A SYSTEM AND A METHOD FOR PREDICTING THE PERFORMANCE OF RF FILTER IN REAL-TIME

Abstract

ABSTRACT A SYSTEM AND A METHOD FOR PREDICTING THE PERFORMANCE OF RF FILTER IN REAL-TIME The present system discloses a system (100) that is designed to predict the performance of RF filters (102b) by using an Artificial Neural Network (ANN) module (108) and Full-Wave Electromagnetic (EM) simulation (104). The system (100) consists of several key components: an EM simulation module (104) that generates scattering parameters (S11 and S21), an optimization module (106) that collects, standardizes, and normalizes the data, and an ANN module (108) that builds, trains, and validates a predictive model. The ANN model (108a) predicts scattering parameters based on the design specifications of the RF filter (102b). The system (100) also includes an evaluation module (108e) to assess the accuracy of predictions and an output module (110) that displays the predicted results in various formats, such as graphs and numerical data. The integration of these modules ensures the RF filter's (102b) performance can be accurately predicted and optimized.

Specification

Description:FIELD
The present disclosure generally relates to the field of radio frequency (RF) filters to predict the scattering parameter without any specific tools. More particularly, the present disclosure relates to a system and a method for predicting the performance of RF filters in real-time.
DEFINITIONS
As used in the present disclosure, the following terms are generally intended to have the meaning as set forth below, except to the extent that the context in which they are used indicate otherwise.
Computer Simulation Technology (CST) microwave studio: The CST microwave studio is a high-frequency electromagnetic (EM) simulation software that is used to design, analyze, and optimize various microwave and radio frequency (RF) components and systems. The CST provides a comprehensive set of tools for simulating electromagnetic fields, circuits, and systems, making it an invaluable tool for engineers and researchers in the field of microwave and RF engineering.
High frequency structure simulator (HFSS): The high frequency structure simulator (HFSS) is software used for electromagnetic (EM) simulation of high-frequency components and systems. HFSS is primarily used in the design and analysis of antennas, microwave circuits, and radio frequency (RF) devices.
BACKGROUND
The background information herein below relates to the present disclosure but is not necessarily prior art.
Traditionally, an RF filter was used in electronic devices to block electromagnetic frequencies in a particular range. Conventionally, the design and prediction of the performance of the RF filter was time time-consuming and complex processes. The performance of the bandstop filter was measured by electromagnetic simulators, such as a computer simulation technology (CST) and a high frequency structure simulator (HFSS). The CST and the HFSS are advanced computational tools used to predict the electromagnetic performance of bandstop RF filters.
The limitation of the conventional method for predicting the performance of the RF bandstop filters was that simulator, such as the CST and the HFSS, were required to predict the filter's performance. These simulators run for each iteration of data, increasing the time and computational effort required. Conventionally, the performance prediction does not allow for real-time performance predictions of the RF bandstop filter.
Therefore, there is felt a need for a system and a method for predicting the performance of RF filters in real-time that alleviates the aforementioned drawbacks.
OBJECTS
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 provide a system for predicting the performance of RF filters in real time.
Another object of the present disclosure is to provide a system that increases computational time.
Still another object of the present disclosure is to provide a system that consumes less time.
Yet another object of the present disclosure is to provide a system that does not require specific simulation tools.
Still another object of the present disclosure is to provide a system that predicts the scattering parameters of RF filters.
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.
SUMMARY
The present disclosure is a system and a method for predicting the performance of RF filter in real-time. The system comprises a substrate, a Full-Wave Electromagnetic module, an optimization module, an artificial neural network module, and an output module.
The substrate includes at least one resonator and an RF filter.
The substrate is configured to simulate and measure a signal that is transmitted from at least one resonator. The substrate provides structural support for the RF filter's components.
In an embodiment, the substrate is a FR4 substrate.
The resonator is connected to the RF filter. The resonator is configured to control frequencies that will be blocked or allowed to pass from the RF filter.
The RF filter is connected to a Full-Wave Electromagnetic Simulation module. The RF filter is configured to either pass or block the signals based on the behavior of the resonator.
The Full-Wave Electromagnetic (EM) simulation module is configured to analyze the scattering parameters including S11 (reflection coefficient) and S21 (transmission coefficient) that are reflected and transmitted at desired frequencies from the RF filter.
The optimization module is connected to the Full-Wave Electromagnetic (EM) simulation module. The optimization module receives scattering parameters (S11 and S21).
The optimization module includes a data acquisition module and a pre-processing module.
The data acquisition module is configured to prepare datasets for scattering parameters (S11 and S21).
The pre-processing module is configured to standardize and normalize collected datasets and generate pre-processed data. The pre-processed data is optimized and cleaned for accurate model training and prediction.
The Artificial Neural Network (ANN) module is configured to build an Artificial Neural Network (ANN) model to train scattering parameters (S11 and S21) corresponding to design specifications, adjusting internal parameters.
The ANN module includes a training module, a validation module, a prediction module, and an evaluation module.
The training module is configured to iteratively train the artificial neural network (ANN) model based on the pre-processed data, to minimize the error between predicted and actual outputs.
The validation module is configured to iteratively validate the artificial neural network (ANN) model based on the trained data. The ANN model is capable of producing accurate and reliable results in different conditions.
The prediction module is configured to utilize the ANN model to predict the scattering parameters (S11 and S21) of the RF filter based on the design specification.
The evaluation module is configured to evaluate the ANN model's performance based on accuracy and precision.
The output module is configured to display and/or output predicted scattering parameters of the RF filter in various formats, including graphical patterns, numerical data, performance metrics, and design specifications of the substrate.
In an aspect, the design specifications include a number of resonators, material properties, and operational frequency ranges.
In an aspect, The ANN model is trained on a dataset of RF filters paired with corresponding simulated or measured scattering parameters (S11 and S21). The ANN model predicts the performance of scattering parameters including S11 (reflection coefficient) and S21 (transmission coefficient).
In an aspect, the training module uses Levenberg-Marquardt (LM) algorithm by considering the validation and testing dataset at 15% and the corresponding training dataset at 70% accordingly.
In an aspect, the pre-processing module is configured to standardize and normalize the collected datasets and then generate the processed data. The preprocessed data is optimized and cleaned for accurate model training prediction.
In an aspect, the pre-processing module separates the collected data into input datasets containing design specifications, output datasets containing scattering parameters, and a training dataset comprising both design specifications and scattering parameters (S11 and S21).
The ANN model is configured with an input layer to receive the design specifications and scattering parameters (S11 and S21). The hidden layers comprise a plurality of RF filters with activation function, and an output layer configured to generate the performance of scattering parameters (S11 and S21).
The output module is configured to provide a graphical interface for real-time visualization of the predicted radiation pattern and design specifications, enabling user-driven adjustments to the antenna design process.
The present disclosure further envisages a method for predicting the performance of RF filter. The method comprises the following:
• generating, by a Full-Wave Electromagnetic (EM) simulation module, scattering parameters including S11 (reflection coefficient) and S21 (transmission coefficient) that is reflected and transmitted at desired frequencies from an RF filter;
• connecting the Full-Wave Electromagnetic (EM) simulation module with optimization module iteratively to receive scattering parameters so as to generate input data;
• processing, by a preprocessing module, input data to standardize and normalize collected input data to generate pre-processed data is optimized and cleaned for accurate model training and prediction;
• implementing and training, by a neural network model, an Artificial Neural Network (ANN) model on the pre-processed data to learn the relationship between the input design specifications and the scattering parameters of the RF filter and predicting the characteristics of the RF filter based on the design specification; and
• displaying, by an output module, predicted characteristics of the RF filter in various formats, including graphical patterns, numerical data, performance metrics, and design specifications of the substrate.
BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWING
A system for predicting the performance of the RF filter in real time of the present disclosure will now be described with the help of the accompanying drawing, in which:
Figure 1 illustrates the block diagram of the system to predict the performance of the RF filter in accordance with an embodiment of the present disclosure;
Figure 2A and Figure 2B illustrate a flow chart depicting the steps involved in a method for predicting the scattering parameters in accordance with an embodiment of the present disclosure;
Figure 3 illustrates the RF filter design in a substrate in accordance with an embodiment of the present disclosure;
Figure 4 illustrates the flow chart of the artificial neural network (ANN) model in accordance with an embodiment of the present disclosure;
Figure 5 illustrates the architecture of the artificial neural network (ANN) module in accordance with an embodiment of the present disclosure;
Figure 6A and Figure 6B illustrate the graphical representation of the training, testing and output dataset in accordance with an embodiment of the present disclosure; and
Figure 7 illustrates the simulation of scattering parameters in accordance with an embodiment of the present disclosure.
LIST OF REFERENCE NUMERALS
100 - System
102 - Substrate
102a - Resonator
102b - Radio Frequency (RF) Filter
104 - Electromagnetic Simulation Module
106 - Optimization Module
106a - Data Acquisition Module
106b - Pre-Processing Module
108 - Artificial Neural Network (ANN) Module
108a - Artificial Neural Network (ANN) Model
108b - Training Module
108c - Validation Module
108d - Prediction Module
108e - Evaluation Module
110 - Output Module
112 - First Port
114 - Second Port
116 - Input Layer
118 - Hidden Layer
120 - Output Layer
122 - Predicted Output
DETAILED DESCRIPTION
Embodiments, of the present disclosure, will now be described with reference to the accompanying drawing.
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.
When an element is referred to as being "mounted on," "engaged to," "connected to," or "coupled to" another element, it may be directly on, engaged, connected or coupled to the other element. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed elements.
The terms first, second, third, etc., should not be construed to limit the scope of the present disclosure as the aforementioned terms may be only used to distinguish one element, component, region, layer, or section from another component, region, layer, or section. Terms such as first, second, third, etc., when used herein do not imply a specific sequence or order unless clearly suggested by the present disclosure.
Terms such as "inner," "outer," "beneath," "below," "lower," "above," "upper," and the like, may be used in the present disclosure to describe relationships between different elements as depicted from the figures.
Traditionally, an RF filter was used in electronic devices to block electromagnetic frequencies in a particular range. Conventionally, the design and prediction of the performance of the RF filter were time time-consuming and complex processes. The performance of the bandstop filter was measured by electromagnetic simulators, such as a CST and an HFSS. The CST and the HFSS are advanced computational tools used to predict the electromagnetic performance of bandstop RF filters.
The limitation of the conventional method for predicting the performance of the RF bandstop filters was that simulator, such as the CST and the HFSS, were required to predict the filter's performance. These simulators run for each iteration of data, increasing the time and computational effort required. Conventionally, the performance prediction does not allow for real time performance predictions of the RF bandstop filter.
To address the issues of the existing system and methods, the present disclosure envisages a system for predicting the performance of RF filter in real-time (hereinafter referred to as "system 100") and a method for predicting the performance of RF filter in real-time (hereinafter referred to as "method 200"). The system 100 will now be described with reference to Figure 1 and the method 200 will be described with reference to Figure 2A and Figure 2B.
Referring to Figure 1, the system (100) comprises a substrate (102), a full-wave electromagnetic simulation (104), an optimization module (106), an artificial neural network (ANN) module (108), and an output module (110).
The substrate (102) includes at least one resonator (102a) and an RF filter (102b).
The substrate (102) is configured to simulate and measure a signal that is transmitted from at least one resonator (102a). The substrate (102) provides structural support for the RF filter's (102b) components.
In an embodiment, the substrate (102) is a FR4 substrate (102).
The resonator (102a) is connected to the RF filter (102b). The resonator (102a) is configured to control frequencies that will be blocked or allowed to pass from the RF filter (102b).
The RF filter (102b) is connected to the Full-Wave Electromagnetic Simulation module (104). The RF filter (102b) is configured to either pass or block the signals based on the behavior of the resonator (102a).
The Full-Wave Electromagnetic (EM) simulation module (104) is configured to analyze the scattering parameters including S11 (reflection coefficient) and S21 (transmission coefficient) that are reflected and transmitted at desired frequencies from the RF filter (102b).
The optimization module (106) is connected to the Full-Wave Electromagnetic (EM) simulation module (104). The optimization module (106) receives scattering parameters (S11 and S21).
The optimization module (106) includes a data acquisition module (106a) and a pre-processing module (106b).
The data acquisition module (106a) is configured to prepare datasets for scattering parameters (S11 and S21).
The pre-processing module (106b) is configured to standardize and normalize collected datasets and generate pre-processed data. The pre-processed data is optimized and cleaned for accurate model training and prediction.
In an aspect, the pre-processing module (106b) separates the collected data into input datasets containing design specifications, output datasets containing scattering parameters, and a training dataset comprising both design specifications and scattering parameters (S11 and S21).
In an aspect, the pre-processing module (106b) is configured to standardize and normalize the collected datasets and then generate the processed data. The preprocessed data is optimized and cleaned for accurate model training prediction.
The Artificial Neural Network (ANN) module (108) is configured to build an Artificial Neural Network (ANN) model (108a) to train scattering parameters (S11 and S21) corresponding to design specifications, adjusting internal parameters.
In an aspect, The ANN model (108a) is trained on a dataset of RF filter (102b) paired with corresponding simulated or measured scattering parameters (S11 and S21). The ANN model (108a) predicts the performance of scattering parameters including S11 (reflection coefficient) and S21 (transmission coefficient).
The ANN module (108) includes a training module (108b), a validation module (108c), a prediction module (108d), and an evaluation module (108e).
The training module (108b) is configured to iteratively train the artificial neural network (ANN) model (108a) based on the pre-processed data, to minimize the error between predicted and actual outputs.
In an aspect, the training module (108b) uses Levenberg-Marquardt (LM) algorithm by considering the validation and testing dataset at 15% and the corresponding training dataset at 70% accordingly.
The validation module (108c) is configured to iteratively validate the artificial neural network (ANN) model (108a) based on the trained data. The ANN model (108a) is capable of producing accurate and reliable results in different conditions.
The prediction module (108d) is configured to utilize the ANN model (108a) to predict the scattering parameters (S11 and S21) of the RF filter (102b) based on the design specification.
The evaluation module (108e) is configured to evaluate the ANN model's (108a) performance based on accuracy and precision.
The output module (110) is configured to display and/or output predicted scattering parameters of the RF filter (102b) in various formats, including graphical patterns, numerical data, performance metrics, and design specifications of the substrate (102).
The ANN model (108a) is configured with an input layer to receive the design specifications and scattering parameters (S11 and S21). The hidden layers comprise a plurality of RF filter (102b) with an activation function, and an output layer configured to generate the performance of scattering parameters (S11 and S21).
The output module (110) is configured to provide a graphical interface for real-time visualization of the predicted radiation pattern and design specifications, enabling user-driven adjustments to the antenna design process.
Figure 2A and Figure 2B illustrate a flow chart depicting the steps involved in a method for predicting the scattering parameters in accordance with an embodiment of the present disclosure. The order in which method (200) is described is not intended to be construed as a limitation, and any number of the described method steps may be combined in any order to implement method (200), or an alternative method. Furthermore, method (200) may be implemented by processing resource or computing system(s) through any suitable hardware, non-transitory machine-readable medium/instructions, or a combination thereof. The method (200) comprises the following steps.
At step 202, the method (200) includes generating, by a Full-Wave Electromagnetic (EM) simulation module (104), scattering parameters including S11 (reflection coefficient) and S21 (transmission coefficient) that is reflected and transmitted at desired frequencies from a RF filter (102b).
At step 204, the method (200) includes connecting the Full-Wave Electromagnetic simulation module (104) with the optimization module (106) iteratively to receive the scattering parameters to generate input data.
At step 206, the method (200) includes processing, by a pre-processing module (106b), the input data to standardize and normalize the collected input data to generate pre-processed data is optimized and cleaned for accurate model training and prediction.
At step 208, the method (200) includes implementing and training, by an artificial neural network module (108), an Artificial Neural Network (ANN) model (108a) on the pre-processed data to learn the relationship between the input design specifications and the scattering parameters of the RF filter (102b), predicting the characteristics of the RF filter (102b) based on the design specification.
At step 210, the method (200) includes displaying, by an output module (110) the predicted characteristics of the RF filter (102b) in various formats, including graphical patterns, numerical data, performance metrics, and design specifications of the substrate (102).
Figure 3 illustrates the RF filter design in a substrate in accordance with an embodiment of the present disclosure. The substrate (102) provides the base to at least one resonator (102a) and RF filter (102b) to analyze the electromagnetic properties of the RF filter (102b). The RF filter (102b) has dimensions including L2, W2, W3, and S with design specifications of a length of 175mm x 150mm.
The substrate (102) has relative permittivity and loss tangent, which affect the scattering parameters including S11 (reflection coefficient) and S21 (transmission coefficient). The relative permittivity is 4.3 and loss tangent is 0.02.
The substrate (102) has two horizontal conductive strips with a separation distance "S". The distance 'S' is used as parameter for controlling the frequencies of the RF filter (102b). The distance "S" determines the specific frequencies that are blocked or passed to the RF filter (102b). The distance "S" impacts the electromagnetic coupling between the strips, which influence the S11 (reflection coefficient) and the S21 (transmission coefficient). The strips are positioned relative to the substrate (102). The substrate (102) is used for supporting the conductive strips and insulating them from each other.
The ports (first port 112 and second port 114) are the points where the signal enters and exit from the RF filter (102b). The signal travel through the conductive strips, and the scattering parameters (S11 and S21) are measured at these ports using Full-wave electromagnetic simulation (104).
Figure 4 illustrates flow chart of the artificial neural network (ANN) model in accordance with an embodiment of the present disclosure. The flow chart begins with the design of RF filter (102b). The RF filter's (102b) data is collected and undergoes pre-processing (106b). The data pre-processing module (106b) standardizes and normalize the data to make it for suitable analysis. The pre-processing module (106b) separates normalized data into three datasets: input dataset, testing dataset and output dataset.
The input dataset is fed into the system (100) to define the artificial neural network (ANN) module (108). The ANN model (108a) then undergoes into a training module (108b). If the model does not train (no), it loops back to the input dataset for further refinement. If the model is trained successfully (yes), then it goes to the validation module (108c) to validate the trained data.
The validation module (108c) ensures that the ANN model (108a) is capable of producing accurate and reliable results in different conditions. Then, the validated data goes to the predicted output module (110), which predicts the scattering parameter of the RF filter (102b).
The output predicted by the output module (110) goes to the visual evaluation model (108e). The visual evaluation model (108e) is configured to evaluate the ANN model's (108a) performance-based accuracy and precision.
Figure 5 illustrates the architecture of the ANN model in accordance with an embodiment of the present disclosure. The ANN model (108a) includes an input layer (116), a plurality of hidden layer (118), and an output layer (120). The ANN model (108a) receives the input from the substrate (102). The input layer (116) consists of several nodes. Each node in the input layer (116) represents a different input feature, such as the dimension of the RF filter (102b) like L2, W2, W3.
The data from the input layer (116) is passed to the plurality hidden Layer (118). The plurality of hidden layer (118) contains a larger number of nodes compared to the input layer (116). Each node in the hidden layer (118) applies a set of weight to the input data and then passes the result through an activation function.
The output layer (120) receives the processed data from the hidden layer (118). The output layer (120) has fewer nodes than hidden layer (118). The nodes in the output layer (120) generate the final predictions of the network. The output nodes predict the scattering parameters of the RF filter (102b), specifically the S11 and S21 parameters. These predictions are based on the input dimensions and the learned patterns from the hidden layer (118).
Figure 6A and Figure 6B illustrate the graphical representation of training, testing and output dataset in accordance with an embodiment of the present disclosure. The regression coefficient plots are used to depict the relationship between the predicted values and the actual values of the scattering parameters (S11 and S21). The performance plots display the mean squared error and the coefficient of determination (R-squared) for the predicted outcomes.
In terms of specific measurements, the regression coefficient (R) is reported as 0.999, indicating a strong correlation between the predicted and actual values. The mean squared error is recorded at 0.001, reflecting a low level of error in the predictions. Additionally, the coefficient of determination (R-squared) is also noted as 0.999, further confirming the high accuracy of the model in predicting the performance parameters.
Figure 7 illustrates the simulation of scattering parameters in accordance with an embodiment of the present disclosure. Figure 7 presents a comparison between the predicted and simulated values of the scattering parameters (S11 and S21) for the RF filter (102b). The predicted values are generated using the developed Artificial Neural Network (ANN) model (108a), while the simulated values were obtained through the CST-Full microwave suite. This comparison demonstrates the accuracy of the ANN model (108a) in predicting the performance of the RF filter (102b) by closely matching the simulated results.
In an embodiment, the need for traditional simulations tool is eliminated. The system (100) enables real-time prediction of the RF filter's (102b) performance, based on its dimensions.
In an embodiment, significant benefits of the system (100) are realized in terms of computational requirements, speed, and performance. The system (100) reduces the computational load, resulting in faster processing times and improved overall performance in predicting the RF filter's (102b) parameters.
In an embodiment, the system (100) is used to predict the scattering parameters (S11 and S21) of the RF filter design. By leveraging the trained artificial neural network, the system (100) provides accurate performance predictions based on the RF filter's (102b) dimensions, ensuring reliable and efficient design outcomes without the need for traditional simulation processes.
The present disclosure is further described in light of the following experiments which are set forth for illustration purpose only and not to be construed for limiting the scope of the disclosure. The following experiments can be scaled up to industrial/commercial scale and the results obtained can be extrapolated to industrial scale.
In an exemplary embodiment, the numerical analysis was conducted for the RF filter, and experimental data was collected. A total of 625 (N) iterations were carried out by varying dimensions such as the width of the mini strip (W2), the length and width of the wing strip (L2, W3), and S. These variations were recorded and tabulated in Table 1. Corresponding data for the S11 and S21 parameters, totaling N x 1001 data points, were gathered. The collected data was then divided into three categories: input, output, and testing datasets. The input dataset included dimensions like L2, S, W2, W3, and frequency, the performance to be predicted served as the output dataset. The testing dataset contained 15% of the input data.
Parameter variable Initial value Sweep range
L2 134.5 134.3:0.1:134.7
S 5.5 5.55:0.1:5.95
W3 2 1.7:0.1:2.1
W2 3.5 3.3:0.1:3.7
In an operative configuration, the system (100) comprises of a substrate (102) connected with the Full-Wave Electromagnetic Simulation (104). The substrate (102) is configured to simulate and measure a signal that is transmitted at least one resonator (102a). The substrate (102) provides structural support for RF filter (102b). The resonator (102a) is connected to the RF filter (102b), used to control frequencies that will be blocked or allowed to pass from the RF filter (102b). The RF filter (102b) is connected to the Full-Wave Electromagnetic simulation module (104). The Rf filter (102b) is configured to either pass or block the signal based on the behavior of the resonator (102a). The Full-Wave Electromagnetic module (104) is configured to analyze the scattering parameters (S11 and S21). The optimization module (106) is connected with the Full-wave Electromagnetic Simulation (104). The optimization module (106) receives scattering parameters (S11 and S21). The optimization module (106) includes data acquisition module (106a) and pre-processing module (106b). The data acquisition module (106a) is used to prepare datasets for scattering parameters (S11 and S21). The pre-processing (106b) is used to optimized and clean the data for accurate model training and prediction. The artificial neural network module (108) is configured to build an artificial neural network model (108a) to train scattering parameters (S11 and S21) corresponding to design specification and adjusting internal parameters. The ANN module (108) includes a training module (108b), a validation module (108c), a prediction module (108d) and an evaluation module (108e). The training module (108b) is based on the pre-processed data, to minimize the error between predicted and actual output. The validation module (108c) is based on the trained data. The ANN model (108a) is capable to produce accurate and reliable results in different conditions the predicted module is used to predict the scattering parameters (S11 and S21) of the RF filter (102b). The output module (110) is configured to display and/or output predicted scattering parameters (S11 and S21).
Advantageously, the system (100) predicts the performance of the RF filter (102b) in real time based its dimensions, eliminating the need for simulation tools. The system (100) decreases computational requirements and increase speed, and overall performance. Additionally, the system (100) is capable of learning from new data, enabling it to refine its predictions and evaluate multiple iterations effectively. The system (100) accurately predicts the scattering parameters (S11 and S21) for standard RF filter (102b), enhancing the design process and decision-making efficiency.
The foregoing description of the embodiments has been provided for purposes of illustration and not intended to limit the scope of the present disclosure. Individual components of a particular embodiment are generally not limited to that particular embodiment but are interchangeable. Such variations are not to be regarded as a departure from the present disclosure, and all such modifications are considered to be within the scope of the present disclosure.

TECHNICAL ADVANCEMENTS
The present disclosure described herein above has several technical advantages including, but not limited to, a system and a method for predicting the performance of RF filter in real-time that:
• eliminates the need for simulations;
• increases computational efficiency, speed, and overall performance;
• consume less time; and
• eliminates the need for specific tools.
• predicts the S11 and S21 parameters of bandstop RF filter.
The embodiments herein and the various features and advantageous details thereof are explained with reference to the non-limiting embodiments in the following description. Descriptions of well-known components and processing techniques are omitted so as to not unnecessarily obscure the embodiments herein. The examples used herein are intended merely to facilitate an understanding of ways in which the embodiments herein may be practiced and to further enable those of skill in the art to practice the embodiments herein. Accordingly, the examples should not be construed as limiting the scope of the embodiments herein.
The foregoing description of the specific embodiments so fully reveal the general nature of the embodiments herein that others can, by applying current knowledge, readily modify and/or adapt for various applications such specific embodiments without departing from the generic concept, and, therefore, such adaptations and modifications should and are intended to be comprehended within the meaning and range of equivalents of the disclosed embodiments. It is to be understood that the phraseology or terminology employed herein is for the purpose of description and not of limitation. Therefore, while the embodiments herein have been described in terms of preferred embodiments, those skilled in the art will recognize that the embodiments herein can be practiced with modification within the spirit and scope of the embodiments as described herein.
The use of the expression "at least" or "at least one" suggests the use of one or more elements or ingredients or quantities, as the use may be in the embodiment of the disclosure to achieve one or more of the desired objects or results.
Any discussion of documents, acts, materials, devices, articles or the like that has been included in this specification is solely for the purpose of providing a context for the disclosure. It is not to be taken as an admission that any or all of these matters form a part of the prior art base or were common general knowledge in the field relevant to the disclosure as it existed anywhere before the priority date of this application.
The numerical values mentioned for the various physical parameters, dimensions or quantities are only approximations and it is envisaged that the values higher/lower than the numerical values assigned to the parameters, dimensions or quantities fall within the scope of the disclosure, unless there is a statement in the specification specific to the contrary.
While considerable emphasis has been placed herein on the components and component parts of the preferred embodiments, it will be appreciated that many embodiments can be made and that many changes can be made in the preferred embodiments without departing from the principles of the disclosure. These and other changes in the preferred embodiment as well as other embodiments 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. , Claims:WE CLAIM:
1. A system (100) for predicting the performance of RF filter (102b), said system (100) comprises:
• a Full-Wave Electromagnetic (EM) simulation module (104) configured to generate scattering parameters including S11 (reflection coefficient) and S21 (transmission coefficient) that is reflected and transmitted at desired frequencies from a RF filter (102b); wherein said RF filter (102b) includes:
o a substrate (102) configured to simulate and measure a signal that is transmitted from at least one resonator (102a);
o at least one resonator (102a) configured to analyze scattering parameters to block or pass at desired frequencies;
• an optimization module (106) configured to connect with said Full-Wave Electromagnetic (EM) simulation module (104) iteratively to receive said scattering parameters, said optimization module (106) comprises:
o a data acquisition module (106a) configured to prepare datasets for said scattered parameters;
o a pre-processing module (106b) configured to standardize and normalize said collected datasets and generate pre-processed data, ensuring the pre-processed data is optimized and cleaned for accurate model training and prediction;
• an Artificial Neural Network (ANN) module (108) configured to build an Artificial Neural Network (ANN) model (108a) to train said scattering parameters corresponding to design specification, adjusting internal parameters, wherein said ANN module (108) comprises:
o a training module (108b) configured to iteratively train said ANN model based on the pre-processed data, minimizing the error between predicted and actual outputs;
o a validation module (108c) configured to iteratively validate said ANN model based on the trained data, ensuring that ANN model capable of producing accurate and reliable results in different conditions;
o a prediction module (108d) configured to utilize said ANN model to predict predicting the scattering parameters of the RF filter (102b) based on the design specification;
o an evaluation module (108e) configured to evaluate the ANN model's performance based on its accuracy and precision; and
• an output module (110) configured to display and/or output said predicted scattering parameters of the RF filter (102b) in various formats, including graphical patterns, numerical data, and performance metrics, and design specifications of said substrate (102).
2. The system (100) as claimed in claim 1, wherein said design specifications includes a number of resonators (102a), material properties, and operational frequency range.
3. The system (100) as claimed in claim 1, wherein said RF filter (102b) having dimensions including L2, W2, W3 and S with design specifications of length of 175mm x 150mm.
4. The system (100) as claimed in claim 1, wherein said ANN model is trained on a dataset of RF filter (102b) paired with corresponding simulated or measured signal parameters, allowing the model to predict the performance of scattering parameters including S11 (reflection coefficient) and S21 (transmission coefficient).
5. The system (100) as claimed in claim 1, wherein said training module (108b) uses Levenberg-Marquardt (LM) algorithm by considering the validation and testing dataset at 15% and the corresponding training dataset at 70% accordingly.
6. The system (100) as claimed in claim 1, wherein said pre-processing module (106b) further comprises separating the collected data into input datasets containing design specifications, output datasets containing scattering parameters, and a training dataset comprising both design specifications and scattering parameters (S11 and S21).
7. The system (100) as claimed in claim 1, wherein said ANN model is configured with:
• an input layer (116) to receive the design specifications and scattering parameters (S11 and S21),
• one or more hidden layers (118) comprising a plurality of RF filters (102b) with activation function, and
• an output layer (120) configured to generate the performance of scattering parameters (S11 and S21).
8. The system (100) as claimed in claim 1, wherein said output module (110) is configured to provide a graphical interface for real-time visualization of the predicted radiation pattern and design specifications, enabling user-driven adjustments to the antenna design process.
9. A method for predicting the performance of RF filter (102b), said method (200) comprises the following:
• generating, by a Full-Wave Electromagnetic (EM) simulation module (104), scattering parameters including S11 (reflection coefficient) and S21 (transmission coefficient) that is reflected and transmitted at desired frequencies from a RF filter (102b);
• connecting said Full-Wave Electromagnetic (EM) simulation module (104) with optimization module (106) iteratively to receive said scattering parameters so as to generate input data;
• processing, by a pre-processing module (106b), said input data to standardize and normalize said collected input data to generate pre-processed data is optimized and cleaned for accurate model training and prediction;
• implementing and training, by artificial neural network module (108), an Artificial Neural Network (ANN) model (108a) on the pre-processed data to learn the relationship between the input design specifications and the scattering parameters of the RF filter (102b) and predicting the characteristics of the RF filter (102b) based on the design specification; and
• displaying, by an output module (110), said predicted characteristics of the RF filter (102b) in various formats, including graphical patterns, numerical data, and performance metrics, and design specifications of said substrate (102).

Dated this 29th day of October, 2024

_______________________________
MOHAN RAJKUMAR DEWAN, IN/PA - 25
OF R. K. DEWAN & CO.
AUTHORIZED AGENT OF APPLICANT

TO,
THE CONTROLLER OF PATENTS
THE PATENT OFFICE, CHENNAI

Documents