A displacement sensing system and displacement prediction method based on SNAP microcavity

Abstract

The present invention provides a displacement sensing system and displacement prediction method based on a SNAP microcavity, comprising: a laser, a polarization controller, a coupling waveguide, an optical fiber microcavity, a displacement device, a photodetector and a computer; the laser generates a frequency sweep laser, which enters a SNAP microcavity in the optical fiber microcavity through a polarization controller and a coupling waveguide, and uses a photodetector to detect an optical signal that is not coupled into the SNAP microcavity and converts it into an electrical signal, and then processes the obtained electrical signal through a computer; when the optical fiber microcavity moves, multiple groups of ratio parameters of second-order and above transmittances relative to first-order transmittances and their corresponding displacements are used to train and optimize a preset neural network model, and displacement prediction is performed using the optimized neural network model; the displacement sensing system can reduce the influence of environmental factors such as humidity on displacement sensing, and has good anti-noise capability.

Description

Technical Field

[0001] The present invention relates to the field of optical sensing technology, and more specifically, to a displacement sensing system and a displacement prediction method based on a SNAP microcavity. Background Art

[0002] With the continuous improvement of modern science and technology, micro-displacement measurement technology has been widely used in many fields of production and life. More and more precision microstructures play an important role in aerospace, national defense science and technology, industrial production and other fields, but there are still huge challenges in the internal measurement of these fine structures. As a high-performance optical resonant cavity, SNAP microcavity has the advantages of good processing consistency and easy identification of resonance spectrum, and has broad application prospects in such displacement measurement.

[0003] The current prior art discloses a displacement sensing system based on a double-chain SNAP structure microcavity array, which mainly includes a tunable laser, a polarization controller, a coupled waveguide, a double chain of a SNAP structure microcavity array, a displacement device, a photodetector, and a computer. The tunable laser generates a frequency sweep laser, which enters the double-chain SNAP structure microcavity through the polarization controller and the coupled waveguide, and the photodetector obtains the resonance spectrum and sends it to the computer for processing. When the microcavity moves, displacement sensing of half the length of the SNAP structure is realized based on the resonance spectrum characteristics of a single SNAP structure microcavity; each time the half SNAP structure length is crossed, the full range of displacement sensing is realized by switching the resonance spectrum generated by the SNAP microcavity in the double chain in turn. Although this displacement sensing system based on the double-chain SNAP structure microcavity array can achieve high-precision displacement sensing over a large range, the accuracy of the displacement sensing will also be affected when factors such as environmental humidity change, and the anti-noise performance is poor. Summary of the invention

[0004] In order to overcome the defect of poor noise resistance of the above-mentioned prior art, the present invention provides a displacement sensing system and displacement prediction method based on SNAP microcavity, which can effectively reduce the influence of external environmental factors such as humidity on displacement sensing and enhance the noise resistance of the displacement sensing system.

[0005] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0006] A displacement sensing system based on SNAP microcavity, comprising: a laser, a polarization controller, a coupled waveguide, an optical fiber microcavity, a displacement device, a photodetector and a computer;

[0007] The coupling waveguide is provided with an incident end and an output end, and the incident end and the output end are arranged opposite to each other; one end of the polarization controller is connected to the output end of the laser, and the other end of the polarization controller is connected to the output end of the coupling waveguide; one end of the photodetector is connected to the incident end of the coupling waveguide, and the other end of the photodetector is connected to the computer;

[0008] The optical fiber microcavity is transversely arranged between the incident end and the output end of the coupling waveguide, so that the laser light emitted by the laser is emitted from the output end of the coupling waveguide, passes through the optical fiber microcavity, and then enters the incident end of the coupling waveguide, and part of the laser light in the coupling waveguide enters the optical fiber microcavity through evanescent wave coupling;

[0009] The two ends of the optical fiber microcavity are fixedly arranged on the displacement device, and the displacement device drives the optical fiber microcavity to reciprocate along the axial direction;

[0010] The optical fiber microcavity includes a first connecting optical fiber, a SNAP microcavity, and a second connecting optical fiber that are sequentially connected and penetrated; on the longitudinal section of the optical fiber microcavity along the axial direction, protrusions are mirror-imaged at both ends of the upper surface of the SNAP microcavity, including a first protrusion and a second protrusion; the lower surface and the upper surface of the SNAP microcavity are symmetrical about the axis of the optical fiber microcavity; the first protrusion of the SNAP microcavity is connected to one end of the first connecting optical fiber, and the second protrusion of the SNAP microcavity is connected to one end of the second connecting optical fiber.

[0011] Preferably, the coupling waveguide is a micro-nano tapered optical fiber.

[0012] Preferably, the laser is a tunable laser.

[0013] Preferably, in the longitudinal section of the optical fiber microcavity along the axial direction, the upper surface of the SNAP microcavity is determined by the following function:

[0014] f(x,s)=sx s-2 (1+x s ) -2 [s-1-(1+s)x s ]

[0015]

[0016]

[0017] Among them, Δr eff (z) is the effective radius change of the SNAP microcavity; z is the axial length of the SNAP microcavity; L is the width between the first protrusion and the second protrusion in the SNAP microcavity x is the intermediate variable; λ c The wavelength of the laser emitted by the laser; n ris the refractive index of the SNAP microcavity; r0 is the initial radius of the SNAP microcavity; s is the shape parameter.

[0018] Preferably, the coupling waveguide always keeps in contact with the surface of the optical fiber microcavity.

[0019] The present invention also discloses a displacement prediction method, which uses the above-mentioned SNAP microcavity-based displacement sensing system, and the method comprises the following steps:

[0020] S1: Using the SNAP microcavity-based displacement sensing system to obtain multiple sets of transmittance ratio parameters and their corresponding displacements;

[0021] S2: normalizing the transmittance ratio parameter and the displacement respectively to obtain a normalized transmittance ratio parameter and a normalized displacement;

[0022] S3: sending the normalized transmittance ratio parameter to a preset neural network model for prediction to obtain a predicted displacement;

[0023] S4: setting a loss function according to the predicted displacement and the normalized displacement;

[0024] S5: Optimize the preset neural network model using the loss function to obtain an optimized neural network model;

[0025] S6: Obtain the transmittance ratio parameter to be predicted, and use the optimized neural network model to obtain the displacement corresponding to the transmittance ratio parameter to be predicted.

[0026] Preferably, in step S1, the specific method of obtaining multiple sets of transmittance ratio parameters and their corresponding displacements using the SNAP microcavity-based displacement sensing system is:

[0027] After the laser emits laser light and enters the polarization controller to adjust the polarization state, it is coupled into the SNAP microcavity through the coupling waveguide; the SNAP microcavity generates a resonant light wave signal, and the photodetector converts the resonant light wave signal into an electrical signal, which is processed by a computer into a coupled resonance spectrum, and the displacement of the coupling position and the transmittance of each order of the coupled resonance spectrum are obtained, thereby calculating the transmittance ratio parameter; the displacement device drives the optical fiber microcavity to move axially relative to the coupling waveguide, and the coupled resonance spectrum generated by the SNAP microcavity changes, and multiple sets of transmittance ratio parameters and their corresponding displacements are obtained according to the changed coupled resonance spectrum.

[0028] Preferably, the transmittance ratio parameter is the ratio of the second-order transmittance and above to the first-order transmittance.

[0029] Preferably, the loss function is:

[0030]

[0031] Where N represents the number of displacement groups, X y represents the normalized displacement of the tth group, Y y represents the predicted displacement of the tth group, and MSE represents the loss function value.

[0032] Preferably, in step S6, the transmittance ratio parameter to be predicted is obtained, and the specific method of obtaining the displacement corresponding to the transmittance ratio parameter to be predicted by using the optimized neural network model is:

[0033] The transmittance ratio parameter to be predicted is normalized and input into the optimized neural network model, and the output value of the optimized neural network model is denormalized to obtain a displacement corresponding to the transmittance ratio parameter to be predicted.

[0034] Compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0035] The present invention provides a displacement sensing system and a displacement prediction method based on a SNAP microcavity. A laser, a polarization controller and an output end of a coupling waveguide are connected in sequence, and an input end of the coupling waveguide, a photodetector and a computer are connected in sequence. An optical fiber microcavity is transversely arranged between an input end and an output end of the coupling waveguide, and a displacement device drives the optical fiber microcavity to reciprocate along an axial direction. The coupling waveguide and the optical fiber microcavity are coupled to generate a resonant light wave signal, and the photodetector converts the resonant light wave signal into an electrical signal, which is processed into a coupled resonance spectrum by a computer. Due to the convex parts of the SNAP microcavity at both ends of the optical fiber microcavity mirror-arranged, the first-order transmittance of the resonance spectrum between the convex parts at both ends does not change with the displacement. When the ambient humidity changes, the transmittances of each order shift up and down as a whole, and the transmittance valley values ​​of each order do not change relatively. Therefore, the ratio of the second-order and above transmittances to the first-order transmittance remains unchanged. Based on the invariant characteristic of the transmittance ratio, the anti-noise capability of the displacement sensing system is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a schematic structural diagram of a displacement sensing system based on a SNAP microcavity provided by the present invention.

[0037] Figure 2 The figure is a schematic diagram of the longitudinal cross-section structure of the optical fiber microcavity provided by the present invention along the axial direction.

[0038] Figure 3 This is a mathematical function diagram of the upper surface of the SNAP microcavity along the axial longitudinal section provided by the present invention.

[0039] Figure 4 This is a surface diagram of the first 30-order coupled resonance spectra of the SNAP microcavity provided by the present invention.

[0040] Figure 5This is a schematic diagram of the overall upward and downward shift of the transmittance of each order of the SNAP microcavity provided by the present invention due to environmental influences.

[0041] Figure 6 This is a curve diagram of the first-order transmittance of the SNAP microcavity provided by the present invention changing with displacement.

[0042] Figure 7 This is a curve diagram showing the variation of the first 30-order transmittance of the SNAP microcavity provided by the present invention with displacement.

[0043] Figure 8 This is a curve diagram showing the ratio of the second-order and above transmittance to the first-order transmittance of the SNAP microcavity provided by the present invention as a function of displacement.

[0044] Fig. 9 This is a diagram of the training and prediction errors when the SNAP microcavity provided by the present invention uses the first 30-order transmittance-displacement curves for displacement sensing.

[0045] Fig.10 This is a schematic diagram of the overall upward and downward shift of each order transmittance of a quadratic curve-shaped SNAP microcavity in a displacement sensing system based on a SNAP microcavity provided by the present invention.

[0046] Fig.11 A displacement sensing system based on a SNAP microcavity provided by the present invention uses a quadratic curve-shaped SNAP microcavity for training and prediction of displacement error diagrams.

[0047] Fig.12 A flow chart of a method for displacement prediction using a displacement sensing system based on a SNAP microcavity provided by the present invention.

[0048] Fig.13 This is the predicted displacement and normalized displacement error diagram of the normalized transmittance ratio parameter in the optimized neural network model.

[0049] Fig.14 It is the error diagram between the predicted displacement and the actual displacement of the transmittance ratio parameter to be predicted in the optimized neural network model.

[0050] 1-laser, 2-polarization controller, 3-coupled waveguide, 4-fiber microcavity, 5-displacement device, 6-photodetector, 7-computer, 41-first connecting optical fiber, 42-SNAP microcavity, 43-second connecting optical fiber, 421-first protrusion, 422-second protrusion. DETAILED DESCRIPTION

[0051] The drawings are for illustrative purposes only and should not be construed as limiting the present patent;

[0052] In order to better illustrate the present embodiment, some parts in the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product;

[0053] It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0054] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0055] Example 1

[0056] like Figure 1 As shown, this embodiment provides a displacement sensing system based on SNAP microcavity, including: a laser 1, a polarization controller 2, a coupled waveguide 3, an optical fiber microcavity 4, a displacement device 5, a photodetector 6 and a computer 7;

[0057] The coupling waveguide 3 is provided with an incident end and an output end, and the incident end and the output end are arranged opposite to each other; one end of the polarization controller 2 is connected to the output end of the laser 1, and the other end of the polarization controller 2 is connected to the output end of the coupling waveguide 3; one end of the photodetector 6 is connected to the incident end of the coupling waveguide 3, and the other end of the photodetector 6 is connected to the computer 7;

[0058] The fiber microcavity 4 is laterally arranged between the incident end and the output end of the coupling waveguide 3, so that the laser light emitted by the laser 1 is emitted from the output end of the coupling waveguide 3 through the fiber microcavity 4 and then injected into the incident end of the coupling waveguide 3, and part of the laser light in the coupling waveguide 3 is coupled into the fiber microcavity 4 through evanescent waves;

[0059] The two ends of the optical fiber microcavity 4 are fixedly arranged on the displacement device 5, and the displacement device 5 drives the optical fiber microcavity 4 to reciprocate along the axial direction;

[0060] like Figure 2 As shown, the optical fiber microcavity 4 includes a first connecting optical fiber 41, a SNAP microcavity 42, and a second connecting optical fiber 43 which are sequentially connected; on the longitudinal section of the optical fiber microcavity 41 along the axial direction, protrusions are mirror-imaged at both ends of the upper surface of the SNAP microcavity 42, including a first protrusion 421 and a second protrusion 422; the lower surface and the upper surface of the SNAP microcavity 42 are symmetrical about the axis of the optical fiber microcavity 4; the first protrusion 421 of the SNAP microcavity 42 is connected to one end of the first connecting optical fiber 41, and the second protrusion 422 of the SNAP microcavity 42 is connected to one end of the second connecting optical fiber 43.

[0061] In the specific implementation process, the system is set on a suitable platform, so that the laser generated by the laser 1 is emitted into the polarization controller 2, and the laser generated by the laser 1 is adjusted by the polarization controller 2 and then emitted from the output end of the coupling waveguide 3, and then coupled with the optical fiber microcavity 4, and the laser that meets the resonance condition is coupled into the SNAP microcavity 42, and then the laser that is not coupled into the SNAP microcavity 42 is detected by the photodetector 6, and the laser that is not coupled into the SNAP microcavity 42 is converted from an optical signal into an electrical signal and sent to the computer 7 for processing;

[0062] When the optical fiber microcavity 4 moves axially along with the displacement device 5, the coupling degree of each order axial mode will change, so the valley value of each order axial mode will also change accordingly; when the optical fiber microcavity 4 moves axially along the coupling waveguide 3, different coupling resonance spectra will be generated at different positions, and the coupling resonance spectra at different displacement nodes are recorded by the computer 7, and the transmittance of each order resonance mode can be obtained according to the coupling resonance spectra;

[0063] The middle part of the transmittance of the first-order resonance mode obtained from the recorded coupled resonance spectrum does not change with displacement, and when the coupling conditions of the coupled waveguide 3 change due to environmental factors such as humidity, the transmittance of each order resonance mode will shift up and down as a whole, while the valley value of the transmittance of each order resonance mode is relatively unchanged. Therefore, within a certain range, the ratio of the second-order and higher transmittance to the first-order transmittance can remain basically unchanged, and the anti-noise ability of the displacement sensing system based on the SNAP microcavity can be enhanced by the basically unchanged characteristic of this transmittance ratio.

[0064] Example 2

[0065] This embodiment provides a displacement sensing system based on a SNAP microcavity, comprising: a laser 1, a polarization controller 2, a coupled waveguide 3, an optical fiber microcavity 4, a displacement device 5, a photodetector 6 and a computer 7;

[0066] The coupling waveguide 3 is provided with an incident end and an output end, and the incident end and the output end are arranged opposite to each other; one end of the polarization controller 2 is connected to the output end of the laser 1, and the other end of the polarization controller 2 is connected to the output end of the coupling waveguide 3; one end of the photodetector 6 is connected to the incident end of the coupling waveguide 3, and the other end of the photodetector 6 is connected to the computer 7;

[0067] The fiber microcavity 4 is laterally arranged between the incident end and the output end of the coupling waveguide 3, so that the laser light emitted by the laser 1 is emitted from the output end of the coupling waveguide 3 through the fiber microcavity 4 and then injected into the incident end of the coupling waveguide 3, and part of the laser light in the coupling waveguide 3 is coupled into the fiber microcavity 4 through evanescent waves;

[0068] The two ends of the optical fiber microcavity 4 are fixedly arranged on the displacement device 5, and the displacement device 5 drives the optical fiber microcavity 4 to reciprocate along the axial direction;

[0069] The optical fiber microcavity 4 includes a first connecting optical fiber 41, a SNAP microcavity 42, and a second connecting optical fiber 43 which are sequentially connected and penetrated; on the longitudinal section of the optical fiber microcavity 41 along the axial direction, the upper surface of the SNAP microcavity 42 is provided with convex parts in mirror images at both ends, including a first convex part 421 and a second convex part 422; the lower surface and the upper surface of the SNAP microcavity 42 are symmetrical about the axis of the optical fiber microcavity 4; the first convex part 421 of the SNAP microcavity 42 is connected to one end of the first connecting optical fiber 41, and the second convex part 422 of the SNAP microcavity 42 is connected to one end of the second connecting optical fiber 43;

[0070] The laser 1 is a tunable laser used to generate a laser with a continuous wavelength. In this embodiment, the operating wavelength of the laser 1 is around 1550nm and the line width is 300kHz;

[0071] The coupling waveguide 3 is a micro-nano tapered optical fiber. In this embodiment, the coupling waveguide 3 is a tapered optical fiber with a tapered waist diameter of 2 μm;

[0072] like Figure 3 As shown, the shape of the SNAP microcavity 42 is determined by the following mathematical expression:

[0073] f(x,s)=sx s-2 (1+x s ) -2 [s-1-(1+s)x s ]

[0074]

[0075]

[0076] Among them, Δr eff (z) is the effective radius change of the SNAP microcavity 42; z is the axial length of the SNAP microcavity 42; L is the width between the first protrusion 421 and the second protrusion 422 in the SNAP microcavity 42 x is the intermediate variable; λ c is the wavelength of the laser emitted by laser 1; n r is the refractive index of the SNAP microcavity 42; r0 is the initial radius of the SNAP microcavity 42; s is the shape parameter;

[0077] The coupling waveguide 3 always keeps in contact with the surface of the optical fiber microcavity 4 .

[0078] In a specific implementation process, the system is placed on a suitable platform, so that the laser light generated by the laser 1 is emitted into the polarization controller 2, and the laser light generated by the laser 1 is adjusted by the polarization controller 2 and emitted from the output end of the coupling waveguide 3, and part of the laser light that meets the resonance condition is coupled into the SNAP microcavity 42, and then the laser light that is not coupled into the SNAP microcavity 42 is detected by the photodetector 6, and the laser light that is not coupled into the SNAP microcavity 42 is converted from an optical signal into an electrical signal and sent to the computer 7 for processing;

[0079] When the optical fiber microcavity 4 moves axially with the displacement device 5, the coupling degree of each order axial mode will change, so the valley value of each order axial mode will also change accordingly; Figure 4 As shown, when the optical fiber microcavity 4 moves axially along the coupling waveguide 3, different coupling resonance spectra will be generated at different positions. The different resonance spectra at different displacements are the basis of displacement sensing. The coupling resonance spectra at different displacement nodes are recorded by the computer 7, and the transmittance of each order resonance mode can be obtained according to the coupling resonance spectra.

[0080] like Figure 5 As shown, when the coupling conditions of the coupled waveguide 3 are changed due to environmental factors such as humidity, the transmittance of each order resonance mode will shift up and down as a whole. In this embodiment, the environmental influence is simulated by changing the imaginary part of the coupling parameter D, and the imaginary part of the parameter is changed from 0.011 to 0.0111. The resonance spectrum valley value shift corresponding to the normalized transmittance ratio parameter after the parameter change and the transmittance ratio parameter to be predicted is shown in the figure, which is about 0.01.

[0081] like Figure 6 As shown, the middle part of the first-order transmittance is relatively flat and basically does not change with the displacement. The difference between the transmittance of the middle part and the transmittance of the flatness edge is about 0.007. This characteristic is the basis for enhancing the anti-noise ability of this embodiment;

[0082] like Figure 7 As shown in the figure, the transmittance of each resonant mode changes regularly with the change of displacement, and this characteristic is the basis for realizing displacement sensing; at the same time, due to the symmetry of the optical fiber microcavity 4, the transmittance is symmetrically distributed; therefore, only half of the transmittance data can be used when performing displacement sensing, so this figure only shows the transmittance of each order of displacement of the optical fiber microcavity 4 from 0μm to 500μm;

[0083] like Figure 8 As shown, when the displacement of the optical fiber microcavity 4 changes, the ratio of the second-order and higher transmittances to the first-order transmittance remains basically unchanged. The influence of the upward and downward deviations of the overall transmittance on the displacement sensing can be reduced by the ratio, thereby enhancing the anti-noise capability of the displacement sensing system based on the SNAP microcavity.

[0084] Fig. 9 The error diagram when using the transmittance of the SNAP microcavity 42 instead of the transmittance ratio for displacement sensing. The left side of the figure shows the displacement error predicted by the normalized transmittance parameter using the optimized neural network model, and the right side shows the displacement error predicted by the transmittance parameter to be predicted using the optimized neural network model. It can be seen from the left figure that the prediction error of the left figure can be within 3×10 -3 μm, which shows that the learning ability of the preset neural network model is fine, but the displacement prediction error in the right figure is very large, close to 60μm, which means that displacement sensing is basically impossible using normalized transmittance parameters, and the accuracy and noise resistance are very poor;

[0085] Fig.10 Schematic diagram of the overall up-down shift of the transmittance of each order of the quadratic curve SNAP microcavity. The imaginary part of the coupling parameter is changed from 0.011 to 0.0111 in the same way. It can be seen from the figure that the offset of the transmittance is also about 0.01, which is the same as the offset of the SNAP microcavity 42.

[0086] like Fig.11 As shown in the figure, the left side of the figure shows the displacement error predicted by the optimized neural network model for the normalized transmittance ratio parameter, and the right side shows the displacement error predicted by the optimized neural network model for the transmittance ratio parameter to be predicted. It can be seen that the error in the left figure can be within 3×10 -3 μm, indicating that the model has completed good learning, but the displacement prediction error of the right figure is about within 14 μm. In comparison, the error of the quadratic curve SNAP microcavity when predicted by the transmittance ratio parameter to be predicted is larger, indicating that the anti-noise ability and displacement sensing accuracy of the quadratic curve SNAP microcavity are not as good as the SNAP microcavity 42 provided by the present invention;

[0087] In summary, a displacement sensing system based on a SNAP microcavity utilizes the fact that the middle part of the first-order transmittance of the SNAP microcavity 42 does not change with displacement. When factors such as environmental humidity cause the coupling conditions of the coupled waveguide to change, resulting in an upward and downward shift in the overall transmittance, the relative positions of the valley values ​​between the transmittance orders do not change substantially, thereby substantially maintaining the ratio of the high-order transmittance to the first-order transmittance. This feature can enhance the anti-noise capability of the displacement sensing system.

[0088] Example 3

[0089] like Fig.12 As shown, this embodiment discloses a displacement prediction method, which uses the SNAP microcavity-based displacement sensing system provided in Embodiment 1 or 2, and the method includes the following steps:

[0090] S1: Using the SNAP microcavity-based displacement sensing system to obtain multiple sets of transmittance ratio parameters and their corresponding displacements;

[0091] S2: normalizing the transmittance ratio parameter and the displacement respectively to obtain a normalized transmittance ratio parameter and a normalized displacement;

[0092] S3: sending the normalized transmittance ratio parameter to a preset neural network model for prediction to obtain a predicted displacement;

[0093] S4: setting a loss function according to the predicted displacement and the normalized displacement;

[0094] S5: Optimize the preset neural network model using the loss function to obtain an optimized neural network model;

[0095] S6: obtaining a transmittance ratio parameter to be predicted, and using the optimized neural network model to obtain a displacement corresponding to the transmittance ratio parameter to be predicted;

[0096] In step S1, the specific method of obtaining multiple sets of transmittance ratio parameters and their corresponding displacements using the SNAP microcavity-based displacement sensing system is as follows:

[0097] After the laser 1 emits laser light and enters the polarization controller 2 to adjust the polarization state, it is coupled into the SNAP microcavity 42 through the coupling waveguide 3; the SNAP microcavity 42 generates a resonant light wave signal, and the photodetector 6 converts the resonant light wave signal into an electrical signal, which is processed by the computer 7 into a coupled resonance spectrum, and the displacement of the coupling position and the transmittance of each order of the coupled resonance spectrum are obtained, thereby calculating the transmittance ratio parameter; the displacement device 5 drives the optical fiber microcavity 4 to move axially relative to the coupling waveguide 3, and the coupled resonance spectrum generated by the SNAP microcavity 42 changes, and multiple sets of transmittance ratio parameters and their corresponding displacements are obtained according to the changed coupled resonance spectrum;

[0098] The transmittance ratio parameter is the ratio of the second-order transmittance and above to the first-order transmittance;

[0099] The loss function is:

[0100]

[0101] Where N represents the number of displacement groups, X y represents the normalized displacement of the tth group, Y t represents the predicted displacement of the tth group, and MSE represents the loss function value;

[0102] In step S6, the transmittance ratio parameter to be predicted is obtained, and the specific method of obtaining the displacement corresponding to the transmittance ratio parameter to be predicted by using the optimized neural network model is:

[0103] The transmittance ratio parameter to be predicted is normalized and input into the optimized neural network model, and the output value of the optimized neural network model is denormalized to obtain a displacement corresponding to the transmittance ratio parameter to be predicted.

[0104] In the specific implementation process, the displacement sensing system based on the SNAP microcavity is first used to obtain the coupled resonance spectra at different positions, and then multiple sets of ratio parameters of the second-order and above transmittances relative to the first-order transmittances and their corresponding displacements are obtained according to the coupled resonance spectra; then the obtained transmittance ratio parameters and their corresponding displacements are normalized to obtain normalized transmittance ratio parameters and normalized displacements; then the normalized transmittance ratio parameters are input into a neural network model with a preset input layer of 30, an intermediate hidden layer of 35, and an output layer of 1 to predict the output predicted displacement; then the loss function is set according to the obtained predicted displacement and the normalized displacement:

[0105]

[0106] Where N represents the number of displacement groups, X t represents the normalized displacement of the tth group, Y t represents the predicted displacement of the tth group, and MSE represents the loss function value;

[0107] Then, the preset neural network model is optimized according to the error result calculated by the loss function, and the weights of each neural node in the preset neural network model are adjusted to obtain the optimized neural network model;

[0108] Then, a transmittance ratio parameter to be predicted is obtained, the transmittance ratio parameter to be predicted is normalized and input into the optimized neural network model, and the output value of the optimized neural network model is denormalized to obtain a displacement corresponding to the transmittance ratio parameter to be predicted;

[0109] like Fig.13 As shown in Figure 2, the normalized transmittance ratio parameter is sent to the optimized neural network model for prediction. The error accuracy between the predicted displacement and the normalized displacement can reach 4×10 -3 μm, which means that the preset neural network model can be well learned and optimized;

[0110] like Fig.14 As shown, the transmittance ratio parameter to be predicted is sent to the optimized neural network model for prediction. The error between the displacement corresponding to the transmittance ratio parameter to be predicted and the actual displacement is within 2μm, indicating that the optimized neural network model can well realize high-precision displacement sensing.

[0111] The preset neural network model can update the weights of each neural node of the neural network according to the loss function MSE. By training and optimizing the preset neural network model, the problem of decreased sensing performance caused by the asymmetric shape of the SNAP microcavity due to uncontrollable factors during the manufacturing process can be well solved. Therefore, the optimized neural network model can effectively enhance the anti-noise ability of the displacement sensing system based on the SNAP microcavity.

[0112] The same or similar reference numerals correspond to the same or similar components;

[0113] The terms used in the drawings to describe positional relationships are only used for illustrative purposes and should not be construed as limiting this patent;

[0114] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A displacement sensing system based on SNAP microcavity, characterized in that: include: Laser (1), polarization controller (2), coupled waveguide (3), optical fiber microcavity (4), displacement device (5), photodetector (6) and computer (7); The coupling waveguide (3) is provided with an incident end and an output end, and the incident end and the output end are arranged opposite to each other; one end of the polarization controller (2) is connected to the output end of the laser (1), and the other end of the polarization controller (2) is connected to the output end of the coupling waveguide (3); one end of the photodetector (6) is connected to the incident end of the coupling waveguide (3), and the other end of the photodetector (6) is connected to the computer (7); The optical fiber microcavity (4) is transversely arranged between the output end and the input end of the coupling waveguide (3), so that the laser light emitted by the laser (1) is emitted from the output end of the coupling waveguide (3), passes through the optical fiber microcavity (4), and then is injected into the input end of the coupling waveguide (3), and part of the laser light in the coupling waveguide (3) enters the optical fiber microcavity (4) through evanescent wave coupling; Both ends of the optical fiber microcavity (4) are fixedly arranged on a displacement device (5), and the displacement device (5) drives the optical fiber microcavity (4) to reciprocate along the axial direction; The optical fiber microcavity (4) comprises a first connecting optical fiber (41), a SNAP microcavity (42) and a second connecting optical fiber (43) which are sequentially connected and penetrated; on the longitudinal section of the optical fiber microcavity (4) along the axial direction, convex portions are arranged at both ends of the upper surface of the SNAP microcavity (42) in a mirror-image manner, including a first convex portion (421) and a second convex portion (422); the lower surface and the upper surface of the SNAP microcavity (42) are symmetrical about the axis of the optical fiber microcavity (4); the first convex portion (421) of the SNAP microcavity (42) is connected to one end of the first connecting optical fiber (41), and the second convex portion (422) of the SNAP microcavity (42) is connected to one end of the second connecting optical fiber (43); In the longitudinal section of the optical fiber microcavity (4) along the axial direction, the upper surface of the SNAP microcavity (42) is determined by the following function: f(x,s)=sx s-2 (1+x s ) -2 [s-1-(1+s)x s ] Among them, Δr eff (z) is the change in the effective radius of the SNAP microcavity (42); z is the axial length of the SNAP microcavity (42); L is the width between the first protrusion (421) and the second protrusion (422) in the SNAP microcavity (42); x is the intermediate variable; λ c is the wavelength of the laser emitted by the laser (1); n r is the refractive index of the SNAP microcavity (42); r0 is the initial radius of the SNAP microcavity (42); and s is the shape parameter.

2. A displacement sensing system based on SNAP microcavity according to claim 1, characterized in that: The coupling waveguide (3) is a micro-nano tapered optical fiber.

3. A displacement sensing system based on SNAP microcavity according to claim 2, characterized in that: The laser (1) is a tunable laser.

4. A displacement sensing system based on SNAP microcavity according to claim 1, characterized in that: The coupling waveguide (3) is always in contact with the surface of the optical fiber microcavity (4).

5. A displacement prediction method, characterized in that: A displacement sensing system based on a SNAP microcavity according to any one of claims 1 to 4 is applied, wherein the method comprises the following steps: S1: Using the SNAP microcavity-based displacement sensing system to obtain multiple sets of transmittance ratio parameters and their corresponding displacements; S2: normalizing the transmittance ratio parameter and the displacement respectively to obtain a normalized transmittance ratio parameter and a normalized displacement; S3: sending the normalized transmittance ratio parameter to a preset neural network model for prediction to obtain a predicted displacement; S4: setting a loss function according to the predicted displacement and the normalized displacement; S5: Optimize the preset neural network model using the loss function to obtain an optimized neural network model; S6: Obtain the transmittance ratio parameter to be predicted, and use the optimized neural network model to obtain the displacement corresponding to the transmittance ratio parameter to be predicted.

6. A displacement prediction method according to claim 5, characterized in that: In step S1, the specific method of obtaining multiple sets of transmittance ratio parameters and their corresponding displacements using the SNAP microcavity-based displacement sensing system is as follows: The laser (1) emits laser light, which enters the polarization controller (2) to adjust the polarization state, and then is coupled into the SNAP microcavity (42) through the coupling waveguide (3); the SNAP microcavity (42) generates a resonant light wave signal, and the photoelectric detector (6) converts the resonant light wave signal into an electrical signal, which is processed into a coupled resonance spectrum by a computer (7), and the displacement of the current coupling position and the transmittance of each order of the coupled resonance spectrum are obtained, thereby calculating the transmittance ratio parameter; the displacement device (5) drives the optical fiber microcavity (4) to move axially relative to the coupling waveguide (3), and the coupled resonance spectrum generated by the SNAP microcavity (42) changes, and multiple groups of transmittance ratio parameters and their corresponding displacements are obtained according to the changed coupled resonance spectrum.

7. A displacement prediction method according to claim 6, characterized in that: The transmittance ratio parameter is the ratio of the second-order and above transmittance to the first-order transmittance.

8. A displacement prediction method according to claim 7, characterized in that: The loss function is: Where N represents the number of displacement groups, X t represents the normalized displacement of the tth group, Y t represents the predicted displacement of the tth group, and MSE represents the loss function value.

9. A displacement prediction method according to claim 8, characterized in that: In step S6, the transmittance ratio parameter to be predicted is obtained, and the specific method of obtaining the displacement corresponding to the transmittance ratio parameter to be predicted by using the optimized neural network model is: The transmittance ratio parameter to be predicted is normalized and input into the optimized neural network model, and the output value of the optimized neural network model is denormalized to obtain a displacement corresponding to the transmittance ratio parameter to be predicted.

Images