en) * | 2015-08-04 | 2017-02-04 | Strand S R L | Optical demultiplexing and multiplexing device of a plurality of channels with different wavelength and different orbital angular momentum and relative optical communication system |
| CN110824433B (en) * | 2019-10-09 | 2021-11-23 | 清华大学 | Electromagnetic wave quantum state orbital angular momentum radar detection and method |
| CN111130612B (en) * | 2019-12-24 | 2021-07-09 | 清华大学 | Electromagnetic wave quantum state orbital angular momentum communication system and method |
| CN112465137B (en) * | 2020-10-26 | 2023-09-15 | 深圳技术大学 | Vortex beam mode identification system and method based on photonic neural network |
-
2021
- 2021-04-13 CN CN202110394918.9A patent/CN113065494B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113065494A (en) | 2021-07-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113065494B (en) | Vortex electronic mode identification system, method and device and electronic equipment | |
| Lin et al. | Threats of adversarial attacks in DNN-based modulation recognition | |
| CN109254274B (en) | Radar radiation source identification method based on feature fusion | |
| Mendis et al. | Deep learning based radio-signal identification with hardware design | |
| CN110598530A (en) | Small sample radio signal enhanced identification method based on ACGAN | |
| Wan et al. | Recognizing the HRRP by combining CNN and BiRNN with attention mechanism | |
| CN103425998B (en) | SAR target identification method under circumstance of occlusion | |
| CN106483514B (en) | Airplane motion mode identification method based on EEMD and support vector machine | |
| Golling et al. | Flow-enhanced transportation for anomaly detection | |
| CN110113288A (en) | A kind of design and demodulation method of the ofdm demodulator based on machine learning | |
| CN108449295A (en) | Combined modulation recognition methods based on RBM networks and BP neural network | |
| CN116680608B (en) | Signal modulation identification method based on complex graph convolutional neural network | |
| CN113378644B (en) | Method for defending signal modulation type recognition attack based on generation type countermeasure network | |
| Liu et al. | Radar emitter signal detection with convolutional neural network | |
| CN113962381A (en) | Weak signal data enhancement method | |
| Zhao et al. | Hierarchical equalization loss for long-tailed instance segmentation | |
| Xia et al. | Radar waveform recognition based on a two‐stream convolutional network and software defined radio | |
| Roche et al. | Nanosecond anomaly detection with decision trees for high energy physics and real-time application to exotic Higgs decays | |
| Zha et al. | Intelligent identification technology for high‐order digital modulation signals under low signal‐to‐noise ratio conditions | |
| Zhang et al. | Adversarial learning in transformer based neural network in radio signal classification | |
| Gai et al. | Spectrum sensing method based on residual cellular network | |
| Behera et al. | Image enhancement using accelerated particle swarm optimization | |
| CN114912482A (en) | Method and device for identifying radiation source | |
| Vaselli et al. | End-to-end simulation of particle physics events with Flow Matching and generator Oversampling | |
| Huang et al. | Automatic Recognition of Communication Signal Modulation Based on the Multiple‐Parallel Complex Convolutional Neural Network |
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 |