System and method for generating a design for a reflect array

Abstract

A method (800) for generating a design for a reflect array (106) is disclosed. The method (800) includes receiving (802) a set of design input parameters The method (800) includes determining (804) a phase shift profile matrix of the reflect array (106) based on the set of design input parameters. The method (800) further includes predicting (806) an optimal physical dimension of each unit cell (110) for a given geometric shape based on the determined phase shift profile matrix by applying a trained machine learning model. The method (800) further includes generating (804) a Printed Circuit Board (PCB) file based on the determined optimal physical dimension. The method (800) substantially reduces design time by utilizing the trained machine learning model to predict optimal dimensions without repetitive electromagnetic simulations, thereby enabling rapid reflect array design.

Description

FIELD OF THE INVENTION

[0001] The present disclosure generally relates to antenna design. Moreparticularly, the present disclosure relates to a system and a method forgenerating a design for a reflect array.BACKGROUND

[0002] Aperture antenna technology has evolved through two fundamentalapproaches. The first approach involves array antennas that perform control ofamplitude and phase excitation of each array element to obtain specific radiationcharacteristics. The second approach involves geometrical optics-basedantennas, such as reflector antennas, that perform control of radiationcharacteristics through prescribed surface geometrical shapes.

[0003] Reflector antennas provide structural simplicity but lack precise controlof radiation characteristics. In contrast, array antennas provide enhanced controlbut require complex feeding networks and increased design complexity. As aresult, both approaches face limitations in simultaneously achieving structuralsimplicity and precise radiation control.

[0004] To overcome these conflicting characteristics, researchers developed ahybrid configuration known as a reflect array antenna. A reflect array antennaperforms the function of a reflective surface composed of multiple antennaelements with specific impedance characteristics. Each antenna elementperforms control of the phase of reflected waves across the aperture.Consequently, a reflect array antenna performs the transformation of an incidentwavefront into a desired radiation pattern through a flat structure.

[0005] The reflect array antenna typically provides high directivity and a lowprofile compared to a parabolic reflector antenna. Furthermore, the planarconfiguration of a reflect array antenna facilitates easier integration with moderncommunication platforms. However, the reflect array antenna faces challengesin design, modeling, and fabrication because a large number of unit cells requireindividual phase response control.

[0006] A design procedure of a reflect array antenna typically involvesmultiple computational stages. These stages include electromagnetic analysis,phase response computation, and layout generation. A conventional techniqueperforms these stages through manual iteration and adjustment. Consequently,the design procedure requires significant time and human intervention. Thisleads to inefficiency and limits scalability when handling large apertures orcomplex frequency requirements.

[0007] Furthermore, a conventional technique for reflect array antenna designdepends on extensive full-wave electromagnetic simulations. Such simulationsrequire high computational resources and prolonged processing time. As a result,these techniques are unsuitable for rapid prototyping or iterative optimization.Moreover, each unit cell requires precise geometrical parameter adjustment toachieve the desired reflection phase, further extending the overall design cycle.

[0008] Additionally, the design process of the reflect array antenna facesdifficulty in achieving broadband performance. Each unit cell typically providesan optimized response only within a limited frequency range. Consequently,maintaining consistent phase performance across wide frequency bands becomesdifficult. Furthermore, environmental variations and fabrication tolerances canalter the reflection phase response, which degrades the overall antennaperformance.

[0009] Recent advances in wireless communication demand precise beamcontrol, high gain, and compact configuration. The conventional technique doesnot adequately address these requirements because it involves sequential manualoperations and lengthy validation processes. Each design iteration requirescomputation, evaluation, and verification of numerous parameters, whichrestricts automation. Consequently, achieving fast and reliable reflect arrayantenna design remains a persistent challenge.

[0010] Additionally, the conventional technique separates the designoptimization stage from the physical layout generation stage. Designers have tomanually transfer optimized unit cell parameters into computer-aided designsoftware to generate the final layout. This manual transfer introduces potentialerrors and inconsistencies between simulated and fabricated designs.

[0011] Consequently, the conventional techniques for reflect array antennadesign suffer from fragmented workflows, limited automation, and highcomputational cost. These limitations restrict the efficiency of reflect arrayantenna design for high-frequency and large-aperture applications.

[0012] Therefore, there is a need to overcome one or more of the abovementioned limitations.SUMMARY

[0013] This summary is provided to introduce a selection of concepts, in asimplified format, that are further described in the detailed description of theinvention. This summary is neither intended to identify key or essential inventiveconcepts of the invention nor is it intended for determining the scope of theinvention.

[0014] According to one embodiment of the present disclosure, a method forgenerating a design for a reflect array is disclosed. The method includes receivinga set of design input parameters. The set of design input parameters comprises atleast one of an angle of incidence of an incoming radio signal from a transmittingsource, an angle of reflection at which the reflected signal is directed toward areceiver source, one or more operating frequency values corresponding to thereflect array's frequency bands, incident and reflection elevation and azimuthangles, and physical dimensions of the reflect array substrate, number of unitcells MxN to be arranged on the reflect array surface. Further, the methodincludes determining a phase shift profile matrix of the reflect array based on theset of design input parameters. Furthermore, the method includes predicting anoptimal physical dimension of each unit cell on the reflect array for a givengeometric shape based on the determined phase shift profile matrix. The methodpredicts the optimal physical dimension by applying a trained machine learningmodel. Additionally, the method includes generating a Printed Circuit Board(PCB) file based on the determined optimal physical dimension. The PCB fileindicates a PCB layout. The PCB layout includes a geometric layout of each unitcell for fabrication.

[0015] According to another embodiment of the present disclosure, a systemfor generating a design for a reflect array is disclosed. The system includes aprocessor and a memory coupled with the processor. The processor is configuredto receive a set of design input parameters. The set of design input parameterscomprises at least one of an angle of incidence of an incoming radio signal froma transmitting source, an angle of reflection at which the reflected signal isdirected toward a receiver source, one or more operating frequency valuescorresponding to the reflect array's frequency bands, incident and reflectionelevation and azimuth angles, and physical dimensions of the reflect arraysubstrate, number of unit cells MxN to be arranged on the reflect array surface.Further, the processor is configured to determine a phase shift profile matrix ofthe reflect array based on the set of design input parameters. Furthermore, theprocessor is configured to predict an optimal physical dimension of each unit cellfrom the one or more unit cells for a given geometric shape based on thedetermined phase shift profile matrix. The processor predicts the optimalphysical dimension by applying a trained machine learning model. Additionally,the processor is configured to generate a PCB file based on the determinedoptimal physical dimension. The PCB file indicates the PCB layout. The PCBlayout includes the geometric layout of each unit cell for fabrication.

[0016] To further clarify the advantages and features of the present invention,a more particular description of the invention will be rendered by reference tospecific embodiments thereof, which are illustrated in the appended drawings. Itis appreciated that these drawings depict only typical embodiments of theinvention and are therefore not to be considered limiting of its scope. Theinvention will be described and explained with additional specificity and detailin the accompanying drawings.BRIEF DESCRIPTION OF THE DRAWINGS

[0017] These and other features, aspects, and advantages of the presentinvention will become better understood when the following detailed descriptionis read with reference to the accompanying drawings in which like charactersrepresent like parts throughout the drawings, wherein:

[0018] Figure 1A illustrates an environment for implementation of a systemfor generating a design for a reflect array, according to an embodiment of thepresent disclosure;

[0019] Figure 1B illustrates an exemplary design of the reflect array, accordingto an embodiment of the present disclosure;

[0020] Figure 1C illustrates an exemplary design of an exemplary unit cell ofthe reflect array, according to an embodiment of the present disclosure;

[0021] Figure 2 illustrates a block diagram of the system for generating thedesign of the reflect array, according to an embodiment of the present disclosure;

[0022] Figure 3A illustrates a process flow associated with a phase shift profilematrix determining module of the system, according to an embodiment of thepresent disclosure;

[0023] Figure 3B illustrates an exemplary phase shift profile matrix determinedby the phase shift profile matrix determining module, according to anembodiment of the present disclosure;

[0024] Figure 4 illustrates a process flow associated with a physical dimensionpredicting module of the system, according to an embodiment of the presentdisclosure;

[0025] Figures 5-6 illustrate exemplary process flows for generating the designof the reflect array by the system, according to an embodiment of the presentdisclosure;

[0026] Figure 7 illustrates an exemplary process flow for training the machinelearning model used by the system, according to an embodiment of the presentdisclosure; and

[0027] Figure 8 illustrates a flow chart depicting a method for generating adesign for a reflect array, according to an embodiment of the present disclosure.

[0028] Further, skilled artisans will appreciate that elements in the drawingsare illustrated for simplicity and may not have necessarily been drawn to scale.For example, the flow charts illustrate the method in terms of the most prominentsteps involved to help to improve understanding of aspects of the presentinvention. Furthermore, in terms of the construction of the device, one or morecomponents of the device may have been represented in the drawings byconventional symbols, and the drawings may show only those specific detailsthat are pertinent to understanding the embodiments of the present invention soas not to obscure the drawings with details that will be readily apparent to thoseof ordinary skill in the art having the benefit of the description herein.DETAILED DESCRIPTION

[0029] For the purpose of promoting an understanding of the principles of theinvention, reference will now be made to the various embodiments, and specificlanguage will be used to describe the same. It will nevertheless be understoodthat no limitation of the scope of the invention is thereby intended, suchalterations and further modifications in the illustrated system, and such furtherapplications of the principles of the invention as illustrated therein beingcontemplated as would normally occur to one skilled in the art to which theinvention relates.

[0030] It will be understood by those skilled in the art that the foregoinggeneral description and the following detailed description are explanatory of theinvention and are not intended to be restrictive thereof.

[0031] Reference throughout this specification to "an aspect," "another aspect"or similar language means that a particular feature, structure, or characteristicdescribed in connection with the embodiment is included in at least oneembodiment of the present invention. Thus, appearances of the phrase "in anembodiment," "in another embodiment" and similar language throughout thisspecification may, but do not necessarily, all refer to the same embodiment.

[0032] The terms "comprises", "comprising", or any other variations thereof,are intended to cover a non-exclusive inclusion, such that a process or methodthat comprises a list of steps does not include only those steps but may includeother steps not expressly listed or inherent to such process or method. Similarly,one or more devices or sub-systems or elements or structures or componentsproceeded by "comprises... a" does not, without more constraints, preclude theexistence of other devices or other sub-systems or other elements or otherstructures or other components or additional devices or additional sub-systemsor additional elements or additional structures or additional components.

[0033] The embodiments herein and the various features and advantageousdetails thereof are explained more fully with reference to the non-limitingembodiments that are illustrated in the accompanying drawings and detailed inthe following description. Descriptions of well-known components andprocessing techniques are omitted so as to not unnecessarily obscure theembodiments herein. Also, the various embodiments described herein are notnecessarily mutually exclusive, as some embodiments can be combined with oneor more other embodiments to form new embodiments. The term "or" as usedherein, refers to a non-exclusive or unless otherwise indicated. The examplesused herein are intended merely to facilitate an understanding of ways in whichthe embodiments herein can be practiced and to further enable those skilled inthe art to practice the embodiments herein. Accordingly, the examples should notbe construed as limiting the scope of the embodiments herein.

[0034] As is traditional in the field, embodiments may be described andillustrated in terms of blocks that carry out a described function or functions.These blocks, which may be referred to herein as units or modules or the like,are physically implemented by analog or digital circuits such as logic gates,integrated circuits, microprocessors, microcontrollers, memory circuits, passiveelectronic components, active electronic components, optical components,hardwired circuits, or the like, and may optionally be driven by firmware andsoftware. Each block of the embodiments may be physically separated into twoor more interacting and discrete blocks without departing from the scope of theinvention. Likewise, the blocks of the embodiments may be physically combinedinto more complex blocks without departing from the scope of the invention.

[0035] The accompanying drawings are used to help easily understand varioustechnical features and it should be understood that the embodiments presentedherein are not limited by the accompanying drawings. As such, the presentdisclosure should be construed to extend to any alterations, equivalents, andsubstitutes in addition to those which are particularly set out in the accompanyingdrawings. Although the terms first, second, etc. may be used herein to describevarious elements, these elements should not be limited by these terms. Theseterms are generally only used to distinguish one element from another.

[0036] Figure 1A illustrates an environment 100 for the implementation of asystem 108 for generating a design for a reflect array 106, according to anembodiment of the present disclosure.

[0037] In an embodiment, the reflect array 106 may be a planar antennastructure. The reflect array 106 may include a plurality of unit cells arranged ona reflect array surface. Each unit cell among the plurality of unit cells may beconfigured to manipulate electromagnetic waves. The reflect array 106 mayredirect an incoming radio signal toward a desired direction.

[0038] In an embodiment, the reflect array 106 may be a passive antennaelement. The reflect array 106 may provide beam steering capability. The reflectarray 106 may be utilized for applications such as, but not limited to, wirelesscommunication and radar.

[0039] In an embodiment, the reflect array 106 may be implemented in theenvironment 100 requiring redirection of radio signals. The environment 100may include scenarios where direct line-of-sight communication between atransmitting source 102 and a receiver source 104 may be obstructed. Theenvironment 100 may include scenarios where signal coverage enhancementmay be required. The environment 100 may include scenarios where beamshaping may be necessary.

[0040] The reflect array 106 may receive an incoming radio signal from atransmitting source 102. The transmitting source 102 may be a transmittingantenna or a base station. The incoming radio signal may travel from thetransmitting source 102 to the reflect array 106 along an incident path. Theincoming radio signal may arrive at the reflect array 106 at an angle of incidence.The angle of incidence may be measured relative to a normal of the reflect arraysurface.

[0041] The reflect array 106 may reflect the incoming radio signal as areflected signal. The reflected signal may be directed towards a receiver source104. The receiver source 104 may be an antenna, a user equipment, or acommunication device. The reflected signal may travel from the reflect array 106to the receiver source 104 along a reflection path. The reflected signal may leavethereflectarray106atanangleofreflection.Theangleofreflectionmaybemeasured relative to the normal of the reflect array surface. The angle ofreflection may be a desired angle predetermined based on the location of thereceiver source 104.

[0042] In an embodiment, the system 108 may be configured to generate thedesign of the reflect array 106.

[0043] Figure 1B illustrates an exemplary design of the reflect array 106,according to an embodiment of the present disclosure.

[0044] In an embodiment, the reflect array 106 may include a plurality of unitcells. Each unit cell 110 among the plurality of unit cells may be configured tointroduce a phase shift to the incoming radio signal. Each unit cell 110 amongthe plurality of unit cells may have a geometric shape. The geometric shape maybe a circular shape, a rhombus shape, a square shape, a pentagon shape, and thelike.

[0045] The plurality of unit cells may be arranged in a matrix arrangement onthe reflect array surface. The matrix arrangement may be defined by a numberof unit cells MxN. M may represent a number of rows. N may represent a numberof columns. In an exemplary embodiment shown in Figure 1B, the reflect array106 may include a 5x5 matrix arrangement of the plurality of unit cells.

[0046] In an embodiment, the reflect array 106 may have physical dimensionsdefined along an x-axis and a y-axis. The physical dimensions of the reflect array106 may determine an overall size of the reflect array surface. The reflect array106 may include a reflect array substrate. The reflect array substrate may supportthe plurality of unit cells 110. The reflect array substrate may have physicaldimensions corresponding to the physical dimensions of the reflect array 106.

[0047] In an embodiment, the reflect array 106 may have a metallic reflectingsurface. The metallic reflecting surface may be disposed on the reflect arraysubstrate. The reflect array substrate may include various Printed Circuit Board(PCB) substrates. The PCB substrates may include different dielectric materials.The PCB substrates may have different dielectric constants.

[0048] Figure 1C illustrates an exemplary design of an exemplary unit cell 110of the reflect array 106, according to an embodiment of the present disclosure.

[0049] In an embodiment, the unit cell 110 may include a circular metallicpatch. The circular metallic patch may have a circular shape. The circularmetallic patch may have a radius R. The radius R may be a configurableparameter. The radius R may determine a phase shift introduced by the unit cell110.

[0050] In an embodiment, the circular metallic patch may be surrounded by agap. The gap may be defined by a parameter dg. The parameter dg may representa gap size. The gap may separate the circular metallic patch from surroundingstructures.

[0051] In an embodiment, the circular metallic patch may be surrounded by afixed-width metallic frame. The fixed-width metallic frame may be disposedaround the circular metallic patch. The fixed-width metallic frame may beseparated from the circular metallic patch by the gap. The fixed-width metallicframe may have a width defined by a parameter dt. The parameter dt mayrepresent a thickness of the fixed-width metallic frame.

[0052] In an embodiment, the unit cell 110 may have unit cell dimensions. Theunit cell dimensions may include geometric parameters. The geometricparameters may include the radius R for the circular shape. The geometricparameters may include a patch length, width for other geometric shapes. Thegeometric parameters may include the gap size defined by the parameter dg. Theunit cell dimensions may determine electromagnetic properties of the unit cell110.

[0053] In an embodiment, the plurality of unit cells may be separated by a unitcell gap. The unit cell gap may be defined by the parameters dt and dg. The unitcell gap may provide spacing between adjacent unit cells among the plurality ofunit cells.

[0054] In an embodiment, the design of the reflect array 106 may mean aspecification of the physical dimensions of each unit cell among the plurality ofunit cells 110. The design of the reflect array 106 may include the geometricshape of each unit cell. The design of the reflect array 106 may include the unitcell dimensions for each unit cell. The design of the reflect array 106 may includetheunitcellgapbetweenadjacentunitcells.Thedesignofthereflectarray106may enable fabrication of the reflect array 106 with desired electromagneticproperties.

[0055] The system 108 for generating the design of the reflect array 106 isdescribed further in reference to Figures 2-8.

[0056] Figure 2 illustrates a block diagram of the system 108 for generatingthe design of the reflect array 106, according to an embodiment of the presentdisclosure.

[0057] In an embodiment, the system 108 may include, but may not be limitedto, a processor 204, a memory 202, modules 206, and data 208. The modules 206may be coupled to the processor 204 and the memory 202.

[0058] The processor 204 may be a single processing unit or several units,wherein all units may include multiple computing units. The processor 204 maybe implemented as one or more microprocessors, microcomputers,microcontrollers, digital signal processors, central processing units, statemachines, logic circuitries, and / or any devices that manipulate signals based onoperational instructions. Among other capabilities, the processor 204 may beadapted to fetch and execute computer-readable instructions and data stored inthe memory 202. The processor 204 may be a general-purpose processor, suchas a central processing unit (CPU), an application processor (AP), or a graphicsprocessing unit (GPU), a visual processing unit (VPU), and / or an AI-dedicatedprocessor such as a neural processing unit (NPU). The processor 204 may controlthe processing of the RF signal parameters in accordance with predefined signalprocessing techniques stored in the memory 202.

[0059] The memory 202 may include any non-transitory computer-readablemedium known in the art including, for example, volatile memory, such as staticrandom-access memory (SRAM) and dynamic random-access memory(DRAM), and / or non-volatile memory, such as read-only memory (ROM),erasable programmable ROM, flash memories, hard disks, optical disks, andmagnetic tapes.

[0060] The data 208 may serve as a repository for storing processed data,received data, and generated data by one or more of the modules 206.

[0061] The modules 206, amongst other things, may include routines,programs, objects, components, data structures, which perform particular tasksor implement data types. The modules 206 may also be implemented as signalprocessor(s), state machine(s), logic circuitries, and / or any other device orcomponent that manipulates signals based on operational instructions.

[0062] Further, the modules 206 may be implemented in hardware, instructionsexecuted by a processing unit, or by a combination thereof. The processing unitmay comprise a computer, a processor, a state machine, a logic array, or anyother suitable devices capable of processing instructions. The processing unitmay be a general-purpose processor which executes instructions to cause thegeneral-purpose processor to perform the required tasks or the processing unitmay be dedicated to performing the required functions. In another embodimentof the present disclosure, the processor 204 via the modules 206 may beconfigured to execute machine-readable instructions (software) which performthe working of the system 108 within the scope of the present disclosure asdescribed in forthcoming paragraphs.

[0063] The modules 206 may include an input parameters receiving module210, a phase shift profile matrix determining module 212, a physical dimensionpredicting module 214, a PCB file generating module 216, a designing module218.

[0064] The detailed operational and functional aspects of each module in theset of modules 206 may be described in subsequent paragraphs in reference toadditional figures.

[0065] In an embodiment, the input parameters receiving module 210 may beconfigured to receive a set of design input parameters. The set of design inputparameters may include the angle of incidence of the incoming radio signal fromthe transmitting source 102, the angle of reflection at which the reflected signalis directed toward the receiver source 104, and the one or more operatingfrequency values. The set of design input parameters may include incidentelevation angles, incident azimuth angles, reflection elevation angles, andreflection azimuth angles. The set of design input parameters may include thephysical dimensions of the reflect array, the physical dimensions of the reflectarray substrate, and the number of unit cells MxN to be arranged on the reflectarray surface.

[0066] In an embodiment, the incident elevation angle may define a verticalangular orientation of the incoming radio signal. The incident elevation anglemay be measured from a normal of the reflect array surface. The incident azimuthangle may define a horizontal angular orientation of the incoming radio signal.The incident azimuth angle may be measured from a reference axis on the reflectarray surface.

[0067] In an embodiment, the reflection elevation angle may define a verticalangular orientation of the reflected signal. The reflection elevation angle may bemeasured from the normal of the reflect array surface. The reflection azimuthangle may define a horizontal angular orientation of the reflected signal. Thereflection azimuth angle may be measured from the reference axis on the reflectarray surface.

[0068] In an embodiment, the one or more operating frequency values maycorrespond to frequency bands of the reflect array 106. The frequency bands maydefine operating ranges for electromagnetic wave propagation. The one or moreoperating frequency values may be specified in Gigahertz (GHz).

[0069] In an embodiment, the phase shift profile matrix determining module212 may be configured to determine a phase shift profile matrix of the reflectarray 106 based on the set of design input parameters. The phase shift profilematrix may indicate phase shift values required at each unit cell among theplurality of unit cells 110. The phase shift profile matrix may include a pluralityof phase shift values arranged in a matrix format corresponding to the matrixarrangement of the plurality of unit cells 110. Each phase shift value in the phaseshift profile matrix may correspond to a specific unit cell at a specific positionon the reflect array surface.

[0070] The phase shift profile matrix determining module 212 is describedfurther in reference to Figures 3A-3B.

[0071] Figure 3A illustrates a process flow 300A associated with the phaseshift profile matrix determining module 212, according to an embodiment of thepresent disclosure.

[0072] In an embodiment, the process flow 300A at step 302 includes receivingthe physical dimensions of the reflect array 106, the physical dimensions of thereflect array substrate, and the unit cell spacing. The unit cell spacing mayindicate distances between adjacent unit cells among the plurality of unit cells110.

[0073] In an embodiment, the process flow 300A at step 304 includescomputing an incident wave direction vector and a reflected wave directionvector. The incident wave direction vector may be computed using the incidentelevation angles and the incident azimuth angles. The reflected wave directionvector may be computed using the reflection elevation angles and the reflectionazimuth angles. The incident wave direction vector may represent a propagationdirection of the incoming radio signal. The reflected wave direction vector mayrepresent a propagation direction of the reflected signal.

[0074] In an embodiment, the process flow 300A at step 306 includesdetermining the phase shift profile matrix by correlating the computed incidentwave direction vector, the computed reflected wave direction vector, the physicaldimensions of the reflect array substrate, and the unit cell gap parameters. Theunit cell gap parameters may include the parameter dt and the parameter dg. Thecorrelation may involve computing phase shift values required at each positionon the reflect array surface. The phase shift values may be computed based on apath difference between the incoming radio signal and the reflected signal at eachunit cell position. The path difference may vary across different positions on thereflect array surface. The phase shift profile matrix may compensate for the pathdifference to achieve constructive interference of the reflected signal in adirection toward the receiver source 104.

[0075] In an embodiment, a target phase shift value Φmn may be determined foreach unit cell position. The unit cell position may be identified by coordinates mand n. The coordinate m may represent a row position. The coordinate n mayrepresent a column position. The target phase shift value Φmn may be determinedbased on coordinates xm and yn of each unit cell on the reflect array surface. Thecoordinates xm and yn may be determined based on the unit cell spacing.

[0076] In an embodiment, the target phase shift value Φmn may be determinedusing the incident wave direction vector and the reflected wave direction vector.The determination may ensure that the incoming radio signal impinging on thereflect array 106 from the incident direction may be redirected toward the desiredreflection direction. The redirection may be achieved through phasemanipulation at each unit cell location among the plurality of unit cells.

[0077] Figure 3B illustrates an exemplary phase shift profile matrix 300Bdetermined by the phase shift profile matrix determining module 212, accordingto an embodiment of the present disclosure.

[0078] In an embodiment, the exemplary phase shift profile matrix 300B mayinclude M rows and N columns.

[0079] In an embodiment, each element in the exemplary phase shift profilematrix 300B may represent a target phase shift value for a corresponding unitcell. The target phase shift value may be measured in degrees. The exemplaryphase shift profile matrix 300B may include phase shift values ranging from -100 degrees to 150 degrees. Different phase shift values may be represented bydifferent shades in the exemplary phase shift profile matrix 300B.

[0080] In an embodiment, the exemplary phase shift profile matrix 300B mayexhibit spatial variation of phase shift values. The spatial variation maycorrespond to the path differences across the reflect array surface. The phase shiftprofile matrix determining module 212 may generate the phase shift profilematrix to achieve a desired beam steering toward the receiver source 104.

[0081] In an embodiment, the exemplary phase shift profile matrix 300B maycorrespond to a 5x5 reflect array operating at 3.5 GHz. The exemplary phaseshift profile matrix 300B may be calculated for specific incident and reflectionangles. The incident elevation angle θi may be 38 degrees. The incident azimuthangle φi may be 90 degrees. The reflection elevation angle θr may be 0 degrees.The reflection azimuth angle φr may be 90 degrees.

[0082] Referring again to Figure 2, in an embodiment, the physical dimensionpredicting module 214 may be configured to predict an optimal physicaldimension of each unit cell among the plurality of unit cells 110 on the reflectarray 106 for a given geometric shape based on the determined phase shift profilematrix by applying a trained machine learning model.

[0083] The physical dimension predicting module 214 is now described indetail in reference to Figure 4.

[0084] Figure 4 illustrates a process flow 400 associated with the physicaldimension predicting module 214, according to an embodiment of the presentdisclosure.

[0085] In an embodiment, the process flow 400 at step 402 includes receivingthe target phase shift value (Φmn) corresponding to each of the unit cells from thedetermined phase shift profile matrix. The target phase shift value (Φmn) may beextracted from the phase shift profile matrix for a specific unit cell position. Theunit cell position may be identified by the coordinates m and n.

[0086] In an embodiment, the process flow 400 at step 404 includes inputtingthe target phase shift value, the one or more operating frequency values, and theangle of incidence to the trained machine learning model. The trained machinelearning model may receive the target phase shift value (Φmn), the one or moreoperating frequency values, the angle of incidence as input parameters.

[0087] In an embodiment, the process flow 400 at step 406 includes predictingthe optimal physical dimension of each of the unit cells based on trained machinelearning model data. The trained machine learning model data may includelearned parameters from a training process. The optimal physical dimension maybe the radius R for the circular shape. The optimal physical dimension may be apatch length, or a width for other geometric shapes. The trained machine learningmodel may output the optimal physical dimension for achieving the target phaseshift value (Φmn) at the one or more operating frequency values.

[0088] In an embodiment, the trained machine learning model may begenerated through a training process. The training process may include datasetgeneration, data preprocessing, model training, and validation stages.

[0089] In an embodiment, dataset generation may include generating a sampledataset. The sample dataset may cover a range of phase values. The sampledataset may be generated by varying physical dimensions of a unit cell structure.The unit cell structure may be analyzed using electromagnetic simulationtechniques. The electromagnetic simulation techniques may include, forinstance, a Floquet port analysis. The Floquet port analysis may ensure that theunit cell structure resonates at desired operating frequency values.

[0090] In an embodiment, the physical dimensions may be systematicallyvaried across a range of values. Phase response data may be collected over theoperating frequency range. The phase response data may capture a relationshipbetween the physical dimensions and the resulting reflection phasecharacteristics.

[0091] In an exemplary embodiment, the unit cell structure may have thecircular shape. The unit cell structure may include the circular metallic patch.The circular metallic patch may have the radius R. The radius R may be variedacross a range of values. The parameter dg may be kept constant at a firstpredetermined value, for instance, 1.2 millimeters. The parameter dt may be keptconstant at second predetermined value, for instance, 1.2 millimeters. The reflectarray substrate may include Flame Retardant - 4 (FR4) material.

[0092] In an embodiment, the Floquet port analysis may be performed on theunit cell structure. The Floquet port analysis may model electromagneticbehavior of the unit cell when arranged in a periodic array configuration. Theunit cell structure may be configured to resonate at the operating frequency, forinstance, 3.5 Gigahertz. The sample dataset of reflection phases may begenerated over the frequency range by varying the radius R.

[0093] In an embodiment, data points corresponding to the operating frequencymay be selected for training. The operating frequency may be among thefrequency band for fifth generation (5G) wireless communication applications.The selected data points may form a training dataset.

[0094] In an embodiment, data preprocessing may include handling phasediscontinuities. The phase discontinuities may occur in raw simulation data. Thephasediscontinuitiesmayarisewhenphaseresponsewrapsaroundfrom+180degrees to -180 degrees. A 360-degree phase shift may occur at a discontinuitypoint. The 360-degree phase shift may create challenges for machine learningmodels.

[0095] In an embodiment, the discontinuity may arise at a specific radius value.In an embodiment, data preprocessing may include applying a phase offset. Thephase offset may be 360 degrees. The phase offset may be added to valuescorresponding to radius values below or equal to a specific radius value. Thephase offset application may result in a continuous plot. The continuous plot mayeliminate ambiguity at the discontinuity point. The continuous plot may beeffectively approximated using machine learning techniques.

[0096] In an embodiment, data preprocessing may include datasetstandardization. Dataset standardization may include normalizing input features.Input features may be normalized to have zero mean and unit variance. Datasetstandardization may include scaling features.

[0097] In an embodiment, machine learning model training may includeevaluating various model architectures. The model architectures may includeunivariate spline interpolation, sigmoid functions, modified sigmoid functions,hyperbolic tangent functions, polynomial regression, Artificial Neural Networks(ANNs).

[0098] In an embodiment, model architectures may be evaluated based onperformance metrics. The performance metrics may include Mean AbsoluteError (MAE), Mean Squared Error (MSE). The MAE may measure an averagemagnitude of errors in predictions. The MSE may measure a squared average oferrors in predictions.

[0099] In an embodiment, the trained machine learning model may include theANN. The ANN may include three layers. The three layers may include an inputlayer, a hidden layer, an output layer.

[0100] In an embodiment, the number of neurons in the hidden layer may bedetermined empirically. The number of neurons may be optimized based onprediction error. For example, configurations with fewer than 128 neurons mayresult in unacceptable error levels. A configuration with at least 128 neurons maybe required to achieve acceptable model fit. Beyond 128 neurons, the MAE andMSE metrics may exhibit saturation. Further increasing the number of neuronsmay lead to higher computational overhead.

[0101] In an embodiment, once the machine learning model is trained, thetrained machine learning model may predict the radius R for any given phasevalue. The given phase value may be within a training range. The trainedmachine learning model may enable accurate prediction of physical dimensionsof the reflect array 106. The trained machine learning model may generatedimensions for all unit cells the reflect array 106 based on respective phaserequirements from the phase shift profile matrix.

[0102] In an embodiment, for the exemplary 5x5 reflect array 106, the phaseshift profile matrix may provide 25 phase values. Each phase value maycorrespond to one unit cell among the plurality of unit cells 110. Each phasevalue Φmn may be input to the trained machine learning model. The trainedmachine learning model may output a corresponding radius Rmn for each unitcell. In an advantageous aspect, the prediction process may be significantly fasterthan iterative electromagnetic simulations.

[0103] Referring again to Figure 2, in an embodiment, the PCB file generatingmodule 216 may be configured to generate a PCB file based on the determinedoptimal physical dimension. The PCB file may indicate a PCB layout. The PCBlayout may include a geometric layout of each unit cell for fabrication.

[0104] In an embodiment, the PCB file generation process may include formatselection. Format selection may include selecting an appropriate PCB design fileformat. The PCB design file format may be compatible with Electronic DesignAutomation (EDA) tools.

[0105] In an embodiment, the PCB file generation process may includedefining layout. Layout definition may include translating the number of unitcells MxN with individual dimensions into the PCB layout format. Each unit cellamong the plurality of unit cells 110 may have the optimal physical dimensionpredicted by the physical dimension predicting module 214.

[0106] In an embodiment, the PCB file generation process may includecoordinate system mapping. The coordinate system mapping may includemapping theoretical unit cell positions to physical coordinates on the PCB. Thetheoretical unit cell positions may be derived from phase shift. The physicalcoordinates may represent actual positions on the PCB for fabrication.

[0107] In an embodiment, the PCB file generation process may includegeometric representation. Geometric representation may include accuraterepresentation of each unit cell's geometry with the predicted physicaldimensions.

[0108] In an embodiment, the PCB file generation process may includeconfiguration of layers. The configuration of layers may include defining metallayers, substrate layers, PCB stackup information on the PCB.

[0109] In an embodiment, the PCB file generation process may includeapplying one or more design rules such as manufacturing design rules,manufacturing constraints to ensure the fabricability of the design.

[0110] In an embodiment, the PCB file may be used to create Gerber files.Gerber files may be necessary for PCB manufacturing. Gerber files may beindustry-standard file formats for communicating PCB designs to manufacturers.

[0111] In an exemplary embodiment, the PCB file generating module 216 mayutilize KiCAD for PCB design automation. KiCAD is an open-source tool forelectronic design automation. KiCAD may provide comprehensive features forPCB layout.

[0112] In an embodiment, a template PCB design may be initially created usinga KiCAD editor. The template PCB design may be reverse-engineered tounderstand a file structure. File parameters and element coordinates may bemodified programmatically.

[0113] In an embodiment, a stroke-based method may be used for circular cellimplementation. For the circular metallic patch, a circle with radius R / 2 may becreated. A stroke width of R may be assigned to the circle. The stroke widthapplication may produce a filled circle with an effective radius of R.

[0114] In an embodiment, for the fixed-width metallic frame, a circle withradius R + dg + dt / 2 may be defined. A stroke width of dt may be assigned. Thestroke width application may form a ring of thickness dt. The ring may bepositioned with the gap dg from the circular metallic patch.

[0115] In an embodiment, the stroke-based approach may produce a simplerfile structure. The stroke-based method may produce smooth circles withoutpolygon approximation. The stroke-based method may be easier toprogrammatically generate. The stroke-based method may reduce file size.

[0116] In an embodiment, the designing module 218 may be configured todesign the reflect array 106 based on the PCB layout defined in the generatedPCB file. The designing module 218 may facilitate fabrication of the reflect array106. The designing module 218 may use the PCB file to facilitate themanufacturing of the reflect array 106. The reflect array 106 may be fabricatedaccording to the geometric layout specified in the PCB file. The fabricated reflectarray 106 may achieve desired beam steering characteristics.

[0117] Figures 5-6 illustrate exemplary process flows for generating design ofthe reflect array 106 by the system 108, according to an embodiment of thepresent disclosure.

[0118] In an embodiment, at block 502, the set of input parameters may bereceived. The set of input parameters may include geometry parameters,constraints, and optimization criteria.

[0119] In an embodiment, at block 504, drawing may be automated using thetrained machine learning model. The trained machine learning model maydirectly generate the design.

[0120] In an embodiment, at block 506, confidence in the design may beevaluated. Confidence in the design may represent a level of certainty in thepredicted optimal physical dimensions. A decision may be made whether toproceed with the design or verify the design performance throughelectromagnetic simulation.

[0121] In an embodiment, at block 508, High Frequency Structure Simulator(HFSS) simulation may be skipped when confidence in the design is high. HFSSmay be an electromagnetic simulation technique. At block 510, time and effortmay be saved by skipping electromagnetic simulation.

[0122] In an embodiment, at block 512, HFSS simulation may be run forverification of performance when confidence in the design is low. HFSSsimulation may verify electromagnetic performance of the design.

[0123] In an embodiment, at block 514, de-embedding files may be extractedfrom HFSS or Computer Simulation Technology (CST) tools. De-embeddingfiles may contain electromagnetic characteristics of the unit cell.

[0124] In an embodiment, at block 516, Gerber generation may be performed.Gerber files may be generated based on the design.

[0125] In an embodiment, at block 518, PCB fabrication may be performed.The PCB may be manufactured based on the Gerber files.

[0126] Referring to Figure 6, in an embodiment, at block 602, the incidentwave direction and required direction of reflected wave may be provided asinputs. The incident wave direction may correspond to the angle of incidence.The required direction of the reflected wave may correspond to the angle ofreflection.

[0127] In an embodiment, at block 604, frequency, number of unit cells, andsize of reflect array may be provided as inputs. The frequency may correspondto the one or more operating frequency values. The number of unit cells may bethe number of unit cells MxN. The size of the reflect array 106 may correspondto the physical dimensions of the reflect array 106.

[0128] In an embodiment, at block 606, a codeword generator may generate aphase shift profile matrix. The codeword generator may compute the phase shiftprofile matrix based on the incident wave direction and the required direction ofthe reflected wave.

[0129] In an embodiment, at block 608, the phase shift profile matrix requiredmay be determined. The phase shift profile matrix may include target phase shiftvalues for each unit cell among the plurality of unit cells 110.

[0130] In an embodiment, at block 610, the trained machine learning modelmay process the phase shift profile matrix required.

[0131] In an embodiment, at block 612, dimensions of each unit cell may bepredicted. The dimensions may be the optimal physical dimension. The trainedmachine learning model may output the optimal physical dimension.

[0132] In an embodiment, at block 614, a Gerber file generator may generateGerber files. The Gerber file generator may use the optimal physical dimensionof each unit cell. The Gerber file generator may convert the PCB file into Gerberfile format.

[0133] In an embodiment, at block 616, the Gerber file may be produced.

[0134] In an embodiment, at block 618, design validation and fabrication maybe performed. Design validation may include verification of the designspecifications. The reflect array 106 may be fabricated based on the validateddesign.

[0135] In an embodiment, at block 620, testing may be performed. Thefabricated reflect array 106 may be tested for electromagnetic performance.

[0136] In an embodiment, at block 622, test results may be obtained. The testresults may verify performance of the reflect array 106.

[0137] Figure 7 illustrates an exemplary process flow for training the machinelearning model used by the system 108, according to an embodiment of thepresent disclosure.

[0138] In an embodiment, at block 702, unit cell geometry may be defined. Theunit cell geometry may include the geometric shape, and the unit cell dimensions.

[0139] In an embodiment, at block 704, electromagnetic simulations may beperformed. The electromagnetic simulations may be performed for varyingdimensions of the unit cell. A small number of electromagnetic simulations maybe sufficient to generate the sample dataset.

[0140] In an embodiment, at block 706, dimensions of the unit cell andcorresponding phase shift values may be obtained. The electromagneticsimulations may generate the sample dataset. The sample dataset may include arelationship between the unit cell dimensions and the phase shift values.

[0141] In an embodiment, at block 708, machine learning model training maybe performed. The sample dataset may be used to train the machine learningmodel. The machine learning model training may include data preprocessing,model architecture selection, and parameter optimization.

[0142] In an embodiment, at block 710, testing and validation of the machinelearning model may be performed. The machine learning model may beevaluated for accuracy using performance metrics.

[0143] In an embodiment, at block 712, the trained machine learning modelmay be generated. The trained machine learning model may be ready for use inpredicting optimal physical dimensions based on target phase shift values.

[0144] Figure 8 illustrates a flow chart depicting a method 800 for generatinga design for a reflect array, according to an embodiment of the present disclosure.The method 800 may be a computer-implemented method. The method 800 maybe executed by the system 108 and the modules 206. For the sake of brevity,constructional and operational features of the system 108 that are alreadyexplained in the description of Figures 1-7 are not explained in detail in thedescription of Figure 8.

[0145] The method 800, at step 802, involves receiving the set of design inputparameters. The set of design input parameters may include the angle ofincidence of the incoming radio signal from the transmitting source 102. The setof design input parameters may include the angle of reflection at which thereflected signal is directed toward the receiver source 104. The set of design inputparameters may include the one or more operating frequency valuescorresponding to frequency bands of the reflect array 106. The set of design inputparameters may include the incident elevation angles, the incident azimuthangles, the reflection elevation angles, and the reflection azimuth angles. The setof design input parameters may include the physical dimensions of the reflectarray 106. The set of design input parameters may include the physicaldimensions of the reflect array substrate. The set of design input parameters mayincludethenumberofunitcellsMxNtobearrangedonthereflectarraysurface.

[0146] Thereafter, at step 804, the method 800 involves determining the phaseshift profile matrix of the reflect array 106 based on the set of design inputparameters. The phase shift profile matrix may include target phase shift valuesfor each unit cell among the plurality of unit cells 110.

[0147] Subsequently, at step 806, the method 800 involves predicting theoptimal physical dimension of each unit cell on the reflect array 106 for the givengeometric shape. The prediction may be based on the determined phase shiftprofile matrix. The prediction may be performed by applying the trained machinelearning model. The trained machine learning model may receive the target phaseshift value (Φmn) corresponding to each unit cell. The trained machine learningmodel may receive the one or more operating frequency values. The trainedmachine learning model may receive the angle of incidence. The trained machinelearning model may output the optimal physical dimension for each unit cellamong the plurality of unit cells 110.

[0148] Furthermore, at step 808, the method 800 involves generating the PCBfile based on the determined optimal physical dimension. The PCB file mayindicate the PCB layout. The PCB layout may include the geometric layout ofeach unit cell for fabrication.

[0149] At least by virtue of the aforesaid, the present subject matter at leastprovides the following advantages:

[0150] The present disclosure herein provides an automated reflect arraydesign methodology employing a trained machine learning model to predictoptimal physical dimensions of unit cells based on required phase shift values,thereby eliminating the need for repetitive electromagnetic simulations andsubstantially reducing the tedious computational effort involved in conventionalreflect array design.

[0151] The present disclosure herein enables completion of the reflect arraydesign process in seconds to minutes, compared to conventional methodsrequiring days or weeks, thereby dramatically reducing design time andaccelerating time-to-market for reflect array-based wireless communicationsystems.

[0152] The present disclosure herein achieves submillimeter accuracy indimension prediction with a, thereby ensuring that predicted dimensions meetpractical PCB fabrication tolerances and maintain electromagnetic performancespecifications.

[0153] The present disclosure herein provides complete automation of theworkflow from receiving design input parameters to generating fabricationready Printed Circuit Board files, thereby streamlining the design-tomanufacturing process and reducing overall engineering effort.

[0154] The present disclosure herein utilizes a trained machine learning modelthat requires minimal computational resources for prediction once trained,thereby enabling rapid dimension predictions without requiring access tocomputationally intensive electromagnetic simulation software during the designphase.

[0155] The present disclosure herein enables scalability to larger reflect arrayconfigurations without proportional increase in design time, thereby allowingefficient design of arrays with hundreds or thousands of unit cells whilemaintaining the same rapid prediction capability.

[0156] The present disclosure herein eliminates manual dimension selection,thereby reducing human error in dimension specification and ensuringconsistency across the entire reflect array design.

[0157] The present disclosure herein facilitates rapid exploration of multipledesign variations by enabling quick regeneration of designs with different inputparameters, thereby allowing engineers to optimize reflect array performancethrough efficient design iteration.

[0158] The present disclosure herein generates PCB files directly in industry-standardformats compatible with electronic design automation tools, therebyenabling seamless transition from design to Gerber file generation andsubsequent fabrication without intermediate manual conversion steps.

[0159] The present disclosure herein reduces engineering costs throughsubstantial reduction in design time and computational resource requirements,thereby making reflect array technology more economically viable for a widerrange of wireless communication applications.

[0160] While specific language has been used to describe the disclosure, anylimitations arising on account of the same are not intended. As would be apparentto a person in the art, various working modifications may be made to the methodin order to implement the inventive concept as taught herein.

[0161] The drawings and the forgoing description give examples ofembodiments. Those skilled in the art will appreciate that one or more of thedescribed elements may well be combined into a single functional element.Alternatively, certain elements may be split into multiple functional elements.Elements from one embodiment may be added to another embodiment. Forexample, orders of processes described herein may be changed and are notlimited to the manner described herein.

Claims

1. A method (800) for generating a design for a reflect array (106), the method (800) comprising: receiving (802) a set of design input parameters, the set of design input parameters comprises at least one of an angle of incidence of an incoming radio signal from a transmitting source (102), an angle of reflection at which the reflected signal is directed toward a receiver source (104), one or more operating frequency values corresponding to the reflect array's frequency bands, incident and reflection elevation and azimuth angles, and physical dimensions of the reflect array substrate, number of unit cells MxN to be arranged on the reflect array surface; determining (804) a phase shift profile matrix of the reflect array (106) based on the set of design input parameters; predicting (806) an optimal physical dimension of each unit cell on the reflect array (106) for a given geometric shape based on the determined phase shift profile matrix by applying a trained machine learning model; and generating (808) a Printed Circuit Board (PCB) file based on the determined optimal physical dimension, wherein the PCB file indicates a PCB layout, including a geometric layout of each unit cell for fabrication.

2. The method (800) as claimed in claim 1, further comprising: designing the reflect array based on the PCB layout defined in the generated PCB file.

3. The method (800) as claimed in claim 1, wherein the set of design input parameters includes a geometry shape of the reflect array (106), unit cell gap (dt, dg), and unit cell dimensions.

4. The method (800) as claimed in claim 1, wherein the reflect array (106) has a metallic reflecting surface and various PCB substrates.

5. The method (800) as claimed in claim 1, wherein the geometry shape of the unit cell includes a circular shape, a rhombus shape, a square shape, a pentagon shape.

6. The method (800) as claimed in claim 5, wherein the circular unit cell comprises a circular metallic patch.

7. The method (800) as claimed in claim 6, wherein the circular metallic patch has a radius R, and is surrounded by a fixed-width metallic frame and gap defined by the parameters dt and dg, respectively.

8. The method (800) as claimed in claim 1, wherein the unit cell dimensions comprise geometric parameters such as patch length, width, and the gap size of the reflect array (106) substrate.

9. The method (800) as claimed in claim 1, wherein determining (804) the phase shift profile matrix comprises: receiving the physical dimensions of the reflect array (106), substrate and the unit cell spacing; computing an incident wave direction vector and a reflected wave direction vector using the input incident and reflection elevation and azimuth angles; and determining the phase shift profile matrix by correlating the computed incident wave and the reflected wave direction vector, the physical dimensions of the reflect array substrate, and the unit cell gap parameters.

10. The method (800) as claimed in claim 1, wherein predicting (806) the optimal physical dimension of each of the unit cells comprises: receiving a target phase shift value (Φmn) corresponding to each of the unit cells from the determined phase shift profile matrix; inputting the target phase shift value, the one or more operating frequency values, and the incidence angle to the trained machine learning model; and predicting the optimal physical dimension of each of the unit cells based on the trained machine learning model data.

11. A system (108) for generating a design for a reflect array (106), the system (108) comprising: a processor (204) and a memory (202) coupled with the processor (204), wherein the processor (204) is configured to: receive a set of design input parameters, the set of design input parameters comprises at least one of an angle of incidence of an incoming radio signal from a transmitting source (102), an angle of reflection at which the reflected signal is directed toward a receiver source (104), one or more operating frequency values corresponding to the reflect array's frequency bands, incident and reflection elevation and azimuth angles, and physical dimensions of the reflect array (106), substrate, number of unit cells MxN to be arranged on the reflect array surface; determine a phase shift profile matrix of the reflect array (106) based on the set of design input parameters; predict an optimal physical dimension of each unit cell from the one or more unit cells for a given geometric shape based on the determined phase shift profile matrix by applying a trained machine learning model; and generate a PCB file based on the determined optimal physical dimension, wherein the PCB file indicates the PCB layout, including the geometric layout of each unit cell for fabrication.

12. The system (108) as claimed in claim 11, wherein the processor (204) is further configured to design the reflect array (106) based on the PCB layout defined in the generated PCB file.

13. The system (108) as claimed in claim 11, wherein the set of design input parameters includes a geometry shape of the reflect array (106), unit cell gap (dt, dg), and unit cell dimensions.

14. The system (108) as claimed in claim 11, wherein the reflect array (106) has a metallic reflecting surface and various PCB substrates.

15. The system (108) as claimed in claim 11, wherein the geometry shape of the unit cell includes a circular shape, a rhombus shape, a square shape, a pentagon shape.

16. The system (108) as claimed in claim 15, wherein the circular unit cell comprises a circular metallic patch.

17. The system (108) as claimed in claim 16, wherein the circular metallic patch has a radius R, and is surrounded by a fixed-width metallic frame and gap defined by the parameters dt and dg, respectively.

18. The system (108) as claimed in claim 11, wherein the unit cell dimensions comprise geometric parameters such as patch length, width, and the gap size of the reflect array (106) substrate.

19. The system (108) as claimed in claim 11, wherein to determine the phase shift profile matrix, the processor (204) is configured to: receive the physical dimensions of the reflect array (106), substrate and the unit cell spacing; compute an incident wave direction vector and a reflected wave direction vector using the input incident and reflection elevation and azimuth angles; and determine the phase shift profile matrix by correlating the computed incident wave and the reflected wave direction vector, the physical dimensions of the reflect array substrate, and the unit cell gap parameters.

20. The system (108) as claimed in claim 11, wherein to predict the optimal physical dimension of each of the unit cells, the processor (204) is configured to: receive a target phase shift value (Φmn) corresponding to each of the unit cells from the determined phase shift profile matrix; inputting the target phase shift value, the one or more operating frequency values, and the incidence angle to the trained machine learning model; and predict the optimal physical dimension of each of the unit cells based on the trained machine learning model data.