Multi-core optical fiber neuron and optical fiber sensing mode recognition system

Abstract

The invention belongs to the technical field of intelligent optical fiber devices, and particularly relates to a multi-core optical fiber neuron and optical fiber sensing mode recognition system which comprises a regulation and control pulse light module, an optical fiber sensor module, a seven-core optical fiber neuron module and a photoelectric detection module. The pulse light regulation and control module is used for regulating and controlling the combined continuous laser into pulse laser so as to obtain pulse light, and after feedback pulse light energy is coupled, the pulse light enters the optical fiber sensor module; the optical fiber sensor module is used for sensing external sensing information and then carrying out intensity modulation on incident pulsed light, and the intensity-modulated pulsed light with various wavelengths is combined and then incident into the seven-core optical fiber neuron module. According to the invention, the seven-core optical fiber can be used as the substrate to construct the seven-core optical fiber neuron so as to realize an all-optical mode identification task.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent optical fiber devices, and particularly relates to a multi-core fiber neuron and an optical fiber sensing pattern recognition system. Background Art

[0002] After more than 60 years of development, artificial intelligence has made important breakthroughs in algorithms, computing power, and computing materials. In recent years, it has developed from electronics-driven to photon systems, and neural networks based on optical means have become an inevitable trend.

[0003] Currently, most of the developed optical neural network systems are based on silicon-based waveguides, which have low coupling efficiency for light energy, complex manufacturing processes, and high implementation costs. In addition to the solutions based on silicon-based waveguide optical functional structures, another type of implementation method is to conduct functional design and implementation with active optical devices as the core. The cross-gain modulation based on semiconductor optical amplifiers can construct a LIF optical biological neuron model, the saturable absorption effect based on semiconductor excitatory lasers can realize the pulse release process of biological neurons, and the master-slave architecture based on vertical cavity surface lasers can realize the excitation transmission process of bionic neurons. However, the optical brain neuron solutions based on active optical devices are too complex, with high implementation costs and difficulties, and poor compatibility with current communication systems. Summary of the Invention

[0004] The purpose of the present invention is to provide a multi-core fiber neuron and an optical fiber sensing pattern recognition system, which can construct a seven-core fiber neuron based on a seven-core fiber and realize all-optical pattern recognition tasks.

[0005] The technical solution adopted by the present invention is as follows:

[0006] A multi-core fiber neuron and an optical fiber sensing pattern recognition system, including a regulated pulsed light module, an optical fiber sensor module, a seven-core fiber neuron module, and a photoelectric detection module;

[0007] The regulated pulsed light module is used to regulate the combined continuous laser into pulsed laser to obtain pulsed light, and couple the feedback pulsed light energy and then inject it into the optical fiber sensor module;

[0008] The optical fiber sensor module is used to sense external sensing information and then intensity-modulate the incident pulsed light. The pulsed light of each wavelength after intensity modulation is combined and then injected into the seven-core fiber neuron module;

[0009] The seven-core fiber neuron module is used to evenly distribute the light energy in the single-mode fiber to each core of the seven-core fiber, construct a wavelength division selection area and an evanescent field leakage area using femtosecond processing technology, and then combine the light energy in each core of the seven-core fiber into the single-mode fiber for regulation and inject it into the photoelectric detection module;

[0010] The photoelectric detection module is used to detect and interact with the light energy of each wavelength, and the output state of the seven-core fiber optic neuron.

[0011] The modulation pulse light module includes a plurality of tunable lasers, a first dense wavelength division multiplexer and a second dense wavelength division multiplexer, an intensity modulator, an electrical pulse generator, a first erbium-doped fiber amplifier, a second erbium-doped fiber amplifier, and an optical fiber coupler; the tunable laser is a wavelength tunable laser that emits continuous light of at least seven wavelengths.

[0012] A plurality of the tunable lasers are electrically connected to the first dense wavelength division multiplexer, the intensity modulator is electrically connected to the first dense wavelength division multiplexer, the electrical pulse generator is electrically connected to the intensity modulator, the first erbium-doped fiber amplifier is electrically connected between the intensity modulator and the optical fiber coupler, and the optical fiber coupler is electrically connected to the second dense wavelength division multiplexer.

[0013] The fiber optic sensor module includes a plurality of fiber optic sensors, a plurality of fiber optic circulators, and a dense wavelength division multiplexer; the fiber optic sensor is a fiber optic intensity sensor for sensing external sensing information and modulating the intensity of the incident pulse light.

[0014] A plurality of the fiber optic circulators are all electrically connected to the second dense wavelength division multiplexer, a plurality of the fiber optic circulators are all electrically connected to the dense wavelength division multiplexer, the number of fiber optic circulators is the same as the number of fiber optic sensors, and the fiber optic sensors are respectively connected to the fiber optic circulators.

[0015] The seven-core fiber optic neuron module includes a single-mode fiber-seven-core fiber tapered coupling structure, a phase-shifted fiber grating, a phase change material unit with six circumferential cores, and a phase change material unit with a central core; the single-mode fiber-seven-core fiber tapered coupling structure is prepared by welding a single-mode fiber and a seven-core fiber and using a fused taper mechanism for evenly distributing the energy of the single-mode fiber into the seven-core fiber.

[0016] The dense wavelength division multiplexer is connected to the seven-core fiber optic neuron module, the single-mode fiber-seven-core fiber tapered coupling structure includes a plurality of phase-shifted fiber gratings, and the phase-shifted fiber gratings and the phase change material units with six circumferential cores are both distributed in an annular array.

[0017] The photoelectric detection module includes an industrial computer, a data acquisition card, a plurality of photodetectors, a dense wavelength division multiplexer, and an optical fiber coupler;

[0018] The seven-core fiber optic neuron module is connected to the dense wavelength division multiplexer, a plurality of the photodetectors are all connected to the dense wavelength division multiplexer, a plurality of the photodetectors are all electrically connected to the data acquisition card, and the industrial computer is electrically connected to the data acquisition card.

[0019] The phase change material unit is a chalcogenide compound, containing at least two or more elements among Ge, Sb, and Te.

[0020] The photoelectric detection module includes an industrial control computer, a data acquisition card, multiple photodetectors, a dense wavelength division multiplexer, and an optical fiber coupler;

[0021] The seven-core optical fiber neuron module is connected to the dense wavelength division multiplexer, multiple of the photodetectors are all connected to the dense wavelength division multiplexer, multiple of the photodetectors are all electrically connected to the data acquisition card, and the industrial control computer is electrically connected to the data acquisition card.

[0022] The technical effects achieved by the present invention are as follows:

[0023] A multi-core optical fiber neuron and optical fiber sensing pattern recognition system of the present invention can be well compatible with optical fiber communication and optical fiber sensing systems, and has many advantages such as strong signal anti-interference ability, low loss, high speed, high bandwidth, and scalability.

[0024] A multi-core optical fiber neuron and optical fiber sensing pattern recognition system of the present invention uses photons as the driving force to realize the basic structure and function of artificial neurons. The entire recognition process is executed at the speed of light, and has a faster operation speed and lower energy consumption compared to the solution driven by electrons.

[0025] A multi-core optical fiber neuron and optical fiber sensing pattern recognition system of the present invention Description of the Drawings

[0026] Figure 1 is a schematic structural diagram of an embodiment of the present invention;

[0027] Figure 2 is a schematic diagram of a gesture recognition task of an embodiment of the present invention;

[0028] Among them: a, gesture bending sensor; b, schematic diagram of the optical fiber neuron principle; c, output of the optical fiber neuron gesture pattern recognition. Detailed Embodiments

[0029] In order to make the purpose and advantages of the present invention clearer, the following specifically describes the present invention with reference to embodiments. It should be understood that the following text only describes one or several specific implementation manners of the present invention, and does not strictly limit the scope of protection specifically claimed by the present invention.

[0030] As Figure 1 - Figure 2 shown, a multi-core optical fiber neuron and optical fiber sensing pattern recognition system includes a regulated pulsed light module 1, an optical fiber sensor module 2, a seven-core optical fiber neuron module 3, and a photoelectric detection module 4;

[0031] The modulation pulse light module 1 is used to modulate the combined continuous laser into pulsed laser, so as to obtain pulsed light, and after coupling the feedback pulsed light energy, it is incident into the fiber optic sensor module 2;

[0032] The fiber optic sensor module 2 is used to sense the external sensing information and then intensity-modulate the incident pulsed light. After the intensity-modulated pulsed light of each wavelength is combined, it is incident into the seven-core fiber optic neuron module 3;

[0033] The seven-core fiber optic neuron module 3 is used to evenly distribute the optical energy in the single-mode fiber to each core of the seven-core fiber, construct a wavelength division selection area and an evanescent field leakage area by using femtosecond processing technology, and then combine the optical energy in each core of the seven-core fiber into the single-mode fiber for modulation and incident into the photoelectric detection module 4;

[0034] The photoelectric detection module 4 is used to detect and interact with the optical energy of each wavelength, and the output state of the seven-core fiber optic neuron.

[0035] The modulation pulse light module 1 includes multiple tunable lasers 101-107, a first dense wavelength division multiplexer 108, a second dense wavelength division multiplexer 114, an intensity modulator 109, an electrical pulse generator 110, a first erbium-doped fiber amplifier 111, a second erbium-doped fiber amplifier 112, and a fiber optic coupler 113; the tunable lasers 101-107 are wavelength tunable lasers that emit at least seven wavelengths of continuous light.

[0036] As Figure 1 shown, among them, the multiple tunable lasers are divided into a tunable laser 101, a tunable laser 102, a tunable laser 103, a tunable laser 104, a tunable laser 105, a tunable laser 106, and a tunable laser 107;

[0037] The multiple tunable lasers are electrically connected to the first dense wavelength division multiplexer 108, the intensity modulator 109 is electrically connected to the first dense wavelength division multiplexer 108, the electrical pulse generator 110 is electrically connected to the intensity modulator 109, the first erbium-doped fiber amplifier 111 is electrically connected between the intensity modulator 109 and the fiber optic coupler 113, and the fiber optic coupler 113 is electrically connected to the second dense wavelength division multiplexer 114.

[0038] The tunable lasers 101 - 107 in the modulation pulse light module 1 emit continuous lasers with different wavelengths. After being multiplexed and combined by the dense wavelength division multiplexer 108, they are modulated by the intensity modulator 109. The electrical pulse generator 110 provides a pulse modulation signal to modulate the combined continuous laser into a pulsed laser. After passing through the first erbium-doped fiber amplifier 111, pulsed light with sufficient energy is obtained. The pulsed light energy of the feedback is coupled through the fiber coupler 113, demultiplexed by the dense wavelength division multiplexer 114, and then incident into the fiber sensor module 2.

[0039] The intensity modulator 109, the electrical pulse generator 110, and the first erbium-doped fiber amplifier 111 are used to generate pulsed light with sufficient intensity and different pulse widths. Among them, the narrow pulse width pulsed light is used to initialize the state of the weights or soma units in the seven-core fiber neuron module 3, and the wide pulse width pulsed light is used to modulate the state of the weights or soma units in the seven-core fiber neuron module 3.

[0040] The fiber sensor module 2 includes multiple fiber sensors, multiple fiber circulators, and a dense wavelength division multiplexer 213; the fiber sensors are fiber intensity sensors, which are used to sense external sensing information and perform intensity modulation on the incident pulsed light, and the design and selection of the fiber intensity sensors can be carried out according to the specific tasks of fiber sensing pattern recognition.

[0041] As Figure 2 shown in a, after polymer encapsulation, each of the fiber sensors 201 - 206 is coupled with a finger. When the finger is in different bending states, the pulsed light of each wavelength emitted from the fiber sensors 201 - 206 has different intensity states. Among them, each wavelength represents the information of a finger, and for the gesture recognition task, five cores in the circumference of the seven-core fiber neuron 3 are used.

[0042] Multiple fiber circulators are all electrically connected to the second dense wavelength division multiplexer 114, and multiple fiber circulators are all electrically connected to the dense wavelength division multiplexer 213. The number of fiber circulators is the same as that of the fiber sensors, and the fiber sensors are respectively connected to the fiber circulators.

[0043] Among them, the multiple fiber sensors are divided into fiber sensor 201, fiber sensor 202, fiber sensor 203, fiber sensor 204, fiber sensor 205, and fiber sensor 206;

[0044] The multiple fiber circulators are divided into fiber circulator 207, fiber circulator 208, fiber circulator 209, fiber circulator 210, fiber circulator 211, and fiber circulator 212.

[0045] Pulses of light of six wavelengths pass through the fiber optic circulators 207 - 212 on their respective optical paths and then enter each of the fiber optic sensors 201 - 206. After each of the fiber optic sensors 201 - 206 senses the external sensing information, it intensity - modulates the incident pulsed light. The fiber optic sensors 201 - 206 can be regarded as reflective fiber optic intensity sensors. After the intensity - modulated pulsed light of each wavelength passes through the fiber optic circulators 207 - 212 again, it enters the dense wavelength division multiplexer 213 together with the pulsed light of the seventh wavelength for beam combination, and then enters the seven - core fiber neuron module 3.

[0046] The seven - core fiber neuron module 3 includes a single - mode fiber - seven - core fiber tapered coupling structure 301, a phase - shifted fiber grating, a phase - change material unit of six circumferential cores, and a phase - change material unit 316 of the central core; the single - mode fiber - seven - core fiber tapered coupling structure is prepared by welding a single - mode fiber and a seven - core fiber and then using a fused tapering mechanism, and is used to evenly distribute the energy of the single - mode fiber into the seven - core fiber.

[0047] The dense wavelength division multiplexer 213 is connected to the seven - core fiber neuron module 3. The single - mode fiber - seven - core fiber tapered coupling structure 301 includes multiple phase - shifted fiber gratings, and the phase - shifted fiber gratings and the phase - change material unit of six circumferential cores are both distributed in an annular array.

[0048] Among them, the phase - shifted fiber gratings are divided into phase - shifted fiber grating 302, phase - shifted fiber grating 303, phase - shifted fiber grating 304, phase - shifted fiber grating 305, phase - shifted fiber grating 306, phase - shifted fiber grating 307, and phase - shifted fiber grating 308;

[0049] The phase - change material units are divided into phase - change material unit 309, phase - change material unit 310, phase - change material unit 311, phase - change material unit 312, phase - change material unit 313, and phase - change material unit 314.

[0050] The seven-core fiber optic neuron module 3 is based on a seven-core optical fiber. A single-mode optical fiber is fused and tapered with the seven-core optical fiber. After the pulsed light of each wavelength combined by the fiber optic sensor module 2 is incident, the optical energy in the single-mode optical fiber is evenly distributed to each core of the seven-core optical fiber through the single-mode fiber-seven-core fiber tapered coupling structure 301. The femtosecond processing technology is used to construct a wavelength division region and an evanescent field leakage region. The phase-shifted fiber gratings 302-308 in the six circumferential cores only allow one wavelength of pulsed light to pass through, and then pass through the phase change material units 309-314 of the six circumferential cores respectively, that is, the weight units. Among them, the high-energy narrow pulse width pulsed light causes the phase change material to be amorphized, and at this time the transmittance of the core becomes smaller. The low-energy wide pulse width pulsed light causes the phase change material to crystallize, and at this time the transmittance of the core becomes larger. After being regulated by the phase change material units 309-314 of the six surrounding cores, the optical energy in each core of the seven-core optical fiber is combined into the single-mode optical fiber through the single-mode fiber-seven-core fiber tapered coupling structure 315 and is incident into the phase change material unit 316 in the central core, that is, the cell body unit. After the combined pulsed light regulates the cell body unit, it is incident into the photoelectric detection module 4.

[0051] Among them, the phase-shifted fiber gratings 302-308 are written into the cores in different focal planes in sequence through the femtosecond direct writing process, and the central wavelengths of their transmission windows are different;

[0052] The photoelectric detection module 4 includes an industrial control computer 401, a data acquisition card 402, multiple photodetectors, a dense wavelength division multiplexer 410, and an optical fiber coupler 411;

[0053] The seven-core fiber optic neuron module 3 is connected to the dense wavelength division multiplexer 410. Multiple photodetectors are all connected to the dense wavelength division multiplexer 410. Multiple photodetectors are all electrically connected to the data acquisition card 402. The industrial control computer 401 is electrically connected to the data acquisition card 402.

[0054] As Figure 1 shown, among them, the photodetectors are divided into a photodetector 403, a photodetector 404, a photodetector 405, a photodetector 406, a photodetector 407, a photodetector 408, and a photodetector 409.

[0055] After the combined pulsed light is split by the fiber coupler 411, a part of the pulsed light energy is injected into the regulated pulsed light module 1. After being amplified by the second erbium-doped fiber amplifier 112, it is fed back through the fiber coupler 113. Another part of the pulsed light energy is demultiplexed by the dense wavelength division multiplexer 410, and the photodetectors 403 - 409 detect the light energy of each wavelength. The industrial control computer 401 interacts with the data acquisition card 402, and collects and records the output signals of the photodetectors 403 - 409, records the states of each weight unit and the output state of the seven-core fiber neuron;

[0056] The seven-core fiber neuron module 3 is used to identify the state patterns of the fiber sensors 201 - 206 in the fiber sensor module 2. As Figure 2 shown in b, for the state patterns of the target fiber sensors 201 - 206 to be learned, the regulated pulsed light module 1 needs to emit narrow pulse width pulsed light with wavelengths of λ1 to λ7, and initialize the states of each weight and the cell body to a low transmittance state; under supervised learning, the regulated pulsed light module 1 emits wide pulse width pulsed light with wavelengths of λ1 to λ6 and an energy distribution identical to the target pattern, which is incident into the seven-core fiber neuron module 3 to adjust the weights. After being combined by the single-mode fiber - seven-core fiber tapered coupling structure 315, the state of the cell body unit is regulated to achieve output. At this time, it indicates that the supervised learning process is completed, and there is no need to use the feedback loop in this state; under unsupervised learning, the fiber sensors 201 - 206 need to be in this state pattern first. The regulated pulsed light module 1 emits wide pulse width pulsed light with the same energy of wavelengths of λ1 to λ6. After being regulated by the fiber sensors 201 - 206 into a pulsed light sequence with an energy distribution of the target pattern, it is incident into the seven-core fiber neuron module 3 to adjust the weights. After being combined by the single-mode fiber - seven-core fiber tapered coupling structure 315, the state of the cell body unit is regulated. The light output by the seven-core fiber neuron module 3 passes through the feedback loop, that is, the fiber coupler 411, the second erbium-doped fiber amplifier 112, and the fiber coupler 113, and then enters the fiber sensor module 2 and the seven-core fiber neuron module 3 again. Each adjustment of the weight and the cell body state represents a training process. Through feedback, multiple training processes are achieved. The weights and the cell body state approach the target pattern after multiple trainings. At this time, it indicates that the unsupervised learning process is completed; after the learning process is completed, the seven-core fiber neuron module 3 can perform all-optical identification of the state patterns of the fiber sensors 201 - 206 in the fiber sensor module 2. Only when the state patterns of the fiber sensors 201 - 206 match the target pattern, the seven-core fiber neuron module 3 has a high output. The state patterns of the other fiber sensors 201 - 206 are all low outputs. That is, the target state of the seven-core fiber neuron 3 - 1 is the gesture "3". When the gesture is in "3", the seven-core fiber neuron 3 - 1 has a high output. For other gestures, the output of the seven-core fiber neuron 3 - 1 is low. AsFigure 2 as shown in c.

[0057] Specifically, for the phase change material units 309 - 314 of the six core fibers on the circumference and the phase change material unit 316 of the central core fiber, by using the femtosecond subtractive process, holes are drilled in the claddings of the six core fibers on the circumference and the cladding of the single-mode fiber until the evanescent field leakage state is reached, and then a phase change material thin film and an anti-oxidation layer thin film are successively deposited by using the radio frequency magnetron sputtering process. The phase change material units 309 - 314 are chalcogenide compounds, containing at least two or more elements among Ge, Sb, and Te. For example, germanium antimony telluride alloy Ge2Sb2Te5, silver indium antimony telluride alloy AgInSbTe, etc. This phase change material thin film has a crystalline state, an amorphous state, and an intermediate state between the crystalline state and the amorphous state, and there are differences in transmittance under different phase states;

[0058] The fiber coupler 411, the erbium-doped fiber amplifier 112, and the fiber coupler 113 form a feedback loop under unsupervised learning and are in a non-operating state under supervised learning;

[0059] The photodetectors 403 - 409 are used to detect the states of the phase change material units 309 - 314 of the six core fibers on the circumference, that is, the states of the weight units;

[0060] The photodetectors 403 - 409 are used to detect the state of the phase change material unit 316 of the central core fiber, that is, the state of the cell body unit. When it is in a high output state, it indicates that the seven-core fiber neuron module 3 is in an effective recognition output;

[0061] As Figure 2 shown, this solution uses a seven-core fiber as the substrate, and uses femtosecond processing technology to construct a seven-core fiber neuron 3, and then conducts pattern learning and recognition on the states of the respective fiber sensors 201 - 206 of the fiber sensor module 2. The signal transmission and weight adjustment processes in the entire pattern learning process and recognition process are completed by the pulsed light provided by the control pulsed light module 1, and the states of each weight and the output state in the seven-core fiber neuron are detected by the photoelectric detection module 4.

[0062] The above is only the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and retouches can be made, and these improvements and retouches should also be regarded as the protection scope of the present invention. The structures, devices, and operation methods not specifically described and explained in the present invention are implemented according to the conventional means in the art without special instructions and limitations.

Claims

1. A multi-core fiber optic neuron and a fiber optic sensing pattern recognition system, characterized in that: It includes a control pulsed light module (1), an optical fiber sensor module (2), a seven-core optical fiber neuron module (3), and an optoelectronic detection module (4); The control pulsed light module (1) is used to modulate the combined continuous laser into pulsed laser to obtain pulsed light, and after coupling the feedback pulsed light energy, it is incident into the optical fiber sensor module (2); The optical fiber sensor module (2) is used to sense the external sensing information and then intensity-modulate the incident pulsed light. After the pulsed light of each wavelength after intensity modulation is combined, it is incident into the seven-core optical fiber neuron module (3); The seven-core optical fiber neuron module (3) is used to evenly distribute the light energy in the single-mode optical fiber to each core of the seven-core optical fiber, construct a wavelength division selection area and an evanescent field leakage area by using femtosecond processing technology, and then combine the light energy in each core of the seven-core optical fiber into the single-mode optical fiber for modulation and incident into the optoelectronic detection module (4); The optoelectronic detection module (4) is used to detect and interact with the light energy of each wavelength and the output state of the seven-core optical fiber neuron.

2. The multi-core fiber optic neuron and fiber optic sensing pattern recognition system according to claim 1, wherein: The control pulsed light module (1) includes a plurality of tunable lasers, a first dense wavelength division multiplexer (108) and a second dense wavelength division multiplexer (114), an intensity modulator (109), an electrical pulse generator (110), a first erbium-doped fiber amplifier (111), a second erbium-doped fiber amplifier (112), and an optical fiber coupler (113); A plurality of the tunable lasers are electrically connected to the first dense wavelength division multiplexer (108), the intensity modulator (109) is electrically connected to the first dense wavelength division multiplexer (108), the electrical pulse generator (110) is electrically connected to the intensity modulator (109), the first erbium-doped fiber amplifier (111) is electrically connected between the intensity modulator (109) and the optical fiber coupler (113), and the optical fiber coupler (113) is electrically connected to the second dense wavelength division multiplexer (114).

3. A multi-core fiber optic neuron and fiber optic sensing pattern recognition system according to claim 2, characterized in that: The optical fiber sensor module (2) includes a plurality of optical fiber sensors, a plurality of optical fiber circulators, and a dense wavelength division multiplexer (213); A plurality of the optical fiber circulators are all electrically connected to the second dense wavelength division multiplexer (114), a plurality of the optical fiber circulators are all electrically connected to the dense wavelength division multiplexer (213), the number of the optical fiber circulators is the same as that of the optical fiber sensors, and the optical fiber sensors are respectively connected to the optical fiber circulators.

4. A multi-core fiber optic neuron and fiber optic sensing pattern recognition system according to claim 3, characterized in that: The seven-core optical fiber neuron module (3) includes a single-mode optical fiber-seven-core optical fiber tapered coupling structure (301), a phase-shifted fiber grating, a phase change material unit of six circumferential cores, and a phase change material unit of the central core (316); The dense wavelength division multiplexer (213) is connected to the seven-core optical fiber neuron module (3), the single-mode optical fiber-seven-core optical fiber tapered coupling structure (301) includes a plurality of phase-shifted fiber gratings, and the phase-shifted fiber gratings and the phase change material unit of six circumferential cores are both distributed in an annular array.

5. A multi-core fiber optic neuron and fiber optic sensing pattern recognition system according to claim 4, characterized in that: The optoelectronic detection module (4) includes an industrial control computer (401), a data acquisition card (402), a plurality of photodetectors, a dense wavelength division multiplexer (410), and an optical fiber coupler (411); The seven-core fiber optic neuron module (3) is connected to the dense wavelength division multiplexer (410), and multiple photodetectors are all connected to the dense wavelength division multiplexer (410). Multiple photodetectors are all electrically connected to the data acquisition card (402), and the industrial control computer (401) is electrically connected to the data acquisition card (402).

6. A multi-core fiber optic neuron and fiber optic sensing pattern recognition system according to claim 2, characterized in that: The tunable laser is a wavelength tunable laser that emits continuous light of at least seven wavelengths.

7. A multi-core fiber optic neuron and fiber optic sensing pattern recognition system according to claim 1, characterized in that: The fiber optic sensor is a fiber optic intensity sensor that is used to sense external sensing information and perform intensity modulation on the incident pulsed light.

8. The multi-core fiber optic neuron and fiber optic sensing pattern recognition system according to claim 1, characterized in that: The single-mode fiber-seven-core fiber tapered coupling structure is prepared by using a fused taper mechanism after welding the single-mode fiber and the seven-core fiber, and is used to evenly distribute the energy of the single-mode fiber into the seven-core fiber.

9. A multi-core fiber optic neuron and fiber optic sensing pattern recognition system according to claim 4, characterized in that: The phase change material unit (316) is a chalcogenide compound containing at least two or more elements among Ge, Sb, and Te.

Images