Micro-ring resonant cavity wavelength converter for artificial neural network

Abstract

The invention discloses an MRR wavelength conversion method and device applied to an artificial neural network. The MRR wavelength conversion device comprises a substrate component; the SOA optical amplifier is integrated on the substrate component and is used for emitting incident light to lasing an internal light source; and the micro-ring resonant cavity silicon-based chip is integrated on the substrate component and is used for carrying out coupling processing based on the incident light and the internal light source and determining target emitted light. According to the invention, integration of receiving, processing and transmitting of signals is realized, so that the signals can be visually processed. In addition, instant wavelength conversion in a full wavelength range can be realized, and the wavelength conversion device is used in an artificial neural network and has the technical effects of simple structure, small size, low power consumption and high speed.

Description

Technical Field

[0001] The present application relates to the field of optical technology, and in particular to a micro-ring resonant cavity wavelength converter for artificial neural networks. Background Art

[0002] The function of a wavelength converter is to convert an optical signal from one wavelength to another. Wavelength converters are mostly used in WDM networks to convert several signals of the same wavelength to different wavelengths before outputting them, thereby avoiding wavelength contention and blocking between several channels of the same wavelength, saving resources (optical fiber, node scale and wavelength), simplifying network management and reducing the complexity of network interconnection. Traditional wavelength converters can be divided into optical-electrical-optical regeneration type, gain saturation type, phase modulation type and four-wave mixing type according to their working principles.

[0003] In addition, some wavelength converters can also be used as optical-optical nonlinear neuron devices in artificial neural networks (ANN). Artificial neural networks are a data processing method that simulates the structure and function of biological neural networks to handle complex problems. The nonlinear characteristics of artificial neural networks refer to the nonlinear relationship between the feedback signal x2 and the received signal x1 of a neuron (e.g. Figure 2 a) in the figure: x2 = 0 (x1 ≤ b / k), x2 = kx1-b (x1 > b / k), that is, when the received signal x1 reaches a certain value b / k, the feedback signal x2 begins to appear; this relationship is called an activation function in artificial neural networks. Nonlinear characteristics and activation functions are the basis of the work of artificial neural networks. Artificial neural networks use nonlinear characteristics to feedback various received signals, and learn the appropriate k and b values ​​from a large number of exercises, so that they can appropriately process various received signals; nonlinear characteristics enable artificial neural networks to have the characteristics of autonomous learning, parallel computing, and strong robustness.

[0004] In order to physically realize artificial neural networks, it is necessary to find a series of nonlinear neuron devices to construct nonlinear relationships. Common nonlinear devices are divided into electric-electric devices (such as diodes and field effect transistors), optical-electric-optical devices (such as optical reporters and optical switches), and optical-optical devices (such as optical-optical diodes and optical threshold switches). For the development of artificial neural networks, people are continuously looking for and developing optical-optical nonlinear neuron devices with simple structures, rapid feedback, and low energy consumption.

[0005] Some wavelength converters can provide nonlinear effects, that is, the device's transmitted signal λ2 and received signal λ1 satisfy the relationship: λ2=0 (λ1≤b / k), λ2=kλ1+b (λ1>b / k), so they can be used as optical-optical nonlinear neuron devices. Summary of the invention

[0006] The present invention proposes for the first time a wavelength converter based on a completely new principle: a microring resonant cavity wavelength converter, which is particularly suitable for use as an optical-optical nonlinear neuron device in an artificial neural network. The principle of the wavelength converter is based on the periodic coupling filtering characteristics of the microring resonant cavity to the laser wavelength. With the help of SOA as an internal auxiliary light source, when the external signal (wavelength λ1) reaches a certain light intensity, the SOA is excited (wavelength range λSOA), and λSOA is filtered by MRR to emit a laser with a wavelength of λ2. Thus, the external signal λ1 is converted to a distant wavelength λ2 (λ2>λ1), realizing instant wavelength conversion within the full wavelength range. The feedback signal λ2 of the wavelength converter and the received signal λ1 have a nonlinear relationship; λ2=0(λ1≤b / k), λ2=kλ1+b(λ1>b / k); therefore, the nonlinear effect of the optical signal can be realized, and it can be used as an optical-optical nonlinear neuron device in an artificial neural network. The microring resonant cavity wavelength converter is used in artificial neural networks, and has the advantages of simple structure, small size, low power consumption and high speed. In a first aspect, the present application provides a micro-ring resonator wavelength converter for an artificial neural network, the wavelength converter comprising:

[0007] Base plate components;

[0008] An SOA optical amplifier, integrated on the substrate component, for lasing an internal light source according to the incident light acquired;

[0009] A micro-ring resonant cavity silicon-based chip is integrated on the substrate component and is used to perform coupling processing based on the incident light and the internal light source to determine the target emitted light.

[0010] In one embodiment, the microring resonator silicon-based chip comprises:

[0011] A waveguide assembly for guiding the incident light and the internal light source;

[0012] A micro-ring resonant cavity, used for performing coupling processing according to the incident light to determine the incident light after coupling, and exciting the SOA optical amplifier to emit the internal light source according to the incident light after coupling, and performing coupling processing according to the internal light source to determine the internal light source after coupling;

[0013] The spectrometer is used to perform spectroscopic processing based on the coupled internal light source to obtain the target emission light.

[0014] In one embodiment, the ratio of PBW / FSR is 0.5-0.8, wherein PBW is the peak-to-base width of the SOA optical amplifier, and FSR is the free spectrum range of the microring resonant cavity.

[0015] In one embodiment, the splitting ratio of the beam splitter is 1:9, wherein the coupled internal light source is subjected to beam splitting processing to obtain 90% of the light intensity as the target emission light.

[0016] In one embodiment, the waveguide assembly includes a curved waveguide, a first straight waveguide, a second straight waveguide, and a third straight waveguide;

[0017] The curved waveguide is used to guide the target emitted light to the first straight waveguide;

[0018] The first straight waveguide is used to guide the target emission light guided by the curved waveguide and emit the target emission light;

[0019] The second straight waveguide is used to transmit the incident light to the microring resonant cavity, and transmit the internal light source coupled by the microring resonant cavity to the spectrometer;

[0020] The third straight waveguide is used to transmit the incident light coupled by the microring resonant cavity to the SOA optical amplifier, and transmit the internal light source emitted by the SOA optical amplifier to the microring resonant cavity.

[0021] In one embodiment, an anti-reflection film is disposed at the input end of the third straight waveguide, and a high-reflection film is disposed at the output end of the third straight waveguide.

[0022] In one embodiment, a magneto-optical isolator is provided between the optical path for acquiring the incident light and the optical splitter.

[0023] In one embodiment, a thermal tuner is provided on the microring resonant cavity, and the thermal tuner is in the shape of two arcs surrounding the microring resonant cavity;

[0024] The thermal tuner is used to adjust the temperature of the microring resonant cavity through the positive and negative electrodes of the thermal tuner to tune the resonance peak of the microring resonant cavity.

[0025] In one embodiment, the microring resonator silicon-based chip further comprises:

[0026] The spot converter is arranged on one side of the SOA optical amplifier to couple the incident light and the internal light source to the micro-ring resonator silicon-based chip, and is arranged on one side of the waveguide component to couple out the laser light conducted by the waveguide component.

[0027] A second aspect of the present application provides a method for using a micro-ring resonant cavity wavelength converter for an artificial neural network, the method being applied to the above-mentioned micro-ring resonant cavity wavelength converter for an artificial neural network, the method comprising:

[0028] Obtain incident light;

[0029] Based on the incident light, a pre-tuned micro-ring resonator silicon-based chip is used for coupling processing to determine the incident light after coupling;

[0030] Based on the coupled incident light, an internal light source emitted by the SOA optical amplifier is excited and obtained;

[0031] Based on the internal light source, a pre-tuned micro-ring resonator silicon-based chip is used for coupling processing to determine the internal light source after coupling;

[0032] The target emission light is obtained by performing spectroscopic processing based on the coupled internal light source.

[0033] A micro-ring resonant cavity wavelength converter for an artificial neural network provided in an embodiment of the present application has at least the following technical effects.

[0034] This application is a wavelength converter based on a completely new principle: a microring resonant cavity wavelength converter, and is particularly suitable for use as an optical-optical nonlinear neuron device in an artificial neural network. This wavelength converter can achieve instant wavelength conversion within the entire wavelength range, and is used in artificial neural networks, with the technical effects of simple structure, small size, low power consumption, and high speed. In addition, this application is the first to use the periodic characteristics of a microring resonant cavity silicon-based chip for wavelength conversion, which can achieve long-distance wavelength transfer. Since the structure of the device is essentially an optical chip, it is small, compact, and simple. .

[0035] Details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0037] Figure 1 A schematic diagram of the structure of a micro-ring resonant cavity wavelength converter for an artificial neural network provided in an embodiment of the present application;

[0038] Figure 2 The nonlinear effects of optical-optical nonlinear neuron components;

[0039] Figure 3 A schematic diagram of the optical path of a micro-ring resonator wavelength converter for an artificial neural network provided in the modulation step 1 of the present application;

[0040] Figure 4 A schematic diagram of each light path exit provided for the modulation step 1;

[0041] Figure 5 A schematic diagram of the optical path of a micro-ring resonator wavelength converter for an artificial neural network provided in the modulation step 2 of the present application;

[0042] Figure 6 A schematic diagram of each light path exit provided for the modulation step 2;

[0043] Figure 7 A schematic diagram of the optical path of a micro-ring resonator wavelength converter for an artificial neural network provided in modulation step three of the present application;

[0044] Figure 8 Schematic diagram of the various light path exits provided for modulation step three.

[0045] Description of Reference Numerals

[0046] 1-substrate component, 2-SOA optical amplifier, 3-microring resonator silicon-based chip, 31-waveguide component, 311-bent waveguide, 312-first straight waveguide, 313 second straight waveguide, 314-third straight waveguide, 32-microring resonator, 321-thermal tuner, 3211-positive electrode of thermal tuner, 3212-negative electrode of thermal tuner, 4-first optical path entrance, 5-second optical path entrance, 6-first optical path exit, 7-second optical path exit, 8-third optical path exit, 9-DFB laser. DETAILED DESCRIPTION

[0047] The following description of at least one exemplary embodiment is actually only illustrative and is by no means intended to limit the present application and its application or use. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0048] Obviously, the drawings described below are only some examples or embodiments of the present application. For ordinary technicians in this field, the present application can also be applied to other similar scenarios based on these drawings without creative work. In addition, it can also be understood that although the efforts made in this development process may be complicated and lengthy, for ordinary technicians in this field related to the content disclosed in this application, some changes in design, manufacturing or production based on the technical content disclosed in this application are just conventional technical means, and should not be understood as insufficient content disclosed in this application.

[0049] Reference to "embodiment" in this application means that the specific features, structures or characteristics described in conjunction with the embodiment may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those of ordinary skill in the art that the embodiments described in this application can be combined with other embodiments without conflict.

[0050] Unless otherwise defined, the technical terms or scientific terms involved in this application should be understood by people with ordinary skills in the technical field to which this application belongs. The words "one", "a", "a", "the" and the like involved in this application do not indicate a quantitative limitation, and may represent the singular or plural. The terms "include", "comprise", "have" and any of their variations involved in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there may be three relationships, for example, "A and / or B" can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects before and after are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0051] An embodiment of the present application provides a micro-ring resonant cavity wavelength converter for an artificial neural network.

[0052] In a first aspect, an embodiment of the present application provides a micro-ring resonator wavelength converter for an artificial neural network. Figure 1 A schematic diagram of the structure of a micro-ring resonant cavity wavelength converter for an artificial neural network provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, the wavelength converter comprises:

[0053] The substrate component 1 provides a platform for integrating and combining the SOA optical amplifier 2 and the micro-ring resonant cavity silicon-based chip 3. The SOA optical amplifier 2 is integrated on the substrate component 1 and is used to laser the internal light source according to the incident light obtained by the first optical path entrance 4. The micro-ring resonant cavity silicon-based chip 3 is integrated on the substrate component 1 and is used to perform coupling processing based on the incident light and the internal light source to determine the target emission light. The micro-ring resonant cavity silicon-based chip 3 uses the waveguide characteristics and resonance effect of light to achieve light coupling and processing in a small space.

[0054] Preferably, the external light source is a distributed feedback laser (DFB), which is a semiconductor laser with a periodic diffraction grating. This grating structure generates feedback in the cavity of the DFB laser, so that the DFB laser only emits monochromatic light of a specific wavelength, and can provide stable and narrow-bandwidth light output. SOA optical amplifier 2, that is, semiconductor optical amplifier (SOA), is a semiconductor optical amplifier that can amplify and modulate optical signals by injecting current.

[0055] The micro-ring resonant cavity silicon-based chip 3 includes: a waveguide component 31, which is used to conduct the incident light and the internal light source; a micro-ring resonant cavity 32, which is used to perform coupling processing according to the incident light to determine the incident light after coupling, and to excite the SOA optical amplifier 2 to emit the internal light source through the second optical path entrance 5 according to the incident light after coupling, and to determine the internal light source after coupling according to the internal light source; a spectrometer, which is used to perform spectroscopic processing based on the coupled internal light source to obtain the target emission light.

[0056] It should be noted that the SOA optical amplifier 2 needs to rely on the incident light obtained for lasing to emit the internal light source, and then couples through the micro-ring resonator 32 to obtain the coupled internal light source. The micro-ring resonator 32 is a ring structure with specific resonance characteristics, which can enhance or suppress optical signals of specific wavelengths. Finally, the target emission light is obtained through spectroscopic processing, and the wavelength conversion of the incident light is realized to obtain the target emission light, and the incident light and the target emission light are synchronized.

[0057] The ratio of PBW / FSR is 0.5-0.8. Among them, PBW (Peak base width) is the peak base width of the SOA optical amplifier 2. FSR (Free Spectral Range) represents the free spectrum range of the microring resonant cavity 32, that is, the frequency interval between two consecutive resonance peaks in the microring resonant cavity 32; it is an indicator of the frequency selection performance of the microring resonant cavity 32. In the embodiment of the present application, the ratio of PBW / FSR is limited to between 0.5 and 0.8. In order to ensure that within the spectral range of the SOA optical amplifier 2, the selectivity of the microring resonant cavity 32 to a specific frequency can be controlled, and there is only one coupling peak of the microring resonant cavity 32, ensuring that only the target emission light of a specific wavelength is coupled and emitted.

[0058] The splitting ratio of the beam splitter is 1:9, wherein the coupled internal light source is subjected to a splitting process to obtain 90% of the light intensity as the target emission light.

[0059] The waveguide assembly 31 includes a curved waveguide 311, a first straight waveguide 312, a second straight waveguide 313 and a third straight waveguide 314;

[0060] The curved waveguide 311 is used to conduct the target emission light to the first straight waveguide 312; the first straight waveguide 312 is used to conduct the target emission light conducted through the curved waveguide 311 and emit the target emission light; the second straight waveguide 313 is used to conduct the incident light to the microring resonant cavity 32, and to conduct the internal light source coupled through the microring resonant cavity 32 to the spectrometer; the third straight waveguide 314 is used to conduct the incident light coupled through the microring resonant cavity 32 to the SOA optical amplifier 2, and to conduct the internal light source emitted by the SOA optical amplifier 2 to the microring resonant cavity 32.

[0061] An anti-reflection film is disposed at the input end of the third straight waveguide 314 , and a high-reflection film is disposed at the output end of the third straight waveguide 314 .

[0062] Specifically, an anti-reflection film is arranged at the input end of the third straight waveguide 314, and its main function is to enhance the transmission efficiency of light. The working principle of the anti-reflection film is to increase the transmittance of light by interfering with light of a specific wavelength. It can effectively improve the transmission efficiency of light and reduce the loss of light. A high-reflection film is arranged at the output end of the third straight waveguide 314, and its main function is to increase the reflectivity of light. The working principle of the high-reflection film is to increase the reflectivity of light by interfering with light of a specific wavelength. In this way, light leakage can be effectively prevented, and the light can be better kept in the third straight waveguide 314, thereby achieving more efficient optical coupling and transmission.

[0063] An on-chip magneto-optical isolator is provided between the optical path for obtaining the incident light and the optical splitter, which can achieve the effect of allowing only one-way light to pass, prevent reflected light from negatively affecting the performance and stability of the external light source, and improve the noise suppression capability of the entire system.

[0064] A thermal tuner 321 is provided on the microring resonant cavity 32, and the thermal tuner 321 is in the shape of two arcs at the lower layer of the microring resonant cavity 32; the thermal tuner 321 is used to adjust the temperature of the microring resonant cavity 32 through the positive electrode 3211 of the thermal tuner 321 and the negative electrode 3212 of the thermal tuner 321, so as to tune the resonance peak of the microring resonant cavity 32.

[0065] The thermal tuner 321 is a resonant cavity tuning technology that achieves tuning of the resonance peak of the microring resonant cavity 32 by adjusting the temperature of the microring resonant cavity 32. Specifically, the thermal tuner 321 is usually arranged in the lower layer of the microring resonant cavity 32, and the temperature of the microring resonant cavity 32 is changed by controlling the current through the positive electrode 3211 of the thermal tuner 321 and the negative electrode 3212 of the thermal tuner 321, thereby adjusting the resonance peak of the microring resonant cavity 32. When the current passes through the thermal tuner 321, heat is generated and the temperature of the microring resonant cavity 32 is changed, thereby moving its resonance peak. By controlling the magnitude and direction of the current, the temperature of the microring resonant cavity 32 and the position of the resonance peak can be accurately controlled to achieve tuning of the wavelength. The advantage of the thermal tuner 321 is that high-speed and accurate resonant cavity tuning can be achieved.

[0066] The micro-ring resonant cavity silicon-based chip 3 also includes: a pattern spot converter, which is arranged on one side of the SOA optical amplifier 2 and is used to couple the incident light and the internal light source to the micro-ring resonant cavity silicon-based chip 3, and is arranged on one side of the waveguide component 31 and is used to couple out the laser conducted by the waveguide component 31.

[0067] Specifically, the spot converter receives the optical signal from the SOA optical amplifier 2 and couples it to the microring resonator 32. By adjusting the parameters of the spot converter (such as position, angle, etc.), accurate coupling and efficient transmission of the optical signal can be achieved. The spot converter receives the optical signal from the waveguide component 31 and couples it to the output port. By adjusting the parameters of the spot converter, efficient output and precise positioning of the optical signal can be achieved. That is, the received signal λ1 is converted into the transmitted signal λ2. When the received signal strength L(λ1)>threshold Lth, the transmitted signal strength L(λ2) appears, and the transmitted signal strength L(λ2) increases linearly with the received signal strength L(λ1), as shown in FIG. Figure 2 As shown in b, it is the nonlinear effect of the optical-optical nonlinear neuron device.

[0068] In one example, see Figure 1In this example, the external light source adopts a DFB laser, and the microring resonator silicon-based chip 3 adopts a microring resonator silicon photonic device MRR-SiPh. The DFB laser emits an incident light λ1 through the first optical path entrance 4, and the coupled incident light λ1 is obtained after coupling through the microring resonator 32. The coupled incident light λ1 excites the SOA optical amplifier 2, and the SOA optical amplifier 2 is excited to emit a second incident light λ2. The second incident light λ2 passes through the second optical path entrance 5 and the microring resonator 32 to obtain a coupled second incident light λ2. The coupled second incident light λ2 passes through the spectrometer, wherein 90% of the light intensity of the coupled second incident light λ2 is emitted through the curved waveguide 311 and the first straight waveguide 312 in sequence, and 10% of the light intensity of the coupled second incident light λ2 is absorbed by the magneto-optical isolator.

[0069] Specifically, this example is based on the design principle that since the periodicity of the coupled wavelength of the micro-ring resonator 32 can be infinitely extended, the wavelength conversion of this example can also be achieved in any range of the spectrum. However, the SOA optical amplifier 2 is lasered by the incident light λ1 of the DFB laser. According to the principle that the SOA optical amplifier 2 is lasered by an external light source, the incident light λ1 < the second incident light λ2 should be satisfied. That is, only short wavelengths can be converted to long wavelengths in any range of the spectrum, and the opposite wavelength conversion cannot be achieved. For example, λ1 is 1300nm and λ2 is 1500nm.

[0070] In a second aspect, the present application provides a method for using the MRR wavelength converter. The modulation process of the MRR micro-ring resonant cavity wavelength converter of the present invention is divided into the following three steps.

[0071] Step 1

[0072] Figure 3 Schematic diagram of the optical path of the micro-ring resonator wavelength converter for artificial neural network provided in step 1 of this application, Figure 1 As shown in the basis, Figure 3 As shown, the external light source is a DFB laser 9, and the microring resonator silicon-based chip 3 is a microring resonator silicon photonic device MRR-SiPh. The DFB laser 9 emits the incident light λ1, and at this time the microring resonator 32 is not tuned by the thermal tuner 321, and the incident light λ1 has no matching coupling peak. The incident light λ1 cannot be absorbed by the microring resonator 32, so the incident light λ1 is transmitted from the second optical path exit 7 along the second straight waveguide 313. See Figure 4 This is a schematic diagram of each optical path outlet provided in step 1. It can be seen that only one single peak with a wavelength of λ1 is measured at the second optical path outlet 7. Since the microring resonator 32 is not tuned by the thermal tuner 321, no light is transmitted through the first optical path outlet 6 and the third optical path outlet 8.

[0073] Step 2

[0074] Figure 5 Schematic diagram of the optical path of the micro-ring resonator wavelength converter for artificial neural network provided in step 2 of this application, Figure 1 As shown in the basis, Figure 5 As shown, the external light source uses a DFB laser 9, and the microring resonator silicon-based chip 3 uses a microring resonator silicon photonic device MRR-SiPh. The DFB laser 9 emits incident light λ1, and at this time, the microring resonator 32 is tuned by the thermal tuner 321. Due to the low current on the DFB laser 9, the power of the DFB laser 9 is less than the preset threshold, resulting in the inability to lasing the SOA optical amplifier 2.

[0075] Specifically, the principle of tuning the micro-ring resonant cavity 32 by the thermal tuner 321 is: it is known that the wavelength period that the micro-ring resonant cavity 32 can couple is: Wherein, FSR is the free spectrum range of the microring resonant cavity 32, 2πR is the circumference of the microring resonant cavity 32, and n g is the group refractive index of the material constituting the microring resonator 32, and λ is the wavelength of the laser that can be resonated and absorbed by the microring resonator 32, which is distributed in an equidistant periodic manner. There is a coupling peak of the microring resonator 32 for every wavelength FSR. The phase of the microring resonator 32 is tuned by the thermal tuner 321. When a coupling peak λ1' of the microring resonator 32 is locked with λ1, the positions of other coupling peaks λ0', λ2', λ3', λ4', ... λn' of the microring resonator 32 are also determined at the same time; at this time, the DFB laser 9 emits the incident light λ1 at λn', and the λn' of the microring resonator 32 can be coupled within the range of the incident light λ1, and couple out λn', i.e. λ2. This means that the microring resonator 32 is aligned with the incident wavelength λ1 through a certain coupling peak λ1', and transmits the signal of the incident wavelength λ1 to the distant λn' by virtue of the periodicity of its coupling peak in the wavelength distribution. In this process, the micro-ring resonator 32 can be regarded as a "copy machine", λ1 is the "master", the micro-ring resonator 32 is the "stamp" formed according to the "master" of λ1, and the SOA optical amplifier 2 provides the "ink"; after the micro-ring resonator 32 is aligned and calibrated with λ1, it dips into the "ink" in the SOA optical amplifier 2 and "prints" λ2. Thus, the micro-ring resonator 32 combines the two light sources of the DFB laser 9 and the SOA optical amplifier 2, and with the help of the periodicity of the coupled wavelength, the wavelength converter realizes the function of wavelength conversion, converts the incident laser λ1 into λ2 and emits it (λ1<λ2), and there is a nonlinear relationship between L(λ2) and L(λ1).

[0076] See also Figure 6The schematic diagram of each optical path exit provided for step 2 shows that the micro-ring resonator 32 tuned by the thermal tuner 321 can capture most of the incident light with a wavelength of λ1, but a small part of the light enters the SOA optical amplifier 2 through the third straight waveguide 314. However, since the wavelength of the incident light λ1 is less than the preset threshold, the SOA optical amplifier 2 cannot be excited and cannot generate the second incident light λ2. Therefore, the light with a wavelength of λ1 measured at the second optical path exit 7 is weak, and no light is transmitted through the first optical path exit 6 and the third optical path exit 8.

[0077] Step 3

[0078] Figure 7 Schematic diagram of the optical path of the micro-ring resonator wavelength converter for artificial neural network provided in step 3 of this application, Figure 1 As shown in the basis, Figure 7 As shown, the external light source uses a DFB laser 9, and the microring resonator silicon-based chip 3 uses a microring resonator silicon photonic device MRR-SiPh. The current of the DFB laser 9 is increased so that the power of the DFB laser 9 is greater than a preset threshold, and the SOA optical amplifier 2 is lased. The DFB laser 9 emits an incident light λ1, and the micro-ring resonant cavity 32 tuned by the thermal tuner 321 couples the incident light λ1 to obtain the coupled incident light λ1, and the coupled incident light λ1 excites the SOA optical amplifier 2 through the third straight waveguide 314, and the SOA optical amplifier 2 emits a second incident light λ2 after excitation and enters the third straight waveguide 314, and the second incident light λ2 is coupled through the micro-ring resonant cavity 32 to obtain the coupled second incident light λ2, and the coupled second incident light λ2 enters the spectrometer through the second straight waveguide 313 for spectroscopic processing, and 90% of the light intensity is used as the target emission light, and is sequentially emitted from the first optical path outlet 6 through the curved waveguide 311 and the first straight waveguide 312, and 10% of the light intensity will leak out and be absorbed by the magneto-optical isolator to prevent the light from returning to the DFB laser 9 from the spectrometer.

[0079] See also Figure 8 The schematic diagram of each optical path exit provided for step three shows that due to the periodicity of the micro-ring resonant cavity 32, the micro-ring resonant cavity 32 will have a coupling peak λn' (i.e., the second incident light λ2) within the lasing range of the SOA optical amplifier 2. Therefore, in the third optical path exit 8, there will be an absorption peak of λ2 in the broad peak of the SOA optical amplifier 2, and the λ2 coupled and leaked by the micro-ring resonant cavity 32 will be emitted from the first optical path exit 6.

[0080] In summary, the embodiment of the present application provides a microring resonant cavity wavelength converter for artificial neural networks, which can convert the received signal λ1 into the transmitted signal λ2. When the received signal strength L(λ1)>threshold Lth, the transmitted signal strength L(λ2) appears, and the transmitted signal strength L(λ2) increases linearly with the received signal strength L(λ1). In addition, the present application uses the periodic characteristics of the microring resonant cavity silicon-based chip 3 for wavelength conversion for the first time, which can achieve long-distance wavelength transfer. Since the structure of the device is essentially an optical chip, it is small, compact and simple. This makes the present application an efficient, energy-saving and reliable microring resonant cavity microring resonant cavity wavelength converter for optical-optical nonlinear neuron devices used in artificial neural networks.

[0081] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementation modes, and this embodiment will not be described in detail here.

[0082] Those skilled in the art should understand that the technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0083] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A micro-ring resonator wavelength converter for artificial neural network, characterized in that: The wavelength converter comprises: Base plate components; An SOA optical amplifier, integrated on the substrate component, for lasing an internal light source according to the incident light acquired; A micro-ring resonant cavity silicon-based chip is integrated on the substrate component and is used to perform coupling processing based on the incident light and the internal light source to determine the target emitted light.

2. The micro-ring resonator wavelength converter for artificial neural network according to claim 1, characterized in that: The micro-ring resonator silicon-based chip comprises: A waveguide assembly for guiding the incident light and the internal light source; A micro-ring resonant cavity, used for performing coupling processing according to the incident light to determine the incident light after coupling, and exciting the SOA optical amplifier to emit the internal light source according to the incident light after coupling, and performing coupling processing according to the internal light source to determine the internal light source after coupling; The spectrometer is used to perform spectroscopic processing based on the coupled internal light source to obtain the target emission light.

3. The micro-ring resonator wavelength converter for artificial neural network according to claim 2, characterized in that: The ratio of PBW / FSR is 0.5-0.8, wherein PBW is the peak-to-base width of the SOA optical amplifier, and FSR is the free spectrum range of the microring resonant cavity.

4. The micro-ring resonator wavelength converter for artificial neural network according to claim 2, characterized in that: The splitting ratio of the beam splitter is 1:9, wherein the coupled internal light source is subjected to a splitting process to obtain 90% of the light intensity as the target emission light.

5. The micro-ring resonator wavelength converter for artificial neural network according to claim 2, characterized in that: The waveguide assembly includes a curved waveguide, a first straight waveguide, a second straight waveguide and a third straight waveguide; The curved waveguide is used to guide the target emitted light to the first straight waveguide; The first straight waveguide is used to guide the target emission light guided by the curved waveguide and emit the target emission light; The second straight waveguide is used to transmit the incident light to the microring resonant cavity, and transmit the internal light source coupled by the microring resonant cavity to the spectrometer; The third straight waveguide is used to transmit the incident light coupled by the microring resonant cavity to the SOA optical amplifier, and transmit the internal light source emitted by the SOA optical amplifier to the microring resonant cavity.

6. The micro-ring resonator wavelength converter for artificial neural network according to claim 5, characterized in that: An anti-reflection film is arranged at the input end of the third straight waveguide, and a high-reflection film is arranged at the output end of the third straight waveguide.

7. The micro-ring resonator wavelength converter for artificial neural network according to claim 2, characterized in that: A magneto-optical isolator is provided between the optical path for acquiring the incident light and the optical splitter.

8. The micro-ring resonator wavelength converter for artificial neural network according to claim 2, characterized in that: A thermal tuner is provided on the microring resonant cavity, and the shape of the thermal tuner is two arcs surrounding the microring resonant cavity; The thermal tuner is used to adjust the temperature of the microring resonant cavity through the positive and negative electrodes of the thermal tuner to tune the resonance peak of the microring resonant cavity.

9. The micro-ring resonator wavelength converter for artificial neural network according to claim 2, characterized in that: The micro-ring resonator silicon-based chip further comprises: The spot converter is arranged on one side of the SOA optical amplifier to couple the incident light and the internal light source to the micro-ring resonator silicon-based chip, and is arranged on one side of the waveguide component to couple out the laser light conducted by the waveguide component.

10. A method for using a micro-ring resonant cavity wavelength converter for an artificial neural network, characterized in that: The method is applied to the micro-ring resonator wavelength converter for artificial neural network according to any one of claims 1 to 9, and the method comprises: Obtain incident light; Based on the incident light, a pre-tuned micro-ring resonator silicon-based chip is used for coupling processing to determine the incident light after coupling; Based on the coupled incident light, an internal light source emitted by the SOA optical amplifier is excited and obtained; Based on the internal light source, a pre-tuned micro-ring resonator silicon-based chip is used for coupling processing to determine the internal light source after coupling; The target emission light is obtained by performing spectroscopic processing based on the coupled internal light source.

Images