Description:A MATCHED FILTERING BASED SYSTEM AND METHOD FOR DETERMINING MULTI-PARAMETERS OF AN OPTICAL SENSOR
FIELD OF THE INVENTION
The present invention relates to the field of determining parameters of optical sensors. More specifically, the present invention relates to a matched filtering based system and method for determining multi-parameters of an optical sensor
BACKGROUND OF THE INVENTION
Traditional optical sensor setups often struggled with limitations in accurately determining multiple parameters concurrently. These systems typically relied on simplistic single-parameter detection methods, which were not well-suited for capturing the complexity of real-world environments where multiple factors influence sensor readings simultaneously. Consequently, the inability to perform comprehensive multi-parameter detection hindered the practical applicability of these systems in various scenarios where nuanced data analysis was essential.
Moreover, prior art systems frequently lacked robustness and adaptability, leading to challenges in maintaining reliable performance under diverse environmental conditions. Fixed configurations and limited scalability made it difficult to accommodate changes in sensor requirements or operating conditions without extensive reconfiguration or system redesign. This inflexibility not only increased costs but also hindered innovation and hindered the development of adaptable sensor solutions capable of meeting evolving needs.
Furthermore, traditional optical sensor setups often suffered from limitations in sensitivity and accuracy, particularly when dealing with subtle variations in environmental conditions. The reliance on simplistic signal processing techniques resulted in challenges in distinguishing relevant signals from noise, leading to compromised accuracy and reliability in sensor measurements. These limitations undermined the practical utility and effectiveness of traditional optical sensor systems in various industrial, scientific, and field applications, where precision and reliability are paramount.
Overall, the proposed invention aims to overcome these shortcomings by introducing a matched filtering-based system and method for determining multi-parameters of an optical sensor. By integrating advanced techniques such as transfer learning-based machine learning and linear regression analysis, the invention offers enhanced accuracy, efficiency, and adaptability in multi-parameter detection. These advancements pave the way for more robust and versatile optical sensor solutions capable of meeting the demands of diverse applications in today's dynamic environments.

SUMMARY OF THE INVENTION
In view of the foregoing disadvantages inherent in the prior art, the general purpose of the present disclosure is to provide a multimodal authentication system and a method thereof, to include all advantages of the prior art, and to overcome the drawbacks inherent in the prior art.
Some of the objects of the present disclosure, which at least one embodiment herein satisfies, are as follows:
An object of the present disclosure is to ameliorate one or more problems of the prior art or to at least provide a useful alternative. An object of the present disclosure is to provide a matched filtering based system for determining multiple parameters of an optical sensor.
Another object of the present disclosure is to provide a matched filtering based method for determining multiple parameters of an optical sensor.
Another object of the present disclosure is to provide a system and method for increasing accuracy and efficiency to determine the multi-parameters of the optical sensor.
Other objects and advantages of the present disclosure will be more apparent from the following description, which is not intended to limit the scope of the present disclosure.
In view of the above objects, in one aspect, the current disclosure provides a multimodal authentication system and a multimodal authentication method that is a robust and novel phishing-resistant authentication tool.
In an aspect of the present disclosure, a matched filtering based system for determining multiple parameters of an optical sensor facilitates a light source emitting light towards a Y-coupler, which directs the light towards primary and secondary fiber Bragg gratings (FBGs) connected in series. Light reflected from the gratings is analyzed by an optical spectrum analyzer, and data is stored in memory.
According to an embodiment of the present disclosure, an AI module utilizes matched filtering and cross-correlation to identify zero-crossing points, followed by multi-parameter detection protocols to evaluate sensor parameters. These parameters may include strain and temperature, determined through techniques such as transfer learning-based machine learning and linear regression analysis. Furthermore, a computing unit is employed to visualize the evaluated data.
In another aspect of the present disclosure, a matched filtering based method for determining multiple parameters of an optical sensor include steps of emitting light towards primary and secondary FBGs via a coupler, applying strain and temperature to the gratings, receiving reflected light, storing spectrum data, and utilizing multi-parameter detection protocols for evaluation. Specific parameters include a light source wavelength range of 1300 nm to 1600 nm, FBG central peak wavelengths between 1547.5 nm to 1552.5 nm, with primary at 1549.5 nm and secondary at 1550.5 nm. Strain and temperature are applied at 566 microstrain and 30 to 50 degrees Celsius, respectively. Multi-parameter detection protocols involve transfer learning-based machine learning and linear regression analysis, with 70% of data used for training and 30% for testing. These methods ensure accurate evaluation of sensor parameters.

BRIEF DESCRIPTION OF DRAWING
The foregoing summary, as well as the following detailed description of various embodiments, is better understood when read in conjunction with the drawings provided herein. For the purposes of illustration, there are shown in the drawings exemplary embodiments; however, the presently disclosed subject matter is not limited to the specific methods and instrumentalities disclosed.
Figure 1 illustrates a block diagram of a matched filtering based system for determining multi-parameters of an optical sensor as disclosed in the present disclosure;
Figure 2 illustrates a block diagram an AI module associated with the proposed system a according to one of embodiments of the present disclosure;
Figure 3 illustrate a schematic diagram of network structure for detection measurand parameter as disclosed in an embodiment of the present disclosure;
Figure 4 illustrate the graphical representation of simulation reflected spectrum of two cascaed FBG as disclosed in an embodiment of thepresent dislosure;
Figure 5 depicts the gaussian signal used as reference signals with the same shape and the same wavelength range but the peak may be differ than the FBGs as disclosed in an embodiment of the present disclosure;
Figure 6 depicts the different correlation coefficient and zero crossing point mentioned by dark bauble due application of strain and temperature, and the intensity response corresponds to the different zero crossing point for primary FBG and secondry FBG as disclosed in present disclosure;
Figure 7 depicts a tabular form of the data related to mean square error for the training, testing and validation performed by AI module;
Figure 8 (a) depicts graphical representation the normalized experimental reflected spectrum and Figure 8 (b) is assumed reference gaussian signal as disclosed in present disclosure;
Figure 9 (a) depicts graphical representation of the zero-crossing point for cross correlation coefficient plots and 9 (b) depicts the intensity response is plotted as disclosed in an embodiment of the present disclosure;
Figure 10 depicts graphical representation of validation performance of training data performed by the AI module;
Figure 11 depicts graphical representation of training state analysis of the AI module; and
Figure 12 depicts graphical representation of the regression analysis in terms of training, testing and validation performance done by AI module.
Like reference numerals refer to like parts throughout the description of several views of the drawing.
DETAILED DESCRIPTION OF THE INVENTION
Embodiments are provided so as to thoroughly and fully convey the scope of the present disclosure to the person skilled in the art. Numerous details are set forth, relating to specific components, and methods, to provide a complete understanding of embodiments of the present disclosure. It will be apparent to the person skilled in the art that the details provided in the embodiments should not be construed to limit the scope of the present disclosure. In some embodiments, well-known processes, well- known apparatus structures, and well-known techniques are not described in detail.
The terminology used, in the present disclosure, is only for the purpose of explaining a particular embodiment and such terminology shall not be considered to limit the scope of the present disclosure. As used in the present disclosure, the forms "a," "an," and "the" may be intended to include the plural forms as well, unless the context clearly suggests otherwise. The terms "comprises," "comprising," "including," and "having," are open-ended transitional phrases and therefore specify the presence of stated features, integers, steps, operations, elements, modules, units and/or components, but do not forbid the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. The particular order of steps disclosed in the method and process of the present disclosure is not to be construed as necessarily requiring their performance as described or illustrated. It is also to be understood that additional or alternative steps may be employed.
The following detailed description should be read with reference to the drawings, in which similar elements in different drawings are identified with the same reference numbers. The drawings, which are not necessarily to scale, depict illustrative embodiments and are not intended to limit the scope of the disclosure.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed. In this application, the use of the singular includes the plural, the word "a" or "an" means "at least one", and the use of "or" means "and/or", unless specifically stated otherwise. Furthermore, the use of the term "including", as well as other forms, such as "includes" and "included", is not limiting. Also, terms such as "element" or "component" encompass both elements and components comprising one unit and elements or components that comprise more than one unit unless specifically stated otherwise.
Furthermore, the term “module”, as used herein, refers to logic embodied in hardware or firmware, or to a collection of software instructions, written in a programming language, such as, Java, C, C++, python, or assembly. One or more software instructions in the modules can be embedded in firmware, such as in an EPROM. The modules described herein can be implemented as either software and/or hardware modules and can be stored in any type of non-transitory computer-readable medium or other storage device. Some non-limiting examples of non-transitory computer-readable media include CDs, DVDs, BLU-RAY, flash memory, and hard disk drives.
In an aspect, the present disclosure facilitates a matched filtering based system for determining multiple parameters of an optical sensor that includes a light source 100 emits light directed towards a Y-coupler 102, which then guides the light towards primary 104 and secondary 108 fiber Bragg gratings (FBGs) arranged in series. The reflected light from these gratings undergoes analysis using an optical spectrum analyzer, with the resulting data being stored in memory (as illustrated in Figure 1).
Furthermore, an AI module 110, which employs matched filtering and cross-correlation techniques to pinpoint zero-crossing points within the analyzed data. Subsequently, multi-parameter detection protocols are applied to assess various sensor parameters. These parameters, such as strain and temperature, are determined through sophisticated methodologies like transfer learning-based machine learning and linear regression analysis. Additionally, a computing unit 112 is incorporated to visually represent the analyzed data.
In an another aspect of the present disclosure, a matched filtering based method for determining multiple parameters of an optical sensor is disclosed involves the steps of directing emitting light towards primary 104 and secondary 108 FBGs via a coupler 102, subsequently subjecting the gratings to controlled strain and temperature conditions. After this, the reflected light is received and its spectrum data stored for further analysis. Utilizing multi-parameter detection protocols, the system evaluates various sensor parameters. Specific parameters include a light source 100 wavelength range spanning from 1300 nm to 1600 nm, with FBG central peak wavelengths falling within the range of 1547.5 nm to 1552.5 nm, where the primary peak is at 1549.5 nm and the secondary at 1550.5 nm. Strain is applied at 566 microstrain, while temperature ranges between 30 to 50 degrees Celsius. The multi-parameter detection protocols involve sophisticated techniques such as transfer learning-based machine learning and linear regression analysis, utilizing 70% of the data for training and 30% for testing to ensure the accuracy of parameter evaluation. Through these methodologies, the system achieves precise assessment of sensor parameters.
In an embodiment of the present disclosure, the AI module 110 performs a sequence of steps as mentioned above for accurately evaluating the multi-parameters of the optical sensor. The optical sensor is preferably fiber bragg grating sensor that is used to measured the multi-parameter such as temperature and starin etc. The variation in the peak wavelength of FBG is depends on the change in grating period which is given as.

?_B=2×n_eff×? (1)
where, ? gratting period, n_eff effective refractive index of core mode and ?_B is the bragg wavelength. The change in peak wavelength is related with the temperature and strain which is descbed in equation-2.
????_B=?_B [(1-P_oe )e+(a+?)?T] (2)
where, ????_B=?_t-?_B (3)
P_oe is photoesatic constant, ? thermo optics , a thermal expaqnssion coefficient, e strain distribution and ?T temperature.
In an another embodiment of the present disclosure, the multiparameter detection methods for Fiber Bragg Gratings (FBGs) utilizing multi-parameters detection protocols that include matched filtering techniques and transfer learning which are fed within the AI module 110. To implement these techniques, two FBGs are cascaded in series with a small gap between them. The differing peak wavelengths of the FBGs are attributed to variations in the coefficient of thermal expansion and optoelastic coefficient.

¦(??_B1@??_B2 )=[¦(C_T1&C_?1@C_T2&C_e2 )] ¦(?T@?e) (4)
??_B1=??T*C?_T1+?e*C_?1 (5)
??_B2=??T*C?_T2+?e*C_?2 (6)
where, the C_T1,C_T2 and C_?1, C_?2 are temperature coefficient and strain coefficient. To determine the ?T and ?e , the change in peak needs to be calculated by the AI module 110 ??_B1 and ??_B2.
The Matched filtering is applied to determine the change in the peak wavelength to effect of temperature and starin. The coupled mode throy is used to simulate the and the reflectivity of FBG is dtermined from Equation-(7) and plotted in Figure 1.
R(?)=(?sinh?^2 (v(k^2-s^2 ) L) )/(?cosh?^2 (v(k^2-s^2 ) L)-s^2/k^2 ) (7)

Where, the ac coupling =2p/?×v×?dn?_eff , ?dn?_eff is the change in effective refractive index.
Furthermore, the reference signal is assumed as gsussian signal having shame shape but different central to the FBG which is mentioned in the Equation-2 and plotted in Figure 4.
I (?)=e^(-A(?(?-?_p)/(??_B ))?^2 ) (8)
where. A is constant and the value is 2.78, ?_p is the central peak of the reference signal and ??_B is the 3db bandwidth having wavelength range 1550 nm – 1552 nm with central peak ?_p is 1551 nm.
Moreover, the matched filtering is based on correlation between the FBG specral signal and reference signal. The least squire method is used to determine the relation between FBG signal and refelence signal given as.
?x(??_t,a)=?_(-8)^8¦?[R^' (?)-ßI^' (?+?_t)?]d? (9)
where, ?_t is the delay in wavelength which is determine the peak position. The maximizing the ?x(??_t,a) the first derivative is appiled with respect to ?_t and a and generates two parameter as.
C_r (?)=?_(-8)^8¦?[R(?)*I^''' (?+?_t)?]d? (10)
ß(?)=(?_(-8)^8¦?R (?) ? I^'' (?+?_t) d?)/(?_(-8)^8¦?[I^' (?)]^2 d??) (11)
where, C_r (?) is the correlation coefficient of cross correlation betwwen reflected spectrum of FBG and third derivative of reference signal and the ß(?) is the height of intensity. The zero crossing point is determined by equating C_r (?) to zero and the corresponding determined zero crossing point such as ?_t1 ?_t2 , the ß(?) will be maximum which is plotted in Figure 5(a) and 5(b) with the desired peak wavelength.
In another embodiment of the present disclosure, the efficient and more accurate detection of ?_t1 ,?_t2 and change in peak wavelength as ??_B1 and ??_B2 due to effect of temperature and strain, the transfer learning based machine learning is used. In this proposed method, the correlation coefficient ß(?_t) is splitted in sub section as ß_1 (?_1) , ß_2 (?_2) ……………… ß_n (?_n)and the variance of each subsection is determined by the AI module 110 which is described as
ß(?)=?_(i=1)^n¦?ß_i (?_i)? (12)
V_i=var(ß_i (?_i )) (13)
where, V_i variance for i=1, 2 , 3,…………….n
V_1=var(ß_1 (?_1 )) (14)
V_2=var(ß_2 (?_2 )) (15)
V_3=var(ß_3 (?_3 )) (16)

V_n=var(ß_n (?_n )) (17)
The V_1 , V_2 , V_3, ……..V_(n ) are the variance of splited intensity response ß(?) for different i value mentioned in the Equation 12.
The protocol used herein is explain with the help of tranining structure draw in Figure 1. The intensity ß(?) is splited in n-different intensity sub level and the variance for respective level is calculted. The each variaance V_1 , V_2 , V_3, ……..V_(n ) are compare to the threshold value ?(?) and select the all ß(?) those having greater variance value to the threshold value. The training process is reapting until the peak wavelengths are detected.
where,
?(?)=?_(i=1)^n¦?V_i×ß_i (?)? (18)
y(?)=?(?,V_(i-max)/?(?) ) (19)
?_t={max?(y(?),V_(i-max))} (20)
Hence , C_r (?_t )=0 (21)
??_B1= ?_t1-?_B1 (22)
??_B2= ?_t2-?_B2 (23)

In another embodiemnt of the present disclosure, to validate the peak wavelength detection, 70% data is used for training and the reamining 30% data is used for testing and validation for desing network. In Figure 6 the numbers of zero crossing point are available but the true zero crossing point is determined by the help of maximum intensity mention in the figure 6. after collecting the zero crossing point it is somehow difficult to determine the exact change in peak. In order to reduce the complexity, the transfer learning-based machine learning method is depicted from the equation 18 to equation 20.

From the table as illustrated in figure 7, it is concluded that the MSE for the training, testing and validation are very low and the regression value is high and approaching to 1. It means, the proposed method is accurately verified for the simulation data. Experimental testing of the proposed multi-parameter detection system, based on matched filtering-based machine learning powered by transfer learning, has been conducted. The experimental setup details are illustrated in figure 1. This setup comprises all necessary components for the experiment, including a broadband light source 100 with a range of 1300 nm to 1600 nm. The broadband light is directed towards port-1 of a Y-coupler 102 and exits through port-2. Subsequently, this light spectrum is incident on FBGs connected in series within the range of 1547.5 nm to 1552.5 nm, with FBG-1's central peak at 1549.5 nm and FBG-2's at 1550.5 nm. Strain up to 566 microstrain is applied to FBG1, while FBG2 is subjected to a temperature range of 30 degrees to 50 degrees.
Furthermore, the reflected spectrum of light, influenced by the strain and temperature applied to the FBGs, is collected through port-2 and redirected to port-3. These shifted peaks in the reflected spectra of FBGs are gathered using an optical spectrum analyzer (OSA). The collected data is then stored in a storage device connected to the OSA via a GPIB cable.
Moreover, the collected spectral data from FBG is plotted in normalized form if Figure 8(a) and the assumed reference signal which is used in simulation case is also plotted in Figure 8(b). The matched filtering method is applied on experimental data and the cross correlation and the zero crossing points are calculated and plotted in Figure 9(a) and the corresponding intensity is also determined which is plotted in Figure 9(b).
The transfer learning-based machine learning is applied on measured zero crossing point wavelengths data and 70% data is used for training and 30% data is used for testing. The performance validation of training. testing and validation is determined in terms of mean square error (mse) and it is found that the best validation performance is 1.7347e-8 at epoch 242 which is mentioned in Figure 10. The least MSE value tells the data set is trained successfully by the AI module 110 and the detected parameter is nearer to desired target value.
Furthermore, the AI module 110 undergoes training using training data 202 sourced from various media outlets, including external database and the pre-fed database. This module 110 may comprise a non-binary classifier, such as a multinomial logistic regression model integrated into a neural network, which is trained to forecast the probability that an input can be associated with one or more classes of a predefined set of classes. These classes correspond to either content tags 204 or user characteristics derived from user metadata 208.
The training process for AI module 110 involves supervised learning techniques applied to labeled sets of training data 202, incorporating content tags 204, content objects 206, and user metadata 208. User metadata 208 encompasses data related to the user, while content objects 206 represent data associated with tasks accomplished by the user. Content tags 204 are drawn from a feature database used to characterize content objects 206 processed by AI module 110.
Training data 202 is fed into a supervised learning subsystem 210, which includes a data input subsystem 212 responsible for receiving the training data 202. During supervised training, this subsystem 212 defines a ground truth by mapping elements of content tags 204 and user characteristics from user metadata 208. These mapped elements are then inputted into a propensity calculator 214 and an error minimization module 216. The error minimization module 216 utilizes an objective function 218, possibly an error function, which measures the disparity between the model output and the ground truth. Adjustments to the weights and coefficients of the propensity calculator 214 are made iteratively until the objective function 218 converges to a global minimum.
In certain scenarios, the input to propensity calculator 214 comprises characteristics of a user set, and the output encompasses a vector of probability values corresponding to predicted content features. This allows the propensity calculator 214 to be trained in mapping content tags 204 from training data 202 to user metadata 208, and once trained, to generate a propensity score. This score indicates the extent to which a user is inclined to release their data to at least one digital platform.
Additionally, the supervised learning subsystem 210 may include hyper parameter tuning alongside supervised learning to optimize AI module 110. This involves fine-tuning one or more terms of objective function 218 and/or AI module 110 by adjusting parameters that are not learned, such as scalar weighting factors.
The training state of proposed system is plotted as depicted in figure 11 and it is found that the parameter gradient is very low as 3.3071e-8 at 242 epochs. The gradient is described the slope of the model function with high rate of changes for assumed function. The mu is control parameter for the proposed system network which is used to optimize and effects the MSE convergence given. The validation check is use to terminate the learning process of proposed system structure and the number of validation check is proportional to the iteration used for successful trained neural network. The validation check is zero at 242 epochs, hence, the MSE value is very less nearly approaching zero.
After validation performance and validation check, the linear regression analysis method is applied to validate the peak detection which is plotted in figure 12. The R-value for training, testing and validation is maximum as 1, it means, the fitting data is approaching to the target value and the proposed multiparameter detection method is accurate.
After training and validation of regression, transfer learning method is applied to determine the peak wavelength as mentioned in the equation 5, equation 6 equation 22 and equation 23 are used to calculated the strain and temperature on primary FBG 104 and secondary FBG 108 as 566.12 per micro-strain and 49.566 ?
While considerable emphasis has been placed herein on the specific features of the preferred embodiment, it will be appreciated that many additional features can be added and that many changes can be made in the preferred embodiment without departing from the principles of the disclosure. These and other changes in the preferred embodiment of the disclosure will be apparent to those skilled in the art from the disclosure herein, whereby it is to be distinctly understood that the foregoing descriptive matter is to be interpreted merely as illustrative of the disclosure and not as a limitation.
While the invention has been described in connection with what is presently considered to be the most practical and various embodiments, it is to be understood that the invention is not to be limited to the disclosed embodiments, but on the contrary, is intended to cover various modifications and equivalent arrangements.
The embodiments described above are intended only to illustrate and teach one or more ways of practicing or implementing the present invention, not to restrict its breadth or scope. The actual scope of the invention, which embraces all ways of practicing or implementing the teachings of the invention, is defined only by the following claims and their equivalents.
, Claims:CLAIMS:
1. A matched filtering based system and method for determining multi-parameters of an optical sensor, wherein the system comprises:
a light source 100 to emit light of desired wavelength towards a first port of a Y-coupler 102 paired with the source, wherein the coupler 102 directs the light emitted by the source towards a second port of the coupler 102;
a primary 104 and secondary 108 fiber bragg grating connected with the coupler 102 in series combination, wherein the light emitted from the second port incident over the primary 104 as well as secondary 108 gratings and gets reflected from the primary 104 and secondary 108 gratings towards a third port of the coupler 102;
an optical spectrum analyser coupled with the coupler 102 to receive the data related to the light reflected from the primary 104 and secondary 108 fiber bragg grating and store the data in a memory unit paired with the optical spectrum analyser; and
a matched filtering based AI module 110 coupled with the optical spectrum analyser to analyse the received data to evaluate zero crossing points by using matched filtering and cross correlation, wherein upon evaluating zero crossing points, multi-parameter detection protocols are applied over received data to evaluate the multi-parameters of the sensor.
.
2. The matched filtering based system determining multi-parameters of an optical sensor as claimed in claim 1, wherein the multi-parameter includes but not limited to strain and temperature.
3. The matched filtering based system determining multi-parameters of an optical sensor as claimed in claim 1, wherein the multi-parameters detection protocols includes but not limited to transfer learning based machine learning and linear regression analysis.
4. The matched filtering based system determining multi-parameters of an optical sensor as claimed in claim 1, a computing unit 112 is paired with the AI module 110 to illustrate the data evaluated by the AI module 110.
5. A method for determining multi-parameters of an optical sensor, wherein the method comprising the steps of:
emitting light towards a primary 104 and secondary 108 fiber bragg grating via a coupler 102 attached with the gratings in series;
applying a predefined amount of strain and temperature over the primary 104 and secondary 108 fiber bragg grating respectively;
receiving the light reflected from the primary 104 and secondary 108 gratings via the coupler 102;
storing the data related to reflected spectrum of the light; and
evaluating multi-parameters of the sensor by utilizing multi-parameter detection protocols.
6. The method as claimed in claim 5, wherein the light source 100 having range of 1300 nm to 1600 nm.
7. The method as claimed in claim 5, wherein the fiber bragg gratings (FBGs) connected as in series with the range of 1547.5 nm to 1552.5 nm with the central peak as 1549.5 nm for primary FBG 104 and 1550.5 nm for secondary FBG 108.
8. The method as claimed in claim 5, wherein the predefined amount strain is applied on primary FBG 104 refers to 566 micro strain and temperature applied on secondary FBG 108 refers to 30 degree to 50 degree.
9. The method as claimed in claim 5, wherein the multi-parameter detection protocols include a transfer learning-based machine learning is applied on measured zero crossing point wavelengths data and 70% data is used for training and 30% data is used for testing.
10. The method as claimed in claim 1, wherein the multi-parameter detection protocols also include linear regression analysis method applied to validate the peak detection results in accurately evaluation of the multi-parameter detection.