A photon pressure sensing system and implementation method for eliminating temperature influence

Abstract

The application discloses a kind of photonic pressure sensing systems and implementation methods of eliminating temperature influence, belong to microwave photon sensing processing technical field.The system includes electro-optical conversion mixing unit, photonic sensing unit, photoelectric conversion processing unit, signal processing unit and deep learning processing unit.The present application is cascaded with local oscillator signal and microwave signal by electro-optical conversion mixing unit to wide spectrum laser in turn Modulation, generate mixing light signal;Mixing light signal input photonic sensing unit, under the action of pressure and temperature, output sensing optical signal, then converted into sensing microwave signal by photoelectric conversion processing unit;Signal parameter is detected and extracted by signal processing unit;According to signal parameter, the pressure value under different temperatures is calculated by deep learning processing unit.The present application solves the problem that pressure cannot be accurately measured due to the cross-sensitivity of pressure and temperature in existing photonic pressure sensing systems, and has the advantages of high precision, fast demodulation, strong adaptability, simple architecture, etc.

Description

Technical Field

[0001] This invention relates to the field of microwave photonic sensing technology, specifically to a photonic pressure sensing system and method for eliminating the effects of temperature. Background Technology

[0002] In recent years, biomedical engineering has employed photonic sensing technology to monitor parameters such as human body temperature, pressure, heart rate, and respiration. Compared to electronic sensing technology, photonic sensing offers advantages such as high sensitivity, small size, good biosafety, resistance to electromagnetic interference, and high reliability, making it particularly suitable for special environments such as intensive care units (ICUs) and magnetic resonance imaging (MRI). Traditional photonic sensing systems typically use fiber optic gratings or fiber optic interferometers to construct photonic sensing units, measuring physical quantities such as pressure or temperature by detecting wavelengths through optical wavelength demodulation units. However, due to the cross-sensitivity of fiber optic gratings and fiber optic interferometers to pressure and temperature, temperature changes can severely affect the accuracy of pressure measurements. Furthermore, the wavelength detection accuracy of optical wavelength demodulation units is limited, resulting in low measurement accuracy in traditional photonic sensing systems.

[0003] In recent years, microwave photonic sensing technology has utilized the interaction between microwaves and light waves to measure physical quantities, offering advantages such as high precision, high sensitivity, good stability, and good environmental adaptability. In microwave photonic sensing systems, microwave signals are laser-modulated using a broadband or multi-wavelength light source. The modulated light, after delay and weighting, is then converted by photoelectric conversion to obtain a filtered microwave signal, thus achieving the microwave photonic filtering effect. External physical quantities such as pressure or temperature act on a microwave photonic sensing unit composed of various optical elements. The microwave signal is scanned within a certain frequency range, and the frequency interval corresponding to the maximum output amplitude response of adjacent microwave passbands is detected to obtain the physical quantity to be measured (see patent CN109580038A, Temperature Sensing Demodulation Device and Demodulation Method Based on Microwave Photonic Filter, 2019). However, because microwave photonic sensing systems also suffer from cross-sensitivity to pressure and temperature, accurate pressure measurement is impossible when the temperature of the photonic sensing unit changes, severely limiting its application in practical biomedicine.

[0004] Therefore, the present invention provides a photonic pressure sensing system and implementation method that eliminates the influence of temperature, so as to at least solve some of the above-mentioned technical problems. Summary of the Invention

[0005] The purpose of this invention is to provide a photonic pressure sensing system and implementation method that eliminates the influence of temperature, so as to solve the problem that the pressure cannot be accurately measured due to the cross-sensitivity of pressure and temperature in the existing photonic pressure sensing system, and to achieve high-precision pressure measurement at different temperatures.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] The present invention provides a photonic pressure sensing system that eliminates the influence of temperature, comprising an electro-optical conversion mixing unit, a photonic sensing unit, a photoelectric conversion processing unit, a signal processing unit, and a deep learning processing unit;

[0008] The electro-optical conversion mixer unit has a local oscillator signal input terminal, a microwave signal input terminal, and a mixed optical signal output terminal;

[0009] The photonic sensing unit has an input end and a sensing optical signal output end, and its input end is connected to the mixing optical signal output end of the electro-optical conversion mixer unit.

[0010] The photoelectric conversion processing unit has an input end and a sensing microwave signal output end, and its input end is connected to the sensing optical signal output end of the photonic sensing unit.

[0011] The signal processing unit has an input terminal and a signal parameter output terminal, and its input terminal is connected to the sensing microwave signal output terminal of the photoelectric conversion processing unit;

[0012] The deep learning processing unit has an input terminal and a pressure value output terminal, and its input terminal is connected to the signal parameter output terminal of the signal processing unit.

[0013] Furthermore, the electro-optical conversion mixing unit includes a broadband light source, a first light intensity modulator, and a second light intensity modulator;

[0014] Broadband light sources have a laser output terminal and are used to output broadband lasers.

[0015] The first optical intensity modulator has an input terminal, a local oscillator signal input terminal, a first bias voltage input terminal, and a modulated optical signal output terminal. Its input terminal is connected to the laser output terminal of the broadband light source, and the local oscillator signal input terminal and the first bias voltage input terminal are used to input the local oscillator signal and the first bias voltage, respectively.

[0016] The second optical intensity modulator has an input terminal, a microwave signal input terminal, a second bias voltage input terminal, and a mixing optical signal output terminal. Its input terminal is connected to the modulation optical signal output terminal of the first optical intensity modulator. The microwave signal input terminal and the second bias voltage input terminal are used to input microwave signals and second bias voltages, respectively.

[0017] Furthermore, the photonic sensing unit includes an optical isolator, a 2×2 optical coupler, a reference arm, a sensing arm, a first optical reflector, and a second optical reflector;

[0018] The optical isolator has an input terminal and an output terminal, and its input terminal is connected to the mixed optical signal output terminal of the electro-optical conversion mixer unit;

[0019] The 2×2 optical coupler has an input terminal, a first output terminal, and a second output terminal, and its input terminal is connected to the output terminal of the optical isolator.

[0020] The reference arm has an input end and an output end, and its input end is connected to the first output end of the 2×2 optocoupler;

[0021] The sensing arm has an input end and an output end. Its input end is connected to the second output end of a 2×2 optocoupler. External pressure and temperature act on the sensing arm.

[0022] The first light reflector has an input terminal, which is connected to the output terminal of the reference arm;

[0023] The second light reflector has an input end, which is connected to the output end of the sensing arm;

[0024] The 2×2 optical coupler also has a sensing optical signal output terminal.

[0025] Furthermore, the photoelectric conversion processing unit includes a photodetector, a microwave filter, and a microwave amplifier connected in sequence. The input end of the photodetector is connected to the sensing light signal output end of the photonic sensing unit, and the output end of the microwave amplifier is the sensing microwave signal output end.

[0026] The signal processing unit includes an analog-to-digital converter and a high-speed digital signal processor connected together. The input terminal of the analog-to-digital converter is connected to the sensing microwave signal output terminal of the photoelectric conversion processing unit, and the output terminal of the high-speed digital signal processor is the signal parameter output terminal.

[0027] Furthermore, the deep learning processing unit adopts a convolutional neural tangent kernel prediction model, whose input is the input of the deep learning processing unit and whose output is the stress value output of the deep learning processing unit.

[0028] The deep learning processing unit constructs a convolutional neural tangent kernel prediction model using a training dataset, which is a matrix of signal parameter detected under known pressure and temperature.

[0029] The convolutional neural tangent kernel prediction model consists of an input layer, Q infinite-width convolutional layers, a fully connected layer, and an output layer.

[0030] The input layer has an input end and an output end, and its input end is the input end of the convolutional neural tangent kernel prediction model;

[0031] Q infinite-width convolutional layers are connected sequentially, wherein the input of the first infinite-width convolutional layer is connected to the output of the input layer, and the Qth infinite-width convolutional layer has an output; Q is a positive integer greater than 1, and the filter size of the infinite-width convolutional layer is J×∞, where J is a positive integer greater than or equal to 1;

[0032] The fully connected layer has an input and an output, with its input connected to the output of the Qth infinite-width convolutional layer.

[0033] The output layer has an input and an output. Its input is connected to the output of the fully connected layer, and its output is the output of the convolutional neural tangent kernel prediction model.

[0034] The implementation method of the photonic pressure sensing system includes the following steps:

[0035] S1. Through the electro-optic conversion mixing unit, the broadband laser is sequentially cascaded and modulated with the local oscillator signal and the microwave signal, and a mixed optical signal is generated under the control of the bias voltage.

[0036] S2. Input the mixed optical signal into the photonic sensing unit. External pressure and temperature act on the photonic sensing unit, and output the sensing optical signal.

[0037] S3. The photoelectric conversion processing unit converts the sensing light signal into a sensing microwave signal.

[0038] S4. The signal processing unit detects the sensed microwave signal and extracts the signal parameters.

[0039] S5. The pressure values ​​at different temperatures are calculated based on the signal parameters using a deep learning processing unit.

[0040] Further, S1 includes: a broadband light source emitting broadband laser light, the broadband laser light being input into a first optical intensity modulator, with a frequency of... f L The local oscillator signal is input to the first optical intensity modulator. At the same time, under the control of the first bias voltage input to the first optical intensity modulator, the first optical intensity modulator modulates the intensity of the broadband laser and outputs a modulated optical signal.

[0041] The modulated optical signal is input to the second optical intensity modulator at a frequency of f R The microwave signal is input to the second optical intensity modulator, and under the control of the second bias voltage input to the second optical intensity modulator, the second optical intensity modulator remodulates the modulated optical signal and outputs a mixed optical signal.

[0042] The DC operating point of the first light intensity modulator is located at the quadrature point, and the amplitudes of the first-order sideband and the second-order sideband of the second light intensity modulator are equal.

[0043] Further, S2 includes: the mixed optical signal is input to the 2×2 optical coupler through the optical isolator and then split into two paths, one path is output through the reference arm and the other path is output through the sensing arm. The two output lights are reflected by the first optical reflector and the second optical reflector respectively, and then pass through the reference arm and the sensing arm respectively. Finally, the sensing optical signal is output by the 2×2 optical coupler.

[0044] Both the reference arm and the sensing arm contain optical fibers. The sensing arm is constructed by winding optical fibers on a regular geometry, and there is a difference in the length of the optical fibers between the reference arm and the sensing arm.

[0045] External pressure and temperature act on the sensing arm, and the output sensing light signal has a time delay. And satisfy:

[0046] ;

[0047] in, This is the initial delay time. k p This is the pressure delay coefficient. k T This is the temperature delay coefficient. P For the pressure of action, T The temperature at which it is applied;

[0048] Further, S3 includes: a photodetector that senses light signals input to a photodetector and outputs a converted microwave signal; the converted microwave signal is then passed through a microwave filter to remove frequencies in the converted microwave signal that are... f L - f R and f L -2 f R The component signal is output as a filtered microwave signal, which includes a frequency of [frequency value missing]. f L + f R and f L +2 f R The component signal; the filtered microwave signal is amplified by a microwave amplifier to output the sensing microwave signal;

[0049] S4 includes: a sensing microwave signal input analog-to-digital converter, which converts it into a digital signal; the digital signal is then subjected to a short-time fast Fourier transform by a high-speed digital signal processor, extracting a frequency of... f L + f R and f L +2 f R The amplitude and phase parameters of the component signals are extracted, and the extracted amplitude and phase parameters are output as signal parameters.

[0050] Furthermore, S5 includes: a deep learning processing unit employs a convolutional neural tangent kernel prediction model to calculate signal parameters and obtain pressure values ​​at different temperatures. Specifically: the signal parameters are input into the input layer of the convolutional neural tangent kernel prediction model for normalization. The result of the normalization is used to extract feature parameters through Q infinitely wide convolutional layers of the convolutional neural tangent kernel prediction model. The feature parameters are input into the fully connected layer of the convolutional neural tangent kernel prediction model for weighted summation, and then input into the output layer of the convolutional neural tangent kernel prediction model for output, thus obtaining pressure values ​​at different temperatures.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] The photonic pressure sensing system and implementation method proposed in this invention specifically involve modulating a microwave signal onto a broadband laser. Pressure and temperature are simultaneously applied to the mixed optical signal, which is then processed by photoelectric conversion to obtain a sensed microwave signal. High-precision detection and processing of the sensed microwave signal yields signal parameters, which are then processed by deep learning to calculate the pressure at different temperatures. This invention enables precise pressure measurement at varying temperatures and offers advantages such as high pressure measurement accuracy, insensitivity to temperature changes, high sensitivity, fast demodulation speed, high stability, strong adaptability, and a simple implementation architecture. It has significant application value for high-precision pressure measurement in fields such as biomedical photonic sensing. Attached Figure Description

[0053] Figure 1 This is a block diagram of the overall implementation of the photon pressure sensing system.

[0054] Figure 2 Block diagram for implementing the electro-optical conversion mixer unit.

[0055] Figure 3 Block diagram for the implementation of the photon sensing unit.

[0056] Figure 4 Block diagram for the photoelectric conversion processing unit.

[0057] Figure 5 Block diagram for the signal processing unit implementation.

[0058] Figure 6 A block diagram for implementing the deep learning processing unit.

[0059] Figure 7 Implement a block diagram for the training dataset.

[0060] Figure 8 A block diagram for the implementation of the convolutional neural tangent kernel prediction model. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0062] In the description of this invention, it should be noted that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0063] like Figure 1 As shown, the present invention provides a photonic pressure sensing system that eliminates the influence of temperature, comprising an electro-optical conversion mixing unit, a photonic sensing unit, a photoelectric conversion processing unit, a signal processing unit, and a deep learning processing unit. A broadband laser output from a broadband light source is passed through the electro-optical conversion mixing unit and sequentially cascaded with a local oscillator signal and a microwave signal, generating a mixed optical signal under bias voltage control. The mixed optical signal is input to the photonic sensing unit, where pressure and temperature act simultaneously, outputting a sensing optical signal. The photoelectric conversion processing unit converts the sensing optical signal into a sensing microwave signal and outputs it to the signal processing unit. The signal processing unit detects the sensing microwave signal and extracts signal parameters. Finally, the deep learning processing unit calculates the pressure values ​​at different temperatures based on the signal parameters.

[0064] like Figure 2 As shown, the electro-optical conversion mixing unit includes a broadband light source, a first light intensity modulator, and a second light intensity modulator.

[0065] The output frequency of the broadband light source is f 0 A broadband laser, the first optical intensity modulator receives the broadband laser and the external input frequency as follows: f L The local oscillator signal. The first bias voltage applied to the first light intensity modulator. V B_1 Under the control of [unclear], the first optical intensity modulator modulates the intensity of the input broadband laser using the local oscillator signal and outputs a modulated optical signal. The second optical intensity modulator receives the modulated optical signal output from the first optical intensity modulator and an externally input frequency [unclear]. f R The amplitude of the microwave signal is denoted as . V R The second bias voltage applied to the second light intensity modulator V B_2 Under the control of the second optical intensity modulator, the input modulated optical signal is remodulated using a microwave signal, and a mixed optical signal is output.

[0066] To eliminate the cross-sensitivity effect of temperature on pressure measurement, the present invention sets the operating conditions of the two light intensity modulators as follows.

[0067] The half-wave voltage of the first light intensity modulator is V p_1 The first bias voltage is V B_1 If the DC operating point of the first light intensity modulator is the quadrature point, then the first bias voltage applied to it... V B_1 The following relationship must be satisfied:

[0068] .

[0069] The half-wave voltage of the second light intensity modulator is V p_2 The second bias voltage is V B_2 For the second light intensity modulator, the amplitude of the input microwave signal is adjusted. V R Second bias voltage V B_2 This ensures that the amplitudes of the first-order and second-order sidebands in the output mixed optical signal are equal. The condition for the amplitudes of the first-order and second-order sidebands to be equal is determined by the following relationship:

[0070] .

[0071] In the formula, J 1(·) and J 2(·) represent the first-order Bessel function of the first kind and the second-order Bessel function of the first kind, respectively.

[0072] When the amplitudes of the first-order sideband and the second-order sideband are equal, the mixed optical signal contains components of different frequencies. The frequencies of the first-order sideband include: f 0 -( f L - f R ), f 0 +( f L - f R ), f 0 -( f L + f R )and f 0 +( f L +f R The frequencies of the second-order sideband include: f 0 -( f L -2 f R ), f 0 +( f L -2 f R ), f 0 -( f L +2 f R )and f 0 +( f L +2 f R These component signals of different frequencies, after passing through subsequent photonic sensing units and photoelectric conversion processing units, will generate sensing microwave signals carrying pressure and temperature information. Among them, the frequency is... f L + f R and f L +2 f R The component signals are selected as valid signals for subsequent parameter extraction and deep learning processing, thereby achieving joint sensing and decoupling of temperature and pressure.

[0073] like Figure 3 As shown, the photonic sensing unit includes an optical isolator, a 2×2 optical coupler, a reference arm, a sensing arm, a first optical reflector, and a second optical reflector.

[0074] The input of the optical isolator is connected to the output of the electro-optical conversion mixer unit to receive the mixed optical signal and prevent reflected light from returning to the preceding unit, ensuring unidirectional transmission of the optical signal. The input of the 2×2 optical coupler is connected to the output of the optical isolator. After the mixed optical signal enters the 2×2 optical coupler, it is split into a first output light and a second output light.

[0075] The input end of the reference arm is connected to the first output end of the 2×2 optocoupler to receive the first output light. The input end of the sensing arm is connected to the second output end of the 2×2 optocoupler to receive the second output light.

[0076] The first optical reflector is connected to the end of the reference arm, with its input end connected to the output end of the reference arm, and is used to reflect the first output light back to the reference arm. The second optical reflector is connected to the end of the sensing arm, with its input end connected to the output end of the sensing arm, and is used to reflect the first output light back to the sensing arm.

[0077] Both the reference arm and the sensing arm contain optical fibers. The sensing arm is constructed by winding a certain length of optical fiber on a regular geometric shape (such as a cylinder or cuboid), and there is a difference in the length of the optical fiber between the reference arm and the sensing arm to ensure an initial delay between the two reflected light signals.

[0078] External pressure and temperature act on the sensing arm. When the pressure and temperature change, the length of the optical fiber in the sensing arm changes, causing a change in the relative delay time between the two reflected light signals. This relative delay time is the delay time carried by the sensed light signal output by the photonic sensing unit. Delay time t satisfy:

[0079] ;

[0080] in, This is the initial delay time. k p This is the pressure delay coefficient. k T This is the temperature delay coefficient. P For the pressure of action, T The temperature at which the action takes place.

[0081] like Figure 4 As shown, the photoelectric conversion processing unit includes a photodetector, a microwave filter, and a microwave amplifier.

[0082] The input end of the photodetector is connected to the output end of the photonic sensing unit to receive the sensed light signal, convert the sensed light signal into an electrical signal, and output the converted microwave signal.

[0083] The input terminal of the microwave filter is connected to the output terminal of the photodetector. After the converted microwave signal is input into the microwave filter, the microwave filter filters the converted microwave signal, removing frequencies with frequencies of [insert frequency here]. f L - f R and f L -2 f R The component signal is output as a filtered microwave signal. The filtered microwave signal includes components with a frequency of... f L + f R and fL +2 f R The component signal.

[0084] The input of the microwave amplifier is connected to the output of the microwave filter. It receives the filtered microwave signal, amplifies it, and outputs a sensed microwave signal. The sensed microwave signal includes signals with frequencies of... f L + f R and f L +2 f R The component signal.

[0085] To obtain the response characteristics of the photon sensor, the frequency of the microwave signal... f R Initial delay time with photon sensing unit t The following relationship applies between 0 and 0:

[0086] .

[0087] By changing the frequency of the local oscillator signal f L It can sense the two frequency components of a microwave signal ( f L + f R and f L +2 f R The input frequency falls within the effective operating frequency range of the signal processing unit, thus matching the input requirements of the signal processing unit.

[0088] like Figure 5 As shown, the signal processing unit includes an analog-to-digital converter and a high-speed digital signal processor.

[0089] The input terminal of the analog-to-digital converter is connected to the output terminal of the photoelectric conversion processing unit to receive the sensed microwave signal and convert the sensed microwave signal into a high-precision digital signal.

[0090] The input terminal of the high-speed digital signal processor is connected to the output terminal of the analog-to-digital converter to receive digital signals and perform short-time fast Fourier transform on the digital signals to complete real-time spectrum analysis of the digital signals.

[0091] By using the Short-Time Fast Fourier Transform, the high-speed digital signal processor simultaneously extracts the frequency of the digital signal. f L + f R andf L +2 f R The amplitude and phase parameters of the two component signals are given. The frequency is... f L + f R The amplitude obtained from the component signal detection is denoted as A 1. Phase is denoted as ; frequency is f L +2 f R The amplitude obtained from the component signal detection is denoted as A 2. Phase is denoted as .

[0092] The high-speed digital signal processor will extract the amplitude parameters ( A 1. A 2) and phase parameters ( , The signal parameters are output to the deep learning processing unit.

[0093] The above four parameters ( A 1. A 2. , These constitute four input features, corresponding to the pressure and temperature information, providing the deep learning processing unit with the corresponding input features to perform pressure calculations at different temperatures.

[0094] like Figure 6 As shown, the deep learning processing unit uses a convolutional neural tangent kernel prediction model.

[0095] The input terminal of the deep learning processing unit is connected to the output terminal of the signal processing unit to receive the signal parameters (i.e., the frequency) output by the signal processing unit. f L + f R and f L +2 f R The amplitude parameters of the two component signals A 1. A 2 and phase parameters , Then, the calculated pressure values ​​at different temperatures are output.

[0096] like Figure 7 As shown, the convolutional neural tangent kernel prediction model needs to be trained using a training dataset. The training dataset is constructed as follows:

[0097] By changing the pressure and temperature, the corresponding measurement signal parameters are detected at different known pressure and temperature points. A data matrix is ​​constructed using the measurement signal parameters and the known pressure and temperature to obtain the training dataset.

[0098] Specifically, under known pressure P m (m=1,2,3,…,M) and temperature T n Given (n=1,2,3,…,N), the frequency is f L + f R The amplitude obtained from component signal detection is A mn_1 Phase is The frequency is f L +2 f R The amplitude obtained from the component signal detection is A mn_2 Phase is .

[0099] For M pressures and N temperatures, the training dataset is an MN×6 data matrix, represented as follows:

[0100] ;

[0101] For the training dataset, columns one through four ( A mn_1 , , A mn_2 , ) are the input features, columns five to six ( P m , T n ) represents the output result.

[0102] In the construction of the prediction model, a grid search method is adopted, and the minimum root mean square of the deviation between the predicted pressure and the actual pressure is calculated by gradient descent, thereby obtaining the optimal number and size of filters.

[0103] After training, the convolutional neural tangent kernel prediction model has obtained a complex nonlinear mapping relationship between pressure, temperature, and signal parameters. In actual measurements, this model can obtain the input four-dimensional signal parameters (…). A 1. A 2. , The system effectively separates the cross-sensitivity of temperature to pressure measurement, directly outputting pressure values ​​unaffected by temperature, thereby achieving high-precision pressure measurement under different temperature environments.

[0104] like Figure 8 As shown, the convolutional neural tangent kernel prediction model includes an input layer, Q infinite-width convolutional layers, a fully connected layer, and an output layer.

[0105] The input layer is used to receive the signal parameters output by the signal processing unit and to normalize the input signal parameters.

[0106] Q infinite-width convolutional layers are connected sequentially, with the input of the first infinite-width convolutional layer (infinite-width convolutional layer 1) connected to the output of the input layer. Q is a positive integer greater than 1, and the filter size of the infinite-width convolutional layer is J×∞, where J is a positive integer greater than or equal to 1.

[0107] Q infinite-width convolutional layers are used to extract feature parameters from the normalized signal parameters. Since the input signal parameters contain four dimensions ( A 1. A 2. , (This can be achieved by using convolution operations of infinitely wide convolutional layers) to accurately extract features from the input data.

[0108] The input of the fully connected layer is connected to the output of the Qth infinite-width convolutional layer (i.e., infinite-width convolutional layer Q), which is used to perform weighted summation of the features extracted by the convolutional layer to obtain the regression result.

[0109] The input of the output layer is connected to the output of the fully connected layer to output the regression results obtained from the fully connected layer, thus obtaining the pressure values ​​at different temperatures.

[0110] The present invention also provides a method for implementing a photonic pressure sensing system, comprising the following steps:

[0111] S1. Through the electro-optic conversion mixing unit, the broadband laser is sequentially cascaded and modulated with the local oscillator signal and the microwave signal, and a mixed optical signal is generated under the control of the bias voltage.

[0112] S2. Input the mixed optical signal into the photonic sensing unit. External pressure and temperature act on the photonic sensing unit, and output the sensing optical signal.

[0113] S3. The photoelectric conversion processing unit converts the sensing light signal into a sensing microwave signal.

[0114] S4. The signal processing unit detects the sensed microwave signal and extracts the signal parameters.

[0115] S5. The pressure values ​​at different temperatures are calculated based on the signal parameters using a deep learning processing unit.

[0116] The deep learning processing unit constructs a convolutional neural tangent kernel prediction model using a training dataset, which is a matrix of signal parameters detected under known pressure and temperature.

[0117] Finally, it should be noted that the above embodiments are merely preferred embodiments of the present invention used to illustrate the technical solutions of the present invention, and are not intended to limit the invention, nor are they intended to limit the patent scope of the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention. That is to say, any changes or refinements made to the main design concept and spirit of the present invention that are not of substantial significance, but whose technical problems are still consistent with the present invention, should be included within the protection scope of the present invention. In addition, the direct or indirect application of the technical solutions of the present invention to other related technical fields are similarly included within the patent protection scope of the present invention.

Claims

1. A photonic pressure sensing system that eliminates the influence of temperature, characterized in that, It includes an electro-optical conversion mixing unit, a photonic sensing unit, a photoelectric conversion processing unit, a signal processing unit, and a deep learning processing unit; The electro-optical conversion mixer unit has a local oscillator signal input terminal, a microwave signal input terminal, and a mixed optical signal output terminal; The photonic sensing unit has an input terminal and a sensing optical signal output terminal, and its input terminal is connected to the mixing optical signal output terminal of the electro-optical conversion mixer unit. The photoelectric conversion processing unit has an input end and a sensing microwave signal output end, and its input end is connected to the sensing optical signal output end of the photonic sensing unit. The signal processing unit has an input terminal and a signal parameter output terminal, and its input terminal is connected to the sensing microwave signal output terminal of the photoelectric conversion processing unit; The deep learning processing unit has an input terminal and a pressure value output terminal, and its input terminal is connected to the signal parameter output terminal of the signal processing unit. The electro-optic conversion mixing unit includes a broadband light source, a first light intensity modulator, and a second light intensity modulator; Broadband light sources have a laser output terminal and are used to output broadband lasers. The first optical intensity modulator has an input terminal, a local oscillator signal input terminal, a first bias voltage input terminal, and a modulated optical signal output terminal. Its input terminal is connected to the laser output terminal of the broadband light source, and the local oscillator signal input terminal and the first bias voltage input terminal are used to input the local oscillator signal and the first bias voltage, respectively. The second optical intensity modulator has an input terminal, a microwave signal input terminal, a second bias voltage input terminal, and a mixing optical signal output terminal. Its input terminal is connected to the modulation optical signal output terminal of the first optical intensity modulator, and the microwave signal input terminal and the second bias voltage input terminal are used to input microwave signals and second bias voltages, respectively. The photonic sensing unit includes an optical isolator, a 2×2 optical coupler, a reference arm, a sensing arm, a first optical reflector, and a second optical reflector. The optical isolator has an input terminal and an output terminal, and its input terminal is connected to the mixed optical signal output terminal of the electro-optical conversion mixer unit; The 2×2 optical coupler has an input terminal, a first output terminal, and a second output terminal, and its input terminal is connected to the output terminal of the optical isolator. The reference arm has an input end and an output end, and its input end is connected to the first output end of the 2×2 optocoupler; The sensing arm has an input end and an output end. Its input end is connected to the second output end of a 2×2 optocoupler. External pressure and temperature act on the sensing arm. The first light reflector has an input terminal, which is connected to the output terminal of the reference arm; The second light reflector has an input end, which is connected to the output end of the sensing arm; The 2×2 optical coupler also has a sensing optical signal output terminal.

2. The photonic pressure sensing system according to claim 1, characterized in that, The photoelectric conversion processing unit includes a photodetector, a microwave filter, and a microwave amplifier connected in sequence. The input end of the photodetector is connected to the sensing light signal output end of the photonic sensing unit, and the output end of the microwave amplifier is the sensing microwave signal output end. The signal processing unit includes an analog-to-digital converter and a high-speed digital signal processor connected together. The input terminal of the analog-to-digital converter is connected to the sensing microwave signal output terminal of the photoelectric conversion processing unit, and the output terminal of the high-speed digital signal processor is the signal parameter output terminal.

3. The photonic pressure sensing system according to claim 1, characterized in that, The deep learning processing unit uses a convolutional neural tangent kernel prediction model, whose input is the input of the deep learning processing unit and whose output is the stress value output of the deep learning processing unit. The deep learning processing unit constructs a convolutional neural tangent kernel prediction model using a training dataset, which is a matrix of signal parameter detected under known pressure and temperature. The convolutional neural tangent kernel prediction model consists of an input layer, Q infinite-width convolutional layers, a fully connected layer, and an output layer. The input layer has an input end and an output end, and its input end is the input end of the convolutional neural tangent kernel prediction model; Q infinite-width convolutional layers are connected sequentially, wherein the input of the first infinite-width convolutional layer is connected to the output of the input layer, and the Qth infinite-width convolutional layer has an output; Q is a positive integer greater than 1, and the filter size of the infinite-width convolutional layer is J×∞, where J is a positive integer greater than or equal to 1; The fully connected layer has an input and an output, with its input connected to the output of the Qth infinite-width convolutional layer. The output layer has an input and an output. Its input is connected to the output of the fully connected layer, and its output is the output of the convolutional neural tangent kernel prediction model.

4. A method for implementing the photon pressure sensing system as described in any one of claims 1 to 3, characterized in that, Includes the following steps: S1. Through the electro-optic conversion mixing unit, the broadband laser is sequentially cascaded and modulated with the local oscillator signal and the microwave signal, and a mixed optical signal is generated under the control of the bias voltage. S2. Input the mixed optical signal into the photonic sensing unit. External pressure and temperature act on the photonic sensing unit, and output the sensing optical signal. S3. The photoelectric conversion processing unit converts the sensing light signal into a sensing microwave signal. S4. The signal processing unit detects the sensed microwave signal and extracts the signal parameters. S5. The pressure values ​​at different temperatures are calculated based on the signal parameters using a deep learning processing unit.

5. The implementation method according to claim 4, characterized in that, S1 includes: A broadband light source emits broadband laser light, which is input to the first optical intensity modulator at a frequency of [frequency value missing]. f L The local oscillator signal is input to the first optical intensity modulator. At the same time, under the control of the first bias voltage input to the first optical intensity modulator, the first optical intensity modulator modulates the intensity of the broadband laser and outputs a modulated optical signal. The modulated optical signal is input to the second optical intensity modulator at a frequency of f R The microwave signal is input to the second optical intensity modulator, and under the control of the second bias voltage input to the second optical intensity modulator, the second optical intensity modulator remodulates the modulated optical signal and outputs a mixed optical signal. The DC operating point of the first light intensity modulator is located at the quadrature point, and the amplitudes of the first-order sideband and the second-order sideband of the second light intensity modulator are equal.

6. The implementation method according to claim 4, characterized in that, S2 includes: The mixed optical signal is input to the 2×2 optical coupler through the optical isolator and then split into two paths. One path is output through the reference arm and the other path is output through the sensing arm. The two output lights are reflected by the first and second optical reflectors respectively, and then pass through the reference arm and the sensing arm again respectively. Finally, the sensing optical signal is output by the 2×2 optical coupler. Both the reference arm and the sensing arm contain optical fibers. The sensing arm is constructed by winding optical fibers on a regular geometry, and there is a difference in the length of the optical fibers between the reference arm and the sensing arm. External pressure and temperature act on the sensing arm, and the output sensing light signal has a time delay. And satisfy: ； in, This is the initial delay time. k p This is the pressure delay coefficient. k T This is the temperature delay coefficient. P For the pressure of action, T The temperature at which the action takes place.

7. The implementation method according to claim 4, characterized in that, S3 include: The sensed optical signal is input to the photodetector, and the output is converted into a microwave signal; the converted microwave signal is then filtered by a microwave filter to remove frequencies in the converted microwave signal. f L - f R and f L -2 f R The component signal is output as a filtered microwave signal, which includes a frequency of [frequency value missing]. f L + f R and f L +2 f R Component signals; The filtered microwave signal is amplified by a microwave amplifier to output a sensing microwave signal. S4 includes: a microwave signal input analog-to-digital converter, which converts the input microwave signal into a digital signal; The digital signal is subjected to a short-time fast Fourier transform by a high-speed digital signal processor, with an extraction frequency of [frequency value missing]. f L + f R and f L +2 f R The amplitude and phase parameters of the component signals are extracted, and the extracted amplitude and phase parameters are output as signal parameters.

8. The implementation method according to claim 4, characterized in that, S5 includes: a deep learning processing unit that uses a convolutional neural tangent kernel prediction model to calculate signal parameters and obtain pressure values ​​at different temperatures. Specifically: the signal parameters are input into the input layer of the convolutional neural tangent kernel prediction model for normalization. The result of the normalization is used to extract feature parameters through Q infinitely wide convolutional layers of the convolutional neural tangent kernel prediction model. The feature parameters are input into the fully connected layer of the convolutional neural tangent kernel prediction model for weighted summation, and then input into the output layer of the convolutional neural tangent kernel prediction model for output, thus obtaining the pressure values ​​at different temperatures.

Images