Signal enhancement device and method based on assistance of optical neural network

Abstract

The invention provides a signal enhancement device and method based on the assistance of an optical neural network, and relates to the technical field of optical communication, and the device comprises an optical collection unit, a signal processing unit and a signal conversion unit. The optical acquisition unit comprises a plurality of optical fiber sensors and is configured to acquire target environment parameters and convert the target environment parameters into corresponding first optical signals; the signal processing unit comprises an input layer, a hidden layer and an output layer, and the input layer comprises a plurality of optical neurons and is configured to perform optical domain processing on different components of an input first optical signal to obtain a second optical signal; the hidden layer comprises multiple layers of optical neurons connected through an optical interconnection structure and is configured to perform nonlinear transformation on the second optical signal and extract key feature information to obtain a third optical signal; the output layer is configured to integrate and enhance the third optical signal and output a target optical signal; and the signal conversion unit is configured to convert the target optical signal into a corresponding electric signal and output the electric signal to the target port.

Description

Technical Field

[0001] The present invention relates to the field of optical communication technologies, and particularly to a signal enhancement device and method assisted by an optical neural network. Background Art

[0002] Optical fiber sensors have been widely used in many fields such as environmental monitoring, structural health monitoring, and biomedical detection due to their many advantages such as anti-electromagnetic interference, high sensitivity, and remote monitoring. However, in practical applications, the signals collected by optical fiber sensors are often easily affected by various factors, such as external environmental noise, the loss of the optical fiber itself, and the fluctuation of the light source, resulting in a decline in signal quality and affecting the subsequent accurate acquisition and analysis of monitoring information. Although existing signal processing technologies can improve the optical fiber sensor signals to a certain extent, the effect is limited, and complex algorithms and high computational costs are often required. Therefore, a new technical means is urgently needed to more effectively enhance the optical fiber sensor signals and improve their reliability in various complex application scenarios. Summary of the Invention

[0003] In view of this, the present invention proposes a signal enhancement device and method assisted by an optical neural network.

[0004] The technical solution of the present invention is implemented as follows: In the first aspect of the present invention, a signal enhancement device assisted by an optical neural network is provided, including: an optical acquisition unit, a signal processing unit, and a signal conversion unit; wherein,

[0005] The optical acquisition unit includes a plurality of optical fiber sensors, configured to collect target environmental parameters and convert the target environmental parameters into corresponding first optical signals;

[0006] The signal processing unit includes an input layer, a hidden layer, and an output layer. The input layer includes a plurality of optical neurons, configured to perform optical domain processing on different components of the input first optical signal to obtain a second optical signal; the hidden layer includes multiple layers of optical neurons, and each layer of optical neurons is connected through an optical interconnection structure, configured to perform a non-linear transformation on the second optical signal, extract key feature information, and obtain a third optical signal; the output layer is configured to integrate and enhance the third optical signal and output a target optical signal;

[0007] The signal conversion unit is configured to convert the target optical signal into a corresponding electrical signal and output it to a target port.

[0008] Based on the above technical solutions, preferably, the input layer is specifically configured to obtain the component signals included in the first optical signal, modulate, interfere, and / or filter different component signals to obtain a second optical signal; the component signals include at least one of amplitude, phase, optical intensity, frequency, and polarization state.

[0009] Based on the above technical solutions, preferably, each layer of the optical neurons is connected by a waveguide-based optical interconnection structure.

[0010] Based on the above technical solutions, preferably, the hidden layer is specifically configured to perform a non-linear transformation on the second optical signal using the ReLU activation function or the Sigmoid activation function, extract key feature information, and obtain a third optical signal.

[0011] More preferably, a second aspect of the present invention provides a signal enhancement method assisted by an optical neural network, which is applied to the signal enhancement device assisted by an optical neural network described in the first aspect, and includes:

[0012] Obtain target environmental parameters and convert the target environmental parameters into corresponding first optical signals;

[0013] Perform optical domain processing on different components of the first optical signal to obtain a second optical signal, and use multiple layers of optical neurons connected by an optical interconnection structure to perform non-linear transformation on the second optical signal, extract key feature information, and obtain a third optical signal; integrate and enhance the third optical signal to obtain a target optical signal;

[0014] Convert the target optical signal into a corresponding electrical signal and output it to the target port.

[0015] Based on the above technical solutions, preferably, the performing optical domain processing on different components of the first optical signal to obtain a second optical signal includes using the following formula for signal conversion:

[0016] I 1，j = f1(w ij ·I 1，i );

[0017] where I 1，j is the optical signal output by the j-th optical neuron in the input layer, w ij is the weight coefficient between the input layer and the sensor signal, and f1 is a linear or non-linear activation function of the input layer.

[0018] Based on the above technical solutions, preferably, the using multiple layers of optical neurons connected by an optical interconnection structure to perform non-linear transformation on the second optical signal, extract key feature information, and obtain a third optical signal includes using the following formula for signal conversion:

[0019]

[0020] Among them, is the optical signal output by the k-th optical neuron in the l-th hidden layer, is the weight coefficient between the j-th neuron in the (l - 1)-th layer and the k-th neuron in the l-th layer, and f h is the non-linear activation function corresponding to the hidden layer.

[0021] Based on the above technical solutions, preferably, the integration and enhancement of the third optical signal to obtain the target optical signal includes performing signal conversion using the following formula:

[0022]

[0023] where I is the optical signal of the output layer, and w k is the weight coefficient between the hidden layer and the output layer, and f o is the activation function corresponding to the output layer.

[0024] More preferably, in the third aspect of the present invention, an electronic device is provided, including a processor and a memory; the memory stores a computer program, wherein the computer program, when executed by the processor, implements the signal enhancement method based on the optical neural network assistance described in the second aspect.

[0025] More preferably, in the fourth aspect of the present invention, a computer storage medium is provided, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the signal enhancement method based on the optical neural network assistance described in the second aspect.

[0026] The signal enhancement device based on the optical neural network assistance of the present invention has the following beneficial effects compared with the prior art:

[0027] 1. By deeply integrating the optical neural network with fiber optic sensing and directly embedding the optical neural network into the fiber optic sensor signal processing link, the enhancement of all-optical domain signals is realized, avoiding signal loss and electronic computing delay in traditional optoelectronic conversion, and significantly improving the processing efficiency and real-time performance. At the same time, splitting the single-channel sensing signal into multiple optical components for parallel processing realizes ultra-high-speed parallel computing, greatly improving the signal processing efficiency.

[0028] 2. The hidden layer forms an "optical domain feature screening funnel" through cascaded non-linear processing. The first layer extracts the sensitive frequency band of the physical quantity, and the subsequent layers dynamically suppress the characteristic wavelength components of environmental noise. Using the adaptive noise suppression mechanism, through cascaded non-linear processing, the recognition ability for the sensitive frequency band can be further enhanced, improving the accuracy and efficiency of signal processing.

[0029] 3. Use an activation function to perform a non - linear transformation on the optical signal, extract key feature information, and allow the neural network to learn non - linear patterns and features, so that the neural network can better adapt to complex data patterns, improve the expression ability of the model, and ensure its reliability in various complex application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required in the description of the embodiments or the prior art. Obviously, the following - described drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0031] Figure 1 Schematic diagram of the structure of a signal enhancement device assisted by an optical neural network provided by an embodiment of the present invention;

[0032] Figure 2 Schematic diagram of the specific application of the signal enhancement device provided by an embodiment of the present invention;

[0033] Figure 3 Schematic diagram of the structure of the signal processing unit provided by an embodiment of the present invention;

[0034] Figure 4 Schematic diagram of the comparison of the effects before and after signal enhancement provided by an embodiment of the present invention;

[0035] Figure 5 Schematic diagram of the flow of a signal enhancement method assisted by an optical neural network provided by an embodiment of the present invention;

[0036] Figure 6 Schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in combination with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0038] In some embodiments, as Figure 1 shown, Figure 1Schematic diagram of a signal enhancement device assisted by an optical neural network provided by an embodiment of the present invention; A signal enhancement device assisted by an optical neural network provided by the present invention includes: an optical acquisition unit 110, a signal processing unit 120, and a signal conversion unit 130; wherein,

[0039] The optical acquisition unit 110 includes a plurality of fiber optic sensors configured to acquire target environmental parameters and convert the target environmental parameters into corresponding first optical signals.

[0040] The signal processing unit 120 includes an input layer 121, a hidden layer 122, and an output layer 123. The input layer 121 includes a plurality of optical neurons configured to perform optical domain processing on different components of the input first optical signal to obtain a second optical signal; the hidden layer 122 includes multiple layers of optical neurons connected by an optical interconnection structure, configured to perform a non-linear transformation on the second optical signal to extract key feature information and obtain a third optical signal; the output layer 123 is configured to integrate and enhance the third optical signal and output a target optical signal.

[0041] The signal conversion unit 130 is configured to convert the target optical signal into a corresponding electrical signal and output it to a target port.

[0042] In this embodiment, the fiber optic sensors in the optical acquisition unit are responsible for converting target environmental parameters, such as temperature, pressure, vibration, etc., into optical signals. When the external temperature changes, the light wave in the optical fiber will be modulated, and parameters such as its intensity, phase, or wavelength will change. By detecting these parameter changes, the external temperature information can be deduced. Similarly, when the optical fiber is subjected to an external pressure, its length, refractive index, or diameter, etc., may change, thereby affecting the optical transmission characteristics in the optical fiber. By detecting the change in the optical signal, the external pressure information can be sensed. Strain refers to the deformation of an object under the action of an external force. The fiber optic sensor can sense strain information by detecting the change in the optical signal in the optical fiber. For example, when the optical fiber is stretched or compressed, its length will change, resulting in a change in the phase or intensity of the optical signal.

[0043] In some embodiments, the input layer 121 is specifically configured to obtain the component signals included in the first optical signal, perform modulation, interference, and / or filtering on different component signals to obtain a second optical signal; the component signals include at least one of amplitude, phase, optical intensity, frequency, and polarization state.

[0044] The above-mentioned component signals can be extracted through optical components such as gratings, polarizers, interferometers, etc. Modulation encodes the information to be transmitted into the optical signal by changing parameters such as the amplitude, phase, or frequency of the optical signal. Interference can achieve the interference effect of the optical signal by controlling the phase difference and amplitude ratio of the light waves, thereby performing operations such as phase measurement and light intensity measurement on the optical signal. Through optical filters such as grating filters and interference filters, the required optical signal components can be separated from the complex optical signal. Here, it can be to separate the useful signal and the noise signal in the first optical signal to obtain a second optical signal containing the useful signal.

[0045] In some embodiments, each layer of optical neurons is connected through a waveguide-based optical interconnection structure.

[0046] Waveguide optical interconnection technology uses a channel formed by a high-refractive-index medium surrounded by a low-refractive-index medium to confine light waves for propagation. In a multi-layer neural network, each layer of optical neurons is connected to the optical neurons of the adjacent layer through a waveguide optical interconnection structure, enabling information to be transmitted quickly and accurately in the neural network. By optimizing the structure and layout of the waveguide, flexible connection of optical neurons between different layers can be achieved to meet the requirements of the neural network for complex information processing.

[0047] In some embodiments, the hidden layer 122 is specifically configured to perform a non-linear transformation on the second optical signal using a ReLU activation function or a Sigmoid activation function to extract key feature information and obtain a third optical signal.

[0048] In this embodiment, the hidden layer includes several layers of optical neurons. Each layer of optical neurons is connected through a specific optical interconnection structure to achieve the transmission and interaction of optical signals between different neurons. The optical neurons in the hidden layer use a specific optical neuron activation function, such as a ReLU activation function or a Sigmoid activation function, to perform a non-linear transformation on the input optical signal, extract the key feature information in the signal, and at the same time suppress the noise component and enhance the energy of the useful signal. Through the cascaded processing of multiple hidden layers, the extraction and enhancement effect of signal features can be gradually deepened. The specific role of the activation function is to convert the original linear calculation into non-linear, thereby introducing non-linear characteristics and improving the expression ability and generalization ability of the model. It should be noted that in each layer of the neural network, that is, in the input layer 121, the hidden layer 122, and the output layer 123, an activation function can be used to perform a non-linear transformation on the input signal. The choice of the activation function depends on the specific task and data characteristics. After the non-linear transformation, the key feature information in the second optical signal can be captured, thereby obtaining a third optical signal.

[0049] In an alternative embodiment, please refer to Figure 2 , Figure 2Schematic diagram of the specific application of the signal enhancement device provided by the embodiments of the present invention. In practical applications, the optical acquisition unit (corresponding to the sensing head) can be installed in the environment or on the object to be monitored. When the physical quantity to be measured in the external environment changes, the fiber optic sensor module in the optical acquisition unit generates corresponding optical signals. For example, in the scenario of structural health monitoring, the fiber optic sensor is pasted on the surface of the bridge structure. When the bridge undergoes strain due to factors such as vehicle load and wind force, the fiber optic sensor senses the strain information and converts it into an optical signal for output. A high-precision optical detector (corresponding to the polarizer) and a low-loss fiber optic transmission link (corresponding to the light-sending fiber) are used between the optical acquisition unit and the input layer of the signal processing unit to ensure that the optical signal can be accurately and completely transmitted to the optical neural network (metasurface). According to the characteristics of the optical signal, the input layer of the optical neural network decomposes the optical signal into multiple components, and each optical neuron processes one component. For example, the intensity of the optical signal is modulated through an optical intensity modulator to meet the input requirements of the subsequent hidden layer. In the hidden layer, the optical neurons achieve efficient transmission of optical signals through a carefully designed waveguide optical interconnection structure. The activation function of the optical neuron can adopt a non-linear activation function based on the optical Kerr effect. When the optical signal passes through the optical neuron containing the optical Kerr material, the refractive index of the material changes with the light intensity, thereby realizing non-linear transformation of the optical signal, highlighting the key features in the signal, and suppressing noise. The output layer integrates the processing results from the hidden layer. For example, the output optical signals of multiple optical neurons are coupled together through an optical coupler, and then the optical signal is amplified through an optical amplifier to ensure that the output enhanced optical signal has sufficient intensity and good quality, so that the subsequent optoelectronic signal conversion and output module can accurately convert it into an electrical signal. The optoelectronic signal conversion and output module uses a high-performance photodetector to convert the enhanced optical signal into an electrical signal, and this electrical signal can be transmitted to electronic devices such as a data acquisition card and a computer for further data analysis, processing, and display operations, thereby realizing accurate monitoring of the physical quantity to be measured.

[0050] In an alternative embodiment, please refer to Figure 3 , Figure 3Schematic structural diagram of the signal processing unit provided by an embodiment of the present invention. The signal processing unit includes an input layer, a hidden layer, and an output layer. Among them, the hidden layer is a nanostructured metasurface. A metasurface is composed of a series of sub-wavelength unit structures with high degrees of freedom, non-periodicity, and dense arrangement on a two-dimensional plane. It can start from designing a single scattering structure and then expand into a two-dimensional array structure, thereby having the ability to control the polarization direction, amplitude, and phase of the beam. It should be noted that the signal processing unit may include multiple hidden layers. Through the cascaded processing of multiple hidden layers, the extraction of signal features can be continuously deepened and the enhancement effect can be improved. For example, the first hidden layer mainly extracts the primary features of the signal, and the second hidden layer further extracts more advanced features on this basis, while gradually enhancing the energy of the useful signal.

[0051] In an alternative embodiment, the output layer 123 is also composed of optical neurons, receives the optical signal output from the last hidden layer, couples the output optical signals of multiple optical neurons together through an optical coupler, and then amplifies the optical signal through an optical amplifier to ensure that the output enhanced optical signal has sufficient intensity and good quality. This signal has a significant improvement in terms of signal-to-noise ratio, signal integrity, etc. compared with the original fiber optic sensor signal, so that the subsequent optoelectronic signal conversion and output module can accurately convert it into an electrical signal.

[0052] In an example, the output layer 123 integrates an erbium-doped optical waveguide amplifier, automatically adjusts the pump power through closed-loop optical intensity monitoring, compensates for the optical loss in neural network processing, and ensures that the fluctuation of the final output optical signal intensity is less than ±0.5 dB.

[0053] In an alternative embodiment, please refer to Figure 4 , Figure 4 Schematic diagram of the comparison of the effects before and after signal enhancement provided by an embodiment of the present invention. The original signal contains useful signals and noise signals (not shown in the figure). From the visual effect of the waveform diagram, the enhanced signal effectively reduces noise and interference, improves the signal-to-noise ratio of the signal, and thus appears smoother and more stable overall. There are significant differences in waveform characteristics, signal intensity, and signal quality between the fiber optic sensor signals before and after enhancement. These differences indicate that the signal enhancement process has a positive effect on the reliability and accuracy of fiber optic sensor signals, improving the signal quality and usability.

[0054] In some embodiments, please refer to Figure 5 , Figure 5 Schematic flow diagram of a signal enhancement method assisted by an optical neural network provided by an embodiment of the present invention; The present invention provides a signal enhancement method assisted by an optical neural network, which is applied to the above-mentioned signal enhancement device assisted by an optical neural network, including:

[0055] S510. Obtain the target environmental parameters and convert the target environmental parameters into corresponding first optical signals.

[0056] S520. Perform optical domain processing on different components of the first optical signal to obtain a second optical signal, and use multi-layer optical neurons connected through an optical interconnection structure to perform non-linear transformation on the second optical signal, extract key feature information, and obtain a third optical signal; perform integration and enhancement on the third optical signal to obtain a target optical signal.

[0057] S530. Convert the target optical signal into a corresponding electrical signal and output it to the target port.

[0058] In this embodiment, the parameters of the target environment are obtained through an optical fiber sensor. These parameters can be physical quantities such as temperature, pressure, and vibration. After obtaining these parameters, an optical modulator or other related technologies are used to convert these parameters into corresponding first optical signals. The processed first optical signal is sent into multi-layer optical neurons connected through an optical interconnection structure. These optical neurons use optical non-linear effects (such as the optical Kerr effect, photorefractive effect, etc.) to perform non-linear transformation on the optical signal. Through the processing of multi-layer optical neurons, the key feature information in the optical signal can be extracted and a second optical signal is generated. Further, further processing is performed on the second optical signal, such as filtering, amplification, etc., to obtain a third optical signal. In order to further improve the reliability and accuracy of the signal, integration and enhancement processing is performed on the third optical signal, such as can be achieved through technologies such as optical amplifiers and optical filters. The integrated and enhanced optical signal is the target optical signal, which contains the key feature information in the target environmental parameters and has sufficient intensity and stability for subsequent processing.

[0059] In some embodiments, in S520, performing optical domain processing on different components of the first optical signal to obtain a second optical signal includes using the following formula for signal conversion:

[0060] I 1，j =f1(w ij ·I 1，i );

[0061] where, I 1，j is the optical signal output by the optical neuron of the j-th input layer, w ij is the weight coefficient between the input layer and the sensor signal, and f1 is the linear or non-linear activation function of the input layer.

[0062] In this embodiment, the signal conversion is completed by the input layer. The input layer includes multiple optical neurons, and each optical neuron can receive different components of the optical signal transmitted from the optical signal acquisition and transmission unit. By performing preliminary optical domain processing on these components, it converts them into an optical signal form suitable for internal transmission and processing in the optical neural network, and transmits the processed signal to the hidden layer. The optical domain processing includes operations such as optical intensity modulation and phase modulation. The input layer receives the sensor signal, and through multiplication and accumulation with the weight coefficients, passes the result to the next layer. The magnitude of the weight coefficient determines the degree of influence of each sensor signal on the output result. If a certain sensor signal has a greater influence on the final result, then the corresponding weight coefficient of this signal will also be relatively large. Through the backpropagation algorithm, the neural network can adjust the weight coefficients according to the error function to reduce the difference between the output result and the true value.

[0063] In some embodiments, S520, perform a non-linear transformation on the second optical signal using multiple layers of optical neurons connected by an optical interconnection structure to extract key feature information and obtain a third optical signal, including performing signal conversion using the following formula:

[0064]

[0065] Where is the optical signal output by the k-th optical neuron in the l-th hidden layer, is the weight coefficient between the j-th neuron in the (l - 1)-th layer and the k-th neuron in the l-th layer, and f h is the non-linear activation function corresponding to the hidden layer.

[0066] In this embodiment, the signal conversion is completed by the hidden layer. Neurons in different layers are responsible for extracting features at different levels. Shallow neurons may be responsible for extracting low-level features such as edges and textures, while deep neurons are responsible for extracting higher-level features such as objects and scenes. The process of extracting these features depends on the weight coefficients between neurons in different layers. The weight coefficient determines the degree of influence of the output of the previous layer of neurons on the input of the next layer of neurons, and thus affects the overall output result of the hidden layer. The magnitude and sign of the weight coefficient reflect the correlation and influence direction between neurons. A larger weight coefficient indicates a stronger correlation between two neurons, while a smaller weight coefficient indicates a weaker correlation. The sign determines the direction of influence, with a positive weight indicating a positive correlation and a negative weight indicating a negative correlation.

[0067] In some embodiments, S520, integrate and enhance the third optical signal to obtain the target optical signal, including performing signal conversion using the following formula:

[0068]

[0069] where I is the optical signal of the output layer, and w k is the weight coefficient between the hidden layer and the output layer, and f o is the activation function corresponding to the output layer.

[0070] The weight coefficients between the hidden layer and the output layer determine the degree of influence of the output of the hidden layer neurons on the input of the output layer neurons. These weight coefficients transform the output of the hidden layer into the input of the output layer through weighted summation, thereby determining the final output result. Here, the weight coefficients can be continuously adjusted through the backpropagation algorithm to minimize the loss function and improve the prediction accuracy of the model.

[0071] It should be noted that the signal enhancement method assisted by the optical neural network provided in the embodiments of the present application and the signal enhancement device assisted by the optical neural network provided in the embodiments of the present application are based on the same application concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned signal enhancement device assisted by the optical neural network, and the repeated parts will not be elaborated.

[0072] In some embodiments, please refer to Figure 6 , Figure 6 which is a schematic structural diagram of an electronic device provided in an embodiment of the present application. An electronic device 600 provided in an embodiment of the present application includes a processor 610 and a memory 620; the memory 620 stores a computer program, and when the computer program is executed by the processor, the above-mentioned signal enhancement device assisted by the optical neural network is implemented.

[0073] Specifically, the processor 610 may include, for example, a general microprocessor, an instruction set processor, and / or a related chipset, and / or a dedicated microprocessor (such as an application specific integrated circuit (ASIC)), etc. The processor 610 may also include on-board memory for caching purposes. The processor 610 may be a single processing unit or multiple processing units for performing different actions of the method flow according to the embodiments of the present application.

[0074] The memory 620 may be, for example, any medium capable of containing, storing, transmitting, propagating, or transmitting instructions. For example, the memory 620 may include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, components, or propagation media. Specific examples of the memory 620 include: magnetic storage devices, such as magnetic tapes or hard disk drives (HDDs); optical storage devices, such as compact discs (CD-ROMs); it may also be, for example, random access memory (RAM) or flash memory; and / or wired / wireless communication links.

[0075] The present application also provides a computer-readable medium, on which a computer program is stored, and when the program is executed by a processor, it implements the above-described signal enhancement device assisted by an optical neural network. The computer-readable medium may be included in the device / device / system described in the above embodiments; or it may exist alone without being assembled into the device / device / system. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiments of the present application is implemented.

[0076] According to an embodiment of the present application, the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. And in the present application, the computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium may also be any computer-readable medium other than the computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted by any suitable medium, including but not limited to: wireless, wired, optical cable, radio frequency signal, etc., or any suitable combination of the above.

[0077] Those skilled in the art will understand that the features recited in the various embodiments and / or claims of the present application can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly recited in the present application. In particular, without departing from the spirit and teachings of the present application, the features recited in the various embodiments and / or claims of the present application can be combined and / or combined in various ways. All such combinations and / or combinations fall within the scope of the present application. Therefore, the scope of the present application should not be limited to the above embodiments, but should be determined not only by the appended claims, but also by the equivalents of the appended claims. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A signal enhancement device assisted by an optical neural network, characterized in that including: an optical acquisition unit, a signal processing unit, and a signal conversion unit; wherein, the optical acquisition unit includes a plurality of optical fiber sensors configured to acquire target environmental parameters and convert the target environmental parameters into corresponding first optical signals; the signal processing unit includes an input layer, a hidden layer, and an output layer. The input layer includes a plurality of optical neurons configured to perform optical domain processing on different components of the input first optical signal to obtain a second optical signal. The hidden layer includes multiple layers of optical neurons connected by an optical interconnection structure, configured to perform a non-linear transformation on the second optical signal to extract key feature information and obtain a third optical signal. The output layer is configured to integrate and enhance the third optical signal and output a target optical signal; the signal conversion unit is configured to convert the target optical signal into a corresponding electrical signal and output it to a target port.

2. The signal enhancement device based on the optical neural network assistance according to claim 1, wherein The input layer is specifically configured to obtain component signals included in the first optical signal, perform modulation, interference, and / or filtering on different component signals to obtain a second optical signal; the component signals include at least one of amplitude, phase, optical intensity, frequency, and polarization state.

3. The signal enhancement device based on the optical neural network assistance according to claim 1, wherein Each layer of the optical neurons is connected by a waveguide-based optical interconnection structure.

4. The signal enhancement device assisted by an optical neural network according to claim 1, characterized in that The hidden layer is specifically configured to perform a non-linear transformation on the second optical signal using a ReLU activation function or a Sigmoid activation function to extract key feature information and obtain a third optical signal.

5. A signal enhancement method assisted by an optical neural network, applied to the signal enhancement device assisted by an optical neural network described in claims 1-4, characterized in that, including: acquiring target environmental parameters and converting the target environmental parameters into corresponding first optical signals; performing optical domain processing on different components of the first optical signal to obtain a second optical signal, and performing a non-linear transformation on the second optical signal using multiple layers of optical neurons connected by an optical interconnection structure to extract key feature information and obtain a third optical signal; integrating and enhancing the third optical signal to obtain a target optical signal; converting the target optical signal into a corresponding electrical signal and outputting it to a target port.

6. The signal enhancement method based on optical neural network assistance according to claim 5, wherein The performing optical domain processing on different components of the first optical signal to obtain a second optical signal includes performing signal conversion using the following formula: I 1，j = f1(w ij .I 1，i ); Among them, I 1，j is the optical signal output by the j-th optical neuron in the input layer, w ij is the weight coefficient between the input layer and the sensor signal, and f1 is the linear or non-linear activation function of the input layer.

7. The signal enhancement method based on the assistance of an optical neural network according to claim 6, characterized in that, The performing a non-linear transformation on the second optical signal using multiple layers of optical neurons connected by an optical interconnection structure to extract key feature information and obtain a third optical signal includes performing signal conversion using the following formula: Among them, is the optical signal output by the k-th optical neuron in the l-th hidden layer, is the weight coefficient between the j-th neuron in the (l - 1)-th layer and the k-th neuron in the l-th layer, and f h is the non-linear activation function corresponding to the hidden layer.

8. The signal enhancement method based on the optical neural network assistance according to claim 7, wherein The integrating and enhancing the third optical signal to obtain a target optical signal includes performing signal conversion using the following formula: where I is the optical signal of the output layer, w k is the weight coefficient between the hidden layer and the output layer, and f o is the activation function corresponding to the output layer.

9. An electronic device, comprising a processor and a memory; the memory stores a computer program, wherein, When executed by the processor, the computer program implements the signal enhancement method assisted by an optical neural network according to any one of claims 5-8.

10. A computer storage medium, characterized in that, A computer program is stored thereon, wherein when the computer program is executed by a processor, it implements the signal enhancement method assisted by an optical neural network according to any one of claims 5-8.

Images