CN110554006B - Multi-mode measurement method based on self-interference micro-ring resonant cavity optical sensor - Google Patents
Multi-mode measurement method based on self-interference micro-ring resonant cavity optical sensor Download PDFInfo
- Publication number
- CN110554006B CN110554006B CN201910834708.XA CN201910834708A CN110554006B CN 110554006 B CN110554006 B CN 110554006B CN 201910834708 A CN201910834708 A CN 201910834708A CN 110554006 B CN110554006 B CN 110554006B
- Authority
- CN
- China
- Prior art keywords
- neural network
- network model
- transmission
- sensing detection
- sensing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/41—Refractivity; Phase-affecting properties, e.g. optical path length
- G01N21/45—Refractivity; Phase-affecting properties, e.g. optical path length using interferometric methods; using Schlieren methods
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/41—Refractivity; Phase-affecting properties, e.g. optical path length
- G01N21/45—Refractivity; Phase-affecting properties, e.g. optical path length using interferometric methods; using Schlieren methods
- G01N2021/458—Refractivity; Phase-affecting properties, e.g. optical path length using interferometric methods; using Schlieren methods using interferential sensor, e.g. sensor fibre, possibly on optical waveguide
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2201/00—Features of devices classified in G01N21/00
- G01N2201/12—Circuits of general importance; Signal processing
- G01N2201/129—Using chemometrical methods
- G01N2201/1296—Using chemometrical methods using neural networks
Landscapes
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
A multi-mode measurement method based on a self-interference micro-ring resonant cavity sensor comprises the following steps: inputting a detection light source into a self-interference micro-ring resonant cavity optical sensor, detecting various detected substances to obtain transmission extinction values of the detected substances and obtain sensing detection quantity corresponding to the transmission extinction values, training a preset neural network model by using the transmission extinction values and the sensing detection quantity corresponding to the transmission extinction values as training data to obtain a trained neural network model, optimizing each parameter in the neural network model according to the sensing detection quantity to obtain an optimized neural network model, and measuring the sensing detection quantity by using the optimized neural network model. The method provided by the disclosure can effectively extract multi-mode sensing information to improve the detection sensitivity of the self-interference micro-ring resonant cavity, and skillfully avoids a complex multi-mode transmission spectrum acquisition process by combining machine learning algorithms such as an artificial neural network and the like.
Description
Technical Field
The disclosure relates to the field of optical microcavity sensing, in particular to a multi-mode measurement method based on a self-interference micro-ring resonant cavity optical sensor.
Background
The whispering gallery mode optical microcavity refers to an optical component which has a spatial and temporal local enhancement effect and a frequency selection effect on optical waves, and has wide application in the fields of active optical components, optical signal processing, optical interconnection, low-energy nonlinear optics, interaction of light and substances, sensing and the like.
In the patent (201610026981.6), a self-interference micro-ring resonator optical sensor is designed, which includes a micro-ring resonator with a probe arm waveguide coupled twice, and the mechanism is that the probe arm waveguide is affected by an external probe target to generate a small phase change, which results in a significant change in the waveguide and micro-ring out-coupling coefficient, and finally results in a change in the resonance wavelength and transmission extinction on the transmission spectrum. When the self-interference micro-ring resonant cavity optical sensor is used, a detection target is only required to be arranged on a detection arm or a micro-ring, and single-mode dissipative sensing can be realized by measuring the change of transmission extinction at a certain single-mode resonant wavelength. Theoretically, the single-mode dissipative sensing of the self-interference micro-ring resonant cavity optical sensor is proved to be capable of effectively avoiding the influence of frequency noise of a laser and effectively reducing the detection limit.
In a dissipation sensing mechanism of a self-interference micro-ring resonant cavity, in order to realize ultrahigh sensing sensitivity and ultralow detection limit, transmission valleys at different wavelength positions of light need to be measured, which relates to high-sensitivity measurement of transmission spectrum, and an expensive tuning laser or a precise optical spectrometer needs to be used. The use of expensive tuned lasers or precision optical spectrometers results in high sensor costs, which is one of the major constraints impeding commercialization. Therefore, on the premise of keeping high detection accuracy, a new sensing mechanism must be explored to effectively reduce the detection cost of the system, for example, multi-modal sensing is used to replace the original single-modal dissipative sensing measurement.
Disclosure of Invention
The disclosure provides a multi-mode measurement method based on a self-interference micro-ring resonant cavity optical sensor, and realizes a high-sensitivity multi-mode sensing detection method based on the self-interference micro-ring resonant cavity optical sensor.
One aspect of the present disclosure provides a multi-mode measurement method based on a self-interference micro-ring resonator optical sensor, including: inputting a detection light source into a self-interference micro-ring resonant cavity optical sensor, detecting multiple detected substances to obtain transmission extinction values of the detected substances, and obtaining sensing detection quantities corresponding to the transmission extinction values; taking the transmission extinction value and the sensing detection quantity corresponding to the transmission extinction value as training data, and training a preset neural network model to obtain a trained neural network model; optimizing each parameter in the neural network model according to the sensing detection quantity to obtain an optimized neural network model; and measuring the sensing detection quantity by using the optimized neural network model.
Optionally, the training a preset neural network model by using the transmission extinction value and the sensing detection amount corresponding to the transmission extinction value as training data includes: respectively carrying out normalization processing on the transmission extinction value and the sensing detection quantity, taking data obtained after the transmission extinction value is subjected to normalization processing as input, taking data obtained after the sensing detection quantity is subjected to normalization processing as output, training a preset neural network model, and establishing a mapping relation between the transmission extinction value and the sensing detection quantity.
Optionally, the optimizing the parameters in the neural network model according to the sensing detection quantity includes: collecting a plurality of groups of transmission extinction values and real sensing detection quantity corresponding to the transmission extinction values; respectively carrying out normalization processing on the transmission extinction value and the sensing detection quantity; inputting the data after normalization processing of the transmission extinction value into the neural network model to obtain a predicted sensing detection amount; calculating the mean square error between the obtained predicted sensing detection quantity and the real sensing detection quantity through an evaluation function; and adjusting parameters of the neural network model according to the mean square error.
Optionally, the merit function includes: let XtRepresenting the sensing quantities, Y, normalized in the training datatAfter the extinction value subjected to normalization processing is input into the neural network model, the neural network model predicts a sensing measurement value, t represents the number of training data sets, N represents the total number of training data, MSE represents an evaluation function, and then:
optionally, the measuring the sensing detection quantity by using the optimized neural network model includes: normalizing the transmission extinction value, and inputting the transmission extinction value serving as an input value into the neural network model after optimizing parameters; and performing inverse normalization processing on the output value of the neural network model to obtain the sensing detection quantity corresponding to the transmission extinction value.
Optionally, the method further comprises: and after the sensing detection quantity is obtained, judging the type of the detected object according to the sensing detection quantity.
Optionally, the detection light source is a broad spectrum light source.
Optionally, inputting the detection light source into the self-interference micro-ring resonant cavity optical sensor, and detecting the multiple detected substances to obtain the transmission extinction values of the detected substances includes: covering the detected substance on the optical detection arm waveguide of the self-interference micro-ring resonant cavity optical sensor; the wide-spectrum light source light is emitted into the self-interference micro-ring resonant cavity optical sensor, so that the wide-spectrum light source light is subjected to interference in the self-interference micro-ring resonant cavity optical sensor to form multi-mode resonance light waves, and the multi-mode resonance light waves are emitted from the self-interference micro-ring resonant cavity optical sensor; and collecting the multi-mode resonance light waves through a charge coupling device, and transmitting the multi-mode resonance light waves into a computer for processing to obtain a transmission extinction value of the measured substance.
Optionally, the transmission valleys of the transmission spectrum of the multi-modal resonant light waves do not have periodicity.
Optionally, the substance to be measured comprises an optically sensitive material in the form of a gas, liquid, solid.
The at least one technical scheme adopted in the embodiment of the disclosure can achieve the following beneficial effects:
the multi-mode measurement method based on the self-interference micro-ring resonant cavity optical sensor can effectively extract multi-mode sensing information to improve detection sensitivity, and skillfully avoids a complex multi-mode transmission spectrum acquisition process by combining machine learning algorithms such as an artificial neural network and the like. The self-interference micro-ring resonant cavity optical sensor used in the invention adopts a cheap wide-spectrum light source, avoids using a tuning laser or a precision optical spectrometer with high price to carry out emergent frequency spectrum detection on the sensor with high precision on the premise of keeping high detection sensitivity, greatly reduces the cost of a measurement system, and has important significance for promoting commercialization and practicability of the sensor.
Drawings
For a more complete understanding of the present disclosure and the advantages thereof, reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which:
FIG. 1 is a flow chart schematically illustrating a multi-mode measurement method based on a self-interference micro-ring resonator optical sensor according to an embodiment of the present disclosure;
FIG. 2 is a schematic diagram illustrating a self-interference micro-ring resonator optical sensor according to an embodiment of the disclosure;
fig. 3 schematically illustrates a schematic diagram of a transmission spectrum provided by an embodiment of the present disclosure;
FIG. 4 schematically illustrates a schematic diagram of a neural network model provided in accordance with an embodiment of the present disclosure;
FIG. 5 is a schematic diagram illustrating measurement performance of a neural network model in setting different number of hidden layer nodes according to an embodiment of the disclosure;
FIG. 6 is a schematic diagram illustrating measured performance of a neural network model in setting different learning rates according to an embodiment of the present disclosure;
FIG. 7 is a schematic diagram illustrating measured performance of a neural network model in setting different training target errors according to an embodiment of the present disclosure;
FIG. 8 is a schematic diagram illustrating measured performance of a neural network model using different sets of training data according to an embodiment of the present disclosure;
fig. 9 schematically illustrates a comparison graph of a sensing detection amount measured by a neural network model and theoretical simulation data provided according to an embodiment of the present disclosure.
Detailed Description
Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. It should be understood that the description is illustrative only and is not intended to limit the scope of the present disclosure. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present disclosure.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The words "a", "an" and "the" and the like as used herein are also intended to include the meanings of "a plurality" and "the" unless the context clearly dictates otherwise. The terms "comprises," "comprising," and the like, as used herein, specify the presence of stated features, steps, operations, and/or components, but do not preclude the presence or addition of one or more other features, steps, operations, or components.
All terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. It is noted that the terms used herein should be interpreted as having a meaning that is consistent with the context of this specification and should not be interpreted in an idealized or overly formal sense.
Some block diagrams and/or flow diagrams are shown in the figures. It will be understood that some blocks of the block diagrams and/or flowchart illustrations, or combinations thereof, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the instructions, which execute via the processor, create means for implementing the functions/acts specified in the block diagrams and/or flowchart block or blocks.
Fig. 1 schematically shows a flowchart of a multimode measurement method based on a self-interference micro-ring resonator optical sensor according to an embodiment of the present disclosure.
As shown in fig. 1, a multi-mode measurement method based on a self-interference micro-ring resonator optical sensor provided by an embodiment of the present disclosure includes the following steps.
And step S1, inputting a detection light source into the self-interference micro-ring resonant cavity optical sensor, detecting multiple detected substances to obtain the transmission extinction value of the detected substances, and obtaining the sensing detection quantity corresponding to the transmission extinction value.
In the embodiment of the present disclosure, the sensing detection amount may be equivalent to a variation of the arm length of the detection arm in the self-interference micro-ring resonator optical sensor.
The detection light source may be a broad spectrum light source.
The substance to be measured may comprise optically sensitive materials in gaseous, liquid, solid form.
When measuring the transmission extinction value, the following steps are included.
And covering the detected substance on the optical detection arm waveguide of the self-interference micro-ring resonant cavity optical sensor.
Light is emitted into the self-interference micro-ring resonant cavity optical sensor, so that wide-spectrum light source light is interfered in the self-interference micro-ring resonant cavity optical sensor to form multi-mode resonance light waves, and the multi-mode resonance light waves are emitted from the self-interference micro-ring resonant cavity optical sensor.
The charge coupled device is used for collecting multi-mode resonance light waves, and the multi-mode resonance light waves are transmitted to a computer for processing so as to obtain a transmission extinction value of the measured substance.
The transmission valleys of the transmission spectrum of the multi-modal resonant light waves do not have periodicity.
In the embodiment of the disclosure, the equivalent arm length variation of the detection arm of the self-interference micro-ring resonant cavity optical sensor corresponding to the transmission extinction value can be obtained through theoretical simulation.
And step S2, training the preset neural network model by taking the transmission extinction value and the sensing detection quantity corresponding to the transmission extinction value as training data to obtain the trained neural network model.
In this embodiment, a wide-spectrum light source is used as the detection light source, and after entering the self-interference micro-ring resonator optical sensor, multi-mode resonant light waves are generated, and the transmission valleys of the transmission spectrum of the multi-mode resonant light waves are non-periodically distributed, and the variations of the transmission extinction values of the light waves with different resonant wavelengths are also significantly different.
The detection arm waveguide of the self-interference micro-ring resonant cavity optical sensor is influenced by the refractive index of a measured substance, so that light generates tiny phase change during transmission, the change can be equivalently understood as the change of the optical path of the light transmitted in the detection arm, namely the change of the arm length of the detection arm, and the change causes the change of the transmission spectrum of the light source after the light source is emitted from the self-interference micro-ring resonant cavity optical sensor. Therefore, for the self-interference micro-ring resonant cavity optical sensor, when the measured substance changes, the optical path length of the optical detection arm waveguide changes, and the transmission valley value at different resonant wavelengths changes differently.
In the embodiment, the change of the transmission valley value at the multi-resonance wavelength is extracted within a certain wavelength range, the arm length change of the corresponding detection arm of the self-interference micro-ring resonant cavity optical sensor is obtained through experimental simulation, a neural network model is established to realize multi-mode sensing detection, and the measurement of the measured substance can be finally realized.
Respectively carrying out normalization processing on the transmission extinction value and the sensing detection quantity, taking data obtained after the transmission extinction value is subjected to normalization processing as input, taking data obtained after the sensing detection quantity is subjected to normalization processing as output, training a preset neural network model, and establishing a mapping relation between the transmission extinction value and the sensing detection quantity.
The data are normalized, so that the calculated amount can be reduced, and the calculation of the neural network model is more convenient and faster.
By establishing a mapping relation between the transmission extinction value and the sensing detection quantity and utilizing the neural network model, the sensing detection quantity (namely the arm length variation of the detection arm) can be predicted only by transmitting the extinction value, the complex multi-mode transmission spectrum acquisition process is avoided, the detection steps are simplified, and the hardware cost is reduced.
And step S3, optimizing each parameter in the neural network model according to the sensing detection quantity to obtain the optimized neural network model.
After the mapping relation between the transmission extinction value and the sensing detection quantity is established, all parameters in the neural network model can be optimized, so that the detection precision of the neural network model is improved.
Each parameter in the neural network model may include the number of input layer neurons, the number of hidden layer neurons, the number of output layer neurons, the learning rate, the training target error, etc. of the neural network model.
The parameter optimization of the neural network model comprises the following steps.
And collecting a plurality of groups of transmission extinction values and real sensing detection quantity corresponding to the transmission extinction values.
And respectively carrying out normalization processing on the transmission extinction value and the sensing detection quantity.
And inputting the data after normalization processing of the transmission extinction value into a neural network model to obtain the predicted sensing detection quantity.
And calculating the mean square error between the predicted sensing detection quantity and the actually acquired sensing detection quantity through the evaluation function.
The evaluation function includes:
let XtIndicating the normalized sensed quantity, Y, of training datatAfter the extinction value subjected to normalization processing is input to the neural network model, the neural network model predicts a sensing measurement value, t represents the number of training data sets, N represents the total number of training data, MSE represents an evaluation function, and then:
and adjusting parameters of the neural network model according to the mean square error.
According to the method, after the parameters of the neural network model are adjusted each time, the mean square error between the predicted sensing detection quantity and the actually acquired sensing detection quantity is recalculated, and the parameters of the neural network model are adjusted again according to the mean square error. The purpose of adjusting the parameters of the neural network model is to reduce the mean square error, and the smaller the mean square error is, the smaller the error between the sensing detection quantity predicted by the neural network model and the real sensing detection quantity is, and the more accurate the detection precision of the neural network model is.
And step S4, measuring the sensing detection quantity by using the optimized neural network model.
After the optimized neural network model is obtained, the neural network model can be used for measuring the sensing detection quantity, and the using steps comprise the following steps.
And after normalization processing is carried out on the transmission extinction value, the transmission extinction value is used as an input value and is input into the neural network model after the optimization parameters are input.
And performing inverse normalization processing on the output value of the neural network model to obtain the sensing detection quantity corresponding to the transmission extinction value.
And after the sensing detection quantity is obtained, judging the type of the detected object according to the sensing detection quantity.
Since the sensing detection amounts caused by different measured substances are different, what the measured substance is can be deduced reversely through the sensing detection amounts.
Fig. 2 schematically illustrates a schematic diagram of a self-interference micro-ring resonator optical sensor provided in an embodiment of the present disclosure.
As shown in fig. 2, the self-interference type micro-ring resonator optical sensor includes:
the optical sensor comprises an input waveguide 1, a micro-ring resonant cavity 2, an output waveguide 3 and an optical detection arm waveguide 4, wherein the input waveguide 1 and the output waveguide 3 are respectively coupled with the micro-ring resonant cavity 2 and are arranged at two sides of the micro-ring resonant cavity 2, one end of the input waveguide 1 is a light source access end of the whole optical sensor, the other end of the input waveguide 1 is connected with an input end of the optical detection arm waveguide 4 at the coupling position of the input waveguide 1 and the micro-ring resonant cavity 2, the output end of the optical detection arm waveguide 4 is connected with one end of the output waveguide at the coupling position of the output waveguide 3 and the micro-ring resonant cavity 2, and the other end of the output waveguide 3 is a sensing signal exit end.
The self-interference micro-ring resonant cavity optical sensor has the following mechanism: when light propagates through the probe arm waveguide 4, a slight phase change occurs due to the influence of the refractive index of the measured substance, resulting in a change in the resonance wavelength and the transmission extinction value in the transmission spectrum output by the micro-ring resonator.
The detection arm waveguide 4 of the self-interference micro-ring resonant cavity optical sensor is influenced by the refractive index of the measured substance, so that light generates tiny phase change during transmission, and the change can be equivalently understood as the change of the optical path of the light transmitted in the detection arm, namely the change of the arm length of the detection arm, and the change causes the change of the transmission spectrum of the light source after the light source is emitted from the self-interference micro-ring resonant cavity optical sensor.
On the basis, the method as shown in fig. 1, which is proposed by the embodiment of the present disclosure, establishes a neural network model based on the transmission extinction value and the sensing detection amount (i.e., the arm length of the detection arm), and the sensing detection amount can be reversely derived by measuring the transmission extinction value.
In the patent (201610026981.6), the light source used is generated by a laser, and the light source can only generate single-mode resonant light waves from the interference micro-ring resonant structure, so that single-mode dissipative sensing is realized. Theoretically, the single-mode dissipative sensing can effectively avoid the influence of frequency noise of a laser and effectively reduce the detection limit compared with corresponding responsive sensing. However, to achieve ultra-high sensing sensitivity and ultra-low detection limits, the solution provided by this patent involves highly sensitive measurement of the transmission spectrum, requiring an expensive tuned laser or precision optical spectrometer, which results in a high cost of using the sensor provided by this patent.
In the method as shown in fig. 1 provided in the embodiment of the present disclosure, a wide-spectrum light source is used as a detection light source, the cost of equipment for generating the wide-spectrum light source is low, and after the wide-spectrum light source is injected into the self-interference micro-ring resonator optical sensor, a multi-modal resonance light wave is generated, and sensing information carried by the multi-modal resonance light wave is significantly different from sensing information carried by a single-modal resonance light wave. The transmission valleys of the transmission spectrum of the single-mode resonant light wave are periodically distributed, the transmission valleys of the transmission spectrum of the multi-mode resonant light wave are non-periodically distributed, and the variation of the transmission extinction values of the light waves with different resonant wavelengths in the multi-mode resonant light wave is also obviously different. From an information theory perspective, the single-mode measurement may lose part of the effective sensing information. If the traditional frequency sweeping method of the tuned laser is used for detecting the transmission frequency spectrum of multiple modes of the self-interference micro-ring resonant cavity, the detection complexity is greatly increased. Therefore, when multi-mode measurement is carried out, the wide-spectrum light source is used, multi-mode sensing information can be effectively extracted to improve detection sensitivity, the detection process can be effectively simplified, and the cost of a detection system is reduced.
Example one
The working principle and application of the silicon nitride self-interference micro-ring resonant cavity optical sensor are described below by taking the silicon nitride self-interference micro-ring resonant cavity optical sensor as an example.
In the embodiment of the present disclosure, the self-interference micro-ring resonator sensor chip is composed of a Si3N4 waveguide based on SiO2 material, and the parameter settings are shown in table 1.
TABLE 1 silicon nitride self-interference micro-ring resonator parameters
If the measured substance is placed on the surface of the waveguide of the detection arm, the refractive index n of the waveguide of the detection arm is causedLAn additional phase delta phi is generated. The additional phase delta phi of the waveguide of the optical detection arm can be obtained by the optical path difference between the micro-ring resonant cavity and the waveguide of the detection arm, namely:
wherein n isRAnd nLRefractive indices of waveguide of micro-ring resonator and probe arm, respectively, LRAnd LLRespectively representing the lengths of half of the micro-ring resonator and the probe arm waveguide.
By using a transmission matrix method, the coupling coefficient of the whole micro-ring resonant cavity and the detection arm is modulated by the phase delta phi, and the coupling coefficient is as follows:
wherein the content of the first and second substances,
representing the external loss, k, of the micro-ring resonator taking into account the interference occurring in the probe arminRepresenting the intrinsic loss of the micro-ring resonator due to transmission loss, k representing the energy coupling coefficient of the directional coupling region, τrRepresenting the time that light travels back and forth one revolution in the micro-ring cavity.
The transmission coefficient of the self-interference micro-ring resonant cavity sensor is expressed as:
it can be seen that the transmission of light by the self-interference micro-ring resonator sensor is related to the phase change of the detection arm, and the small phase change can cause the sharp change of the transmission spectrum. Further the transmission coefficient may be expressed as,
assuming that the probe arm waveguide and the microring have the same propagation constant betaW＝βR＝(2π/λ)neff-ia, where the initial probe arm waveguide refractive index nL＝nR＝neffAnd α is a waveguide loss coefficient.
If the object to be detected is placed on the detection arm, the length of the detection arm will change (which can be equivalent to the change of the refractive index of the waveguide of the detection arm), and thus the transmission spectrum will change. The resonant wavelength of the self-interference micro-ring resonant cavity cannot be obtained analytically through the formula, so that the transmission spectrum of the self-interference micro-ring resonant cavity can be researched only through numerical calculation. The length of the detection arm satisfies LL＝LW+ L, wherein LW250 μm is the initial length of the probe arm wavelength and l is the change in probe arm length caused by the detection of a substance. When l ═ 10nm, the typical transmission spectrum can be obtained by substituting the parameters shown in table 1 into the above formula.
Fig. 3 schematically illustrates a schematic diagram of a transmission spectrum provided by an embodiment of the present disclosure.
As shown in fig. 3, the intervals between the resonance wavelengths represented by the spectrum are not equidistant, and the transmission extinction values vary in different resonance wavelength modes. For example, when the probe arm length is changed from l ═ 10nm to l ═ 10nm, the transmission extinction value decreases while the resonance wavelength is red-shifted in the mode near the wavelength 1443nm, and the transmission extinction value increases while the resonance wavelength is red-shifted in the mode near the wavelength 1448 nm. Therefore, when the detected target changes, the self-interference micro-ring resonant cavity optical sensor generates a plurality of resonance modes, and compared with a single mode, more effective sensing information is embodied.
Fig. 4 schematically illustrates a schematic diagram of a neural network model provided according to an embodiment of the present disclosure.
In the disclosed embodiments, a Back Propagation (BP) based neural network model is established, which is divided into an input layer, a hidden layer, and an output layer. Training a neural network model by using training data, wherein the neural network model takes a transmission extinction value normalized in the training data as input, and takes a detection arm length change (namely a sensing detection amount) corresponding to the transmission extinction value as unique output, so that the neural network model learns and stores a mapping rule between the two.
The parameters of the neural network model are shown in table 2.
TABLE 2 BP neural network parameters
The main parameters of the BP neural network model comprise an input layer, an output layer, a hidden layer node number, a learning rate and a target error. When the BP neural network model is used for measuring the arm length of the detection arm, the parameters are optimized firstly so as to improve the measurement precision.
In the disclosed embodiment, there are up to 28 transmitted extinction values in the wavelength range 1440nm to 1600nm when the probe arm length/is varied in the range [ -10nm, 10nm ]. Therefore, in order to accurately realize the measurement of the detection arm length variation, the embodiment of the invention determines the number of nodes of the input layer as 28, and when the number of the actually transmitted extinction values is less than the value, other input values are set as zero. Before training, normalization processing is carried out on the transmission extinction value, and the measured data is normalized to the range of [ -1, 1], so that higher measurement precision can be obtained. The number of hidden layer nodes of the established BP neural network model is the number of hidden layer nodes when the mean square error of the predicted sensing detection quantity and the real sensing detection quantity is minimum. When the node number of the hidden layer is too low, the situation that the neural network does not converge in the learning process can occur, and when the node number of the hidden layer is too high, the error of the neural network model can be reduced, the precision is improved, but the neural network is easy to be complicated, so that the training time of the neural network is prolonged, and the tendency of over-fitting is increased. The output of the neural network model is the sensing detection quantity, so the node number of the output layer is 1.
In the embodiment of the disclosure, a data set of the arm length variation of the probe arm is obtained through theoretical simulation, the initial value of the arm length variation l of the probe arm is-9.99688 nm, the maximum value is 9.99688nm, the value intervals are 0.02224nm, and the number of the probe arm is 900 groups. The arm length variation of the detection arm is substituted into the theoretical formula to obtain a corresponding transmission spectrum, so that a corresponding transmission extinction value is obtained. And carrying out normalization processing on the transmission extinction value and the corresponding detection arm length variation to obtain training data.
In the embodiment of the invention, the initial parameters of the neural network model are set as follows: the number of nodes of the hidden layer is 10, the learning rate is 0.01, and the error of the training target is 1 multiplied by 10-6The training data set number is initially determined as 100 groups of data (l is [ -9.999 nm: 0.202 nm: 9.999nm [)]). Because the output value of the sensing data detection model of the BP neural network has random fluctuation every time, a large number of simulations are carried out for determining the optimal parameter in the BP neural network, and the simulation times are set to be 500 times so as to find out the MSE average value under each specific condition.
As shown in fig. 5 to 9, the changes of mean square errors of the detected arm length variation measured by the neural network model and the detected arm length variation of the theoretical simulation when the number of hidden layer nodes, the learning rate, the training target error, and the number of training data of the neural network model are changed are respectively shown.
As shown in fig. 5, the learning rate of the neural network model is fixed to 0.01, and the training target error is fixed to 1 × 10-6When the training set group number is fixed to be 100, the condition of mean square error change when the hidden layer node number changes is obtained. As shown in the figure, when the number of hidden layer nodes is 15, the mean square error takes the minimum value, and therefore, the number of hidden layer nodes of the designed neural network model is set to be 15. On the basis, other neural network parameters are continuously optimized.
As shown in FIG. 6, the number of hidden layer nodes of the neural network model is fixed to 15, and the training target error is fixed to 1 × 10-6The number of training set groups is fixed to 100, and the change in mean square error due to the change in learning rate is obtained. As shown in the figure, when the learning rate is set to 1 × 10-5In this case, the mean square error of the data fitting performed by the neural network model is minimized, and thus the learning rate of the designed neural network model is set to 1 × 10-5. On the basis, other neural network parameters are continuously optimized.
As shown in fig. 7, the number of hidden layer nodes of the neural network model is 15, and the learning rate is 1 × 10-5The number of training set groups is set to 100, resulting in the case where the mean square error changes when the value of the training target error is changed. As shown, when the training target error reaches 1 × 10-10The mean square error of the data fitting performed by the neural network model is the lowest, so the target error is set to 1 × 10-10. On the basis, other neural network parameters are continuously optimized.
As shown in fig. 8, the number of hidden layer nodes of the neural network model is 15, and the learning rate is 1 × 10-5The training target error is set to 1 × 10-10And obtaining the change condition of the mean square error when the training data group number is changed. As shown, the mean square error reduction of the neural network model stabilizes to a minimum value when the number of training sets is 900. Thus, in an embodiment of the present invention, training of the neural network is completed using 900 sets of training data.
After the values of the parameters are determined, the performance of the neural network model is optimal, and the precision of the arm length variation of the detection arm predicted by using the neural network model is highest.
Fig. 9 schematically illustrates a comparison graph of a sensing detection amount measured by a neural network model and theoretical simulation data provided according to an embodiment of the present disclosure.
And (3) emitting a wide-spectrum light source into the self-interference micro-ring resonant cavity sensor, collecting light waves emitted from the self-interference micro-ring resonant cavity sensor through a charge coupling device, and transmitting the light waves into a computer for processing to obtain a transmission extinction value of the measured substance. And normalizing the collected transmission extinction value to be used as test data and inputting the test data into the neural network model. And performing inverse normalization processing on the output value of the neural network model to obtain a measured value of the arm length variation l of the detection arm corresponding to the transmission extinction value.
As shown in fig. 9, after training the neural network, the mean of the mean square error in the embodiment of the present invention reaches 1.56 × 10 after 500 times of simulation tests using three different sets of test data-13. The measured value (i.e., the predicted sensing amount) is compared with the arm length l of the probe arm (i.e., the true sensing amount) obtained by theoretical simulation, and it can be seen from the figure that the predicted sensing amount substantially coincides with the data point of the true sensing amount. The embodiment shows that the invention has good measurement performance, and the measured arm length change of the detection arm is completely consistent with the actual measurement performance. Because the arm length change of the detection arm has the generality of sensing measurement, high-sensitivity sensing measurement can be realized by using the model.
Those skilled in the art will appreciate that various combinations and/or combinations of features recited in the various embodiments and/or claims of the present disclosure can be made, even if such combinations or combinations are not expressly recited in the present disclosure. In particular, various combinations and/or combinations of the features recited in the various embodiments and/or claims of the present disclosure may be made without departing from the spirit or teaching of the present disclosure. All such combinations and/or associations are within the scope of the present disclosure.
While the disclosure has been shown and described with reference to certain exemplary embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the disclosure as defined by the appended claims and their equivalents. Accordingly, the scope of the present disclosure should not be limited to the above-described embodiments, but should be defined not only by the appended claims, but also by equivalents thereof.
Claims (8)
1. A multimode measurement method based on a self-interference micro-ring resonant cavity optical sensor is characterized by comprising the following steps:
inputting a wide-spectrum light source into a self-interference micro-ring resonant cavity optical sensor, detecting multiple detected substances to obtain transmission extinction values of the detected substances, and obtaining sensing detection quantity corresponding to the transmission extinction values, wherein the method comprises the following steps:
covering the detected substance on the optical detection arm waveguide of the self-interference micro-ring resonant cavity optical sensor;
the wide-spectrum light source light is emitted into the self-interference micro-ring resonant cavity optical sensor, so that the wide-spectrum light source light is subjected to interference in the self-interference micro-ring resonant cavity optical sensor to form multi-mode resonance light waves, and the multi-mode resonance light waves are emitted from the self-interference micro-ring resonant cavity optical sensor;
collecting the multi-mode resonance light waves through a charge coupling device, and transmitting the multi-mode resonance light waves into a computer for processing to obtain a transmission extinction value of the measured substance;
taking the transmission extinction value and the sensing detection quantity corresponding to the transmission extinction value as training data, and training a preset neural network model to obtain a trained neural network model;
optimizing each parameter in the neural network model according to the sensing detection quantity to obtain an optimized neural network model;
and measuring the sensing detection quantity by using the optimized neural network model.
2. The method according to claim 1, wherein the training a preset neural network model using the transmission extinction value and the sensing detection amount corresponding to the transmission extinction value as training data comprises:
respectively carrying out normalization processing on the transmission extinction value and the sensing detection quantity, taking data obtained after the transmission extinction value is subjected to normalization processing as input, taking data obtained after the sensing detection quantity is subjected to normalization processing as output, training a preset neural network model, and establishing a mapping relation between the transmission extinction value and the sensing detection quantity.
3. The method of claim 1, wherein optimizing the parameters in the neural network model based on the sensing metrics comprises:
collecting a plurality of groups of transmission extinction values and real sensing detection quantity corresponding to the transmission extinction values;
respectively carrying out normalization processing on the transmission extinction value and the sensing detection quantity;
inputting the data after normalization processing of the transmission extinction value into the neural network model to obtain a predicted sensing detection amount;
calculating the mean square error between the obtained predicted sensing detection quantity and the real sensing detection quantity through an evaluation function;
and adjusting parameters of the neural network model according to the mean square error.
4. The method of claim 3, wherein the merit function comprises:
let XtRepresenting the sensing quantities, Y, normalized in the training datatAfter the extinction value subjected to normalization processing is input into the neural network model, the neural network model predicts a sensing measurement value, t represents the number of training data sets, N represents the total number of training data, MSE represents an evaluation function, and then:
5. the method of claim 1, wherein measuring the sensing metric using the optimized neural network model comprises:
normalizing the transmission extinction value, and inputting the transmission extinction value serving as an input value into the neural network model after optimizing parameters;
and performing inverse normalization processing on the output value of the neural network model to obtain the sensing detection quantity corresponding to the transmission extinction value.
6. The method of claim 5, further comprising:
and after the sensing detection quantity is obtained, judging the type of the detected object according to the sensing detection quantity.
7. The method of claim 1, wherein the transmission valleys of the transmission spectrum of the multi-modal resonant light waves do not have periodicity.
8. The method of claim 1, wherein the substance to be measured comprises a gas, liquid, solid form of optically sensitive material.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910834708.XA CN110554006B (en) | 2019-09-04 | 2019-09-04 | Multi-mode measurement method based on self-interference micro-ring resonant cavity optical sensor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910834708.XA CN110554006B (en) | 2019-09-04 | 2019-09-04 | Multi-mode measurement method based on self-interference micro-ring resonant cavity optical sensor |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110554006A CN110554006A (en) | 2019-12-10 |
| CN110554006B true CN110554006B (en) | 2021-10-19 |
Family
ID=68738931
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910834708.XA Active CN110554006B (en) | 2019-09-04 | 2019-09-04 | Multi-mode measurement method based on self-interference micro-ring resonant cavity optical sensor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110554006B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111523376B (en) * | 2020-03-09 | 2023-07-28 | 浙江工业大学 | Self-interference micro-ring resonant cavity sensing classification identification method based on unsupervised learning |
| CN111721336A (en) * | 2020-03-09 | 2020-09-29 | 浙江工业大学 | Self-interference micro-ring resonant cavity sensing classification identification method based on supervised learning |
| CN111348611B (en) * | 2020-03-10 | 2023-05-09 | 西南大学 | Silicon-based microcavity-based neuron-like light pulse output system |
| US20240142303A1 (en) * | 2021-03-17 | 2024-05-02 | Agency For Science, Technology And Research | Integrated circuit spectrometer |
| CN114199785B (en) * | 2021-11-18 | 2023-09-26 | 国网浙江省电力有限公司诸暨市供电公司 | Echo wall microcavity sensing method based on GAN data enhancement |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015001421A3 (en) * | 2013-06-25 | 2015-05-21 | Alcatel Lucent | Wavelength tunable transmitter for twdm-pon and onu |
| CN105675545A (en) * | 2016-01-15 | 2016-06-15 | 浙江工业大学 | High-sensitivity intensity detection method based on self-interference type micro resonator cavity light sensor |
| WO2017055606A1 (en) * | 2015-09-30 | 2017-04-06 | Consiglio Nazionale Delle Ricerche - Cnr | Method for measuring the concentration of trace gases by scar spectroscopy |
| CN107045637A (en) * | 2016-12-16 | 2017-08-15 | 中国医学科学院生物医学工程研究所 | A kind of blood species identification instrument and recognition methods based on spectrum |
| CN108414448A (en) * | 2018-05-30 | 2018-08-17 | 苏州联讯仪器有限公司 | One kind being based on the cascade optical sensor of dual resonant cavity |
| CN108534986A (en) * | 2018-04-12 | 2018-09-14 | 安徽大学 | A kind of Multi-Longitudinal Mode laser resonant cavity FSR and laser resonant cavity cavity temperature measure of the change device and measurement method |
| CN108801981A (en) * | 2018-06-22 | 2018-11-13 | 集美大学 | Trace quantity liquid refractivity measuring device based on self-mixed interference and measurement method |
| CN109639359A (en) * | 2019-01-07 | 2019-04-16 | 上海交通大学 | Photon neural network convolutional layer chip based on micro-ring resonator |
| CN109855766A (en) * | 2019-01-21 | 2019-06-07 | 浙江工业大学 | A kind of heat dissipation rate measurement method based on the hot light generation of optical microresonator |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10657440B2 (en) * | 2015-10-26 | 2020-05-19 | International Business Machines Corporation | Optical synapse for neuromorphic networks |
-
2019
- 2019-09-04 CN CN201910834708.XA patent/CN110554006B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015001421A3 (en) * | 2013-06-25 | 2015-05-21 | Alcatel Lucent | Wavelength tunable transmitter for twdm-pon and onu |
| WO2017055606A1 (en) * | 2015-09-30 | 2017-04-06 | Consiglio Nazionale Delle Ricerche - Cnr | Method for measuring the concentration of trace gases by scar spectroscopy |
| CN105675545A (en) * | 2016-01-15 | 2016-06-15 | 浙江工业大学 | High-sensitivity intensity detection method based on self-interference type micro resonator cavity light sensor |
| CN107045637A (en) * | 2016-12-16 | 2017-08-15 | 中国医学科学院生物医学工程研究所 | A kind of blood species identification instrument and recognition methods based on spectrum |
| CN108534986A (en) * | 2018-04-12 | 2018-09-14 | 安徽大学 | A kind of Multi-Longitudinal Mode laser resonant cavity FSR and laser resonant cavity cavity temperature measure of the change device and measurement method |
| CN108414448A (en) * | 2018-05-30 | 2018-08-17 | 苏州联讯仪器有限公司 | One kind being based on the cascade optical sensor of dual resonant cavity |
| CN108801981A (en) * | 2018-06-22 | 2018-11-13 | 集美大学 | Trace quantity liquid refractivity measuring device based on self-mixed interference and measurement method |
| CN109639359A (en) * | 2019-01-07 | 2019-04-16 | 上海交通大学 | Photon neural network convolutional layer chip based on micro-ring resonator |
| CN109855766A (en) * | 2019-01-21 | 2019-06-07 | 浙江工业大学 | A kind of heat dissipation rate measurement method based on the hot light generation of optical microresonator |
Non-Patent Citations (4)
| Title |
|---|
| Experimental demonstration of dissipative sensing in a self-interference microring resonator;SHUAI WAN et al.;《Photonics Research》;20180731;第6卷(第7期);681-682页 * |
| Micro ring resonators as building blocks for an all-optical high-speed reservoir-computing bit-pattern-recognition system;Charis Mesaritakis et al.;《J. Opt. Soc. Am. B》;20131130;第30卷(第11期);3048-3055页 * |
| 电力系统用油水分检测技术讨论;米劲臣等;《2013年中国电机工程学会年会 工程科技II辑》;20131120;1-7页 * |
| 超小型带宽可调谐微环谐振腔;董文等;《光子学报》;20180831;第47卷(第8期);0823001页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110554006A (en) | 2019-12-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110554006B (en) | Multi-mode measurement method based on self-interference micro-ring resonant cavity optical sensor | |
| CN113853563B (en) | Apparatus and method for Gaussian bosch subsampling | |
| JP6260911B2 (en) | Coupled waveguides for low-speed optical sensor applications | |
| US11092746B1 (en) | Post-fabrication photonic performance tuning | |
| CN109990975A (en) | Detection system, debugging system and sensor based on optical microcavity mechanical mode | |
| CN101476978B (en) | Method for measuring geometric parameter of single mode fiber | |
| Ren et al. | Dissipative sensing with low detection limit in a self-interference microring resonator | |
| CN109253986B (en) | Double-ring optical sensor of cascade Fourier transform spectrometer | |
| Hochman et al. | Efficient and spurious-free integral-equation-based optical waveguide mode solver | |
| Tian et al. | Deep learning assisted microwave photonic dual-parameter sensing | |
| US7684658B2 (en) | Optimized optical resonator device for sensing applications | |
| Liao et al. | Theoretical study on simultaneous measurement of seawater temperature and salinity based on dual fiber interferometers combined with nonlinear decoupling algorithm | |
| CN104570219B (en) | A kind of integrated optical sensor based on period waveguide microcavity resonance interference effect | |
| CN111523376B (en) | Self-interference micro-ring resonant cavity sensing classification identification method based on unsupervised learning | |
| CN105180978A (en) | Optical sensor based on narrow-band light source and filtering characteristic adjustable element and method thereof | |
| CN111721336A (en) | Self-interference micro-ring resonant cavity sensing classification identification method based on supervised learning | |
| Liu et al. | Quantitative modeling of thermo-optic effects in optical networks-on-chip | |
| Yang et al. | A hybrid method for photonic crystal fiber polarization filter based on artificial neural network and genetic algorithms | |
| CN112255786B (en) | Waveguide parameter extraction method and system, and virtual wafer establishment method and system | |
| CN112432724B (en) | Stress sensor based on vernier effect of optical fiber resonant cavity and stress measurement method | |
| US11714004B2 (en) | Measurements using systems having multiple spectrometers | |
| CN113945154B (en) | Demodulation method of SNAP structure microcavity displacement sensing signal based on neural network | |
| Adamovsky et al. | Morphology-dependent resonances and their applications to sensing in aerospace environments | |
| Pérez-Ocón et al. | Fast single-mode characterization of optical fiber by finite-difference time-domain method | |
| Muhire et al. | Wide-angle emission grating using a supervised genetic optimization for LIDAR integration |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |