Radio frequency matching network design method and device based on material library constraints, equipment and storage medium

Abstract

The application discloses a material library constraint-based radio frequency matching network design method and device, equipment and storage medium, and the method comprises the steps of obtaining a material index space, determining a target impedance and a target matching configuration according to a current matching demand; constructing a current search space based on the target matching configuration and the material index space, and running a differential evolution algorithm in the current search space to obtain an optimal nominal value combination; mapping the optimal nominal value combination to an integer index in the material index space, enumerating all candidate material combinations in the neighborhood of the integer index, determining impedance errors of the candidate material combinations, selecting the candidate material combination with the minimum error as a current optimal discrete solution, converting the current optimal discrete solution back to a continuous value, and returning to iteration until an optimal material combination for designing the target matching network is output. The application embeds the material library discrete constraint in the optimization process, adopts a double-layer search architecture of an outer continuous differential evolution and an inner discrete neighborhood enumeration, eliminates the disconnection between design and material selection, and thus improves the efficiency and accuracy.

Description

Technical Field

[0001] This application relates to the field of radio frequency network technology, and in particular to a method, apparatus, device and storage medium for designing radio frequency matching networks based on a library constraint. Background Technology

[0002] In radio frequency (RF) circuit design, matching networks are used to achieve impedance matching between the source and load ends, maximizing power transfer or optimizing performance metrics such as error vector amplitude and receiver sensitivity. Engineers typically use simulation software to design matching networks, obtaining theoretically optimal component values ​​through schematic simulation. In actual engineering, after the design is completed, the theoretical component values ​​need to be converted into a list of available materials, which engineers then manually select and replace components based on the nominal values ​​in the material library.

[0003] However, the theoretical component values ​​output by simulation software often cannot be precisely matched in standardized material libraries. Engineers need to find approximate nominal values ​​in the material library and perform repeated iterative verification. This process relies on human experience, which prolongs the design cycle. At the same time, because the optimization process searches for the optimal theoretical value in a continuous space, the output result cannot directly correspond to the discrete nominal value in the material library, resulting in a deviation between the actual circuit performance after approximation and the simulation results. Summary of the Invention

[0004] The main objective of this application is to provide a method, apparatus, device, and storage medium for designing radio frequency matching networks based on a material library constraint, aiming to solve the technical problems in existing radio frequency matching network designs, such as extended design cycles and deviations between actual circuit performance and simulation results caused by the mismatch between theoretical component values ​​and discrete nominal values ​​in the material library.

[0005] To achieve the above objectives, this application proposes a method for designing radio frequency matching networks based on library constraints, the method comprising: Obtain the material index space and determine the target impedance and target matching configuration based on the current matching requirements; Based on the target matching configuration and the material index space, a current search space is constructed, and a differential evolution algorithm is run in the current search space with the goal of minimizing impedance error. The optimal nominal value combination is obtained by searching. The impedance error is the distance between the input impedance corresponding to the continuous value in the current search space and the target impedance. The optimal nominal value combination is mapped to an integer index in the material index space, and all candidate material combinations are enumerated in the neighborhood of the integer index. Determine the impedance error corresponding to each candidate material combination, and select the candidate material combination with the smallest impedance error as the current optimal discrete solution; After converting the current optimal discrete solution back to a continuous value, the process returns to the step of running the differential evolution algorithm until the termination condition is met. Then, the optimal material combination corresponding to the current optimal discrete solution is output, and the optimal material combination is used to design the target matching network.

[0006] In one embodiment, prior to the step of obtaining the material index space, the method further includes: Obtain the current bill of materials, which includes each component type and its corresponding nominal values. Each component type includes capacitors, inductors, and resistors. After converting the nominal values ​​of each component type to standard units, arrange them in ascending order to obtain the sorting result; Based on the arrangement result, assign corresponding integer indices to each of the transformed nominal values ​​to establish a material index space.

[0007] In one embodiment, the step of determining the target impedance and target matching configuration based on the current matching requirements includes: Based on the current matching requirements, the measured data is obtained, including several measured impedance values ​​and corresponding performance index values. A discrete data point set is constructed using the measured impedance values ​​and the corresponding performance index values. The discrete data point set is interpolated and fitted using radial basis functions to generate a continuous performance surface; Search for the maximum value of the performance index on the performance surface, and take the impedance value corresponding to the maximum value of the performance index as the target impedance; In response to the user's selection of candidate matching configurations, at least one target matching configuration is determined.

[0008] In one embodiment, the step of constructing the current search space based on the target matching configuration and the material index space includes: The dimension of the continuous parameter space is determined based on the target matching configuration, and the dimension is equal to the number of elements in the target matching configuration; The minimum and maximum nominal values ​​of each type of element in the material index space are used as the boundaries of the continuous parameter space to construct the current search space. Each continuous value in the current search space is used to represent a combination of different element nominal values.

[0009] In one embodiment, the step of running a differential evolution algorithm in the current search space to search for the optimal combination of nominal values ​​with the goal of minimizing impedance error includes: Based on the current matching requirements, obtain actual test data, and determine the initial individual based on the actual test data; An initial population is constructed based on the initial individuals, and a differential evolution algorithm is run in the current search space based on the initial population to search for the optimal combination of nominal values ​​with the goal of minimizing the impedance error. Wherein, the impedance error is the complex Euclidean distance between the input impedance of different individuals and the target impedance obtained by running the differential evolution algorithm, and each individual corresponds to a different continuous value in the current search space.

[0010] In one embodiment, the step of mapping the optimal nominal value combination to an integer index in the material index space and enumerating all candidate material combinations in the neighborhood of the integer index includes: For each component nominal value in the optimal nominal value combination, the corresponding integer index is matched in the material index space according to the principle of the closest nominal value. With each integer index as the center, construct the search interval for each index according to the preset neighborhood radius; Take the Cartesian product of each of the index search intervals to obtain all candidate material combinations, and each candidate material combination corresponds to a component nominal value combination.

[0011] In one embodiment, the candidate material combination corresponds to a component nominal value combination, and the impedance error calculation step includes: Based on the measured data, several measured impedance values ​​and corresponding theoretical calculation values ​​are determined, and fitting parameters are obtained based on each of the measured impedance values ​​and corresponding theoretical calculation values. Based on the target matching configuration, calculate the theoretical input impedance corresponding to each combination of nominal values ​​of the components; The theoretical input impedance is calibrated using the fitting parameters to obtain the input impedance; The complex Euclidean distance between the input impedance and the target impedance is calculated as the impedance error.

[0012] Furthermore, to achieve the above objectives, this application also proposes a radio frequency matching network design apparatus based on a material library constraint, the apparatus comprising: The matching preparation module is used to obtain the material index space and determine the target impedance and target matching configuration based on the current matching requirements. The outer search module is used to construct the current search space based on the target matching configuration and the material index space, and run the differential evolution algorithm in the current search space to search for the optimal nominal value combination with the goal of minimizing impedance error. The impedance error is the distance between the input impedance corresponding to the continuous value in the current search space and the target impedance. The inner search module is used to map the optimal nominal value combination to an integer index in the material index space, and enumerate all candidate material combinations in the neighborhood of the integer index; The inner search module is also used to determine the impedance error corresponding to each candidate material combination, and select the candidate material combination with the smallest impedance error as the current optimal discrete solution. The result output module is used to convert the current optimal discrete solution back to a continuous value and then return to the step of running the differential evolution algorithm until the termination condition is met, and then output the optimal material combination corresponding to the current optimal discrete solution.

[0013] Furthermore, to achieve the above objectives, this application also proposes a library-based RF matching network design device, the device comprising: a memory, a processor, and a library-based RF matching network design program stored in the memory and executable on the processor, the library-based RF matching network design program being configured to implement the steps of the library-based RF matching network design method described above.

[0014] Furthermore, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, storing a library-based RF matching network design program. When the library-based RF matching network design program is executed by a processor, it implements the steps of the library-based RF matching network design method described above.

[0015] This application discloses a method for designing radio frequency matching networks based on material library constraints. The method includes: obtaining a material index space and determining the target impedance and target matching configuration according to the current matching requirements; constructing a current search space based on the target matching configuration and the material index space, and running a differential evolution algorithm in the current search space with the goal of minimizing impedance error, searching for the optimal nominal value combination, where the impedance error is the distance between the input impedance corresponding to the continuous value in the current search space and the target impedance; mapping the optimal nominal value combination to an integer index in the material index space, and enumerating all candidate material combinations in the neighborhood of the integer index; determining the impedance error corresponding to each candidate material combination, and selecting the candidate material combination with the smallest impedance error as the current optimal discrete solution; converting the current optimal discrete solution back to continuous values ​​and returning to the step of running the differential evolution algorithm until the termination condition is met, and outputting the optimal material combination corresponding to the current optimal discrete solution, which is used to design the target matching network.

[0016] Because this application employs the technique of obtaining the material index space and mapping the optimal nominal value combination to an integer index in the material index space, and enumerating candidate material combinations in the neighborhood of the integer index, the optimization process is directly constrained to the actual procureable material library. The output optimal material combination can be directly used to design the target matching network without manual approximation, thereby eliminating the disconnect between design and material selection. At the same time, because it adopts the technique of constructing the current search space based on the target matching configuration and the material index space, running the differential evolution algorithm in the search space to obtain the optimal continuous solution, and converting the current optimal discrete solution back to a continuous value before returning to the iteration, the outer continuous search and the inner discrete search form a closed-loop feedback, and the optimization result continuously approaches the global optimum. Compared with manual traversal or unidirectional optimization methods, this significantly improves the efficiency and accuracy of matching network design. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the first embodiment of the radio frequency matching network design method based on material library constraints of this application. Figure 2 This is a flowchart illustrating the second embodiment of the radio frequency matching network design method based on material library constraints in this application. Figure 3 This is a flowchart illustrating the third embodiment of the radio frequency matching network design method based on material library constraints in this application. Figure 4 This is a schematic diagram illustrating the entire process of the radio frequency matching network design method based on material library constraints in this application; Figure 5 This is a schematic diagram of the module structure of the radio frequency matching network design device based on material library constraints in this application; Figure 6 This is a schematic diagram of the structure of the radio frequency matching network design device based on material library constraints in this application.

[0020] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0022] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0023] This application provides a method for designing radio frequency matching networks based on material library constraints, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the radio frequency matching network design method based on material library constraints according to this application. In this embodiment, the method includes steps S10 to S50: Step S10: Obtain the material index space and determine the target impedance and target matching configuration based on the current matching requirements.

[0024] It should be noted that the method of this embodiment can be applied to the scenario of matching network design during the debugging of radio frequency circuits. The execution subject of this embodiment can be a computing electronic device with data processing, numerical calculation and other functions, such as a personal computer, workstation or server. Here, a matching network design device (referred to as "device") is used as an example to describe this embodiment and the following embodiments.

[0025] It should be understood that the material index space is an integer index space obtained by structuring a procureable material library. This material library can contain the nominal values ​​and packaging information of components such as capacitors, inductors, and resistors. The equipment constructs the material index space by arranging the nominal values ​​of each component in ascending order and assigning them unique integer indices, thus establishing a bidirectional mapping from integer indices to nominal values.

[0026] Furthermore, to illustrate in detail how the material index space is constructed, steps S01~S03 are included before step S10: Step S01: Obtain the current bill of materials, which includes each component type and its corresponding nominal values. Each component type includes capacitors, inductors, and resistors.

[0027] It should be noted that the current bill of materials (BOM) is a list of electronic components that engineers can actually purchase. This BOM can be imported from an enterprise resource planning system, a materials management database, or a spreadsheet file.

[0028] The bill of materials (BOM) can include three basic passive component types: capacitors, inductors, and resistors, with each type corresponding to multiple nominal values. For example, nominal values ​​for capacitors can include 1.0pF, 1.2pF, 1.5pF, 1.8pF, 2.2pF, etc.; nominal values ​​for inductors can include 10nH, 12nH, 15nH, 18nH, 22nH, etc.; and nominal values ​​for resistors can include 10... ,twenty two 33 47 50 wait.

[0029] In addition, the bill of materials can also include the packaging information of each component (such as 0402, 0603, 0805, etc.) to facilitate subsequent actual procurement and soldering.

[0030] In practical implementation, the device can read the bill of materials file from a specified path, parse the component type field, nominal value field and package information field in the file, classify and store them according to component type, and provide basic data for the subsequent steps to establish a material index space.

[0031] Step S02: Convert the nominal values ​​of each component type to standard units and arrange them in ascending order to obtain the arrangement result.

[0032] It should be understood that bills of materials from different sources may use different units to represent the same type of component. For example, capacitors may be measured in farads (F), nanofarads (nF), or picofarads (pF), while inductors may be measured in henries (H), millihenries (mH), or microhenries (μH). The unit is either H or Nahen (nH).

[0033] To establish a unified index space, devices can convert all nominal values ​​to standard units: for example, capacitors can be uniformly converted to picofarads (pF), inductors to nanohenries (nH), and resistors to ohms (pF). ).

[0034] Understandably, after the conversion, the device can also sort the nominal values ​​of each component type in ascending order. For example, after unit conversion and sorting, the nominal values ​​of capacitors result in the sequence: 1.0pF, 1.2pF, 1.5pF, 1.8pF, 2.2pF, 2.7pF, 3.3pF, 3.9pF, 4.7pF, 5.6pF…; after conversion and sorting, the nominal values ​​of inductors result in the sequence: 10nH, 12nH, 15nH, 18nH, 22nH, 27nH, 33nH, 39nH, 47nH, 56nH…

[0035] In the actual implementation, the device can traverse all nominal values ​​for each component type, call the unit conversion function to convert non-standard units to standard units, and then call the sorting algorithm to sort the converted values ​​in ascending order to obtain an ordered list of nominal values.

[0036] Step S03: Based on the arrangement result, assign corresponding integer indices to each of the converted nominal values ​​to establish a material index space.

[0037] It should be understood that the device can assign a unique integer index to each of the permuted nominal values. The index can start from 0 and increment in ascending order.

[0038] For example, in the capacitor library, index 0 corresponds to 1.0pF, index 1 to 1.2pF, index 2 to 1.5pF, index 3 to 1.8pF, index 4 to 2.2pF, and so on. In the inductor library, index 0 corresponds to 10nH, index 1 to 12nH, index 2 to 15nH, index 3 to 18nH, index 4 to 22nH, and so on.

[0039] Through the above allocation, the equipment can establish a bidirectional mapping relationship from integer indices to nominal values. The material index space can be a discrete space composed of the above integer indices, where each integer index uniquely corresponds to a real, procureable component.

[0040] In its implementation, the equipment first obtains the current bill of materials, converts the nominal values ​​of each component type into standard units, and then arranges them in ascending order. Based on this arrangement, it assigns corresponding integer indices to each converted nominal value to establish a material index space. This lays the foundation for subsequently mapping continuous optimization results to actual material combinations.

[0041] It should also be noted that the current matching requirements may include the operating frequency, load impedance, source impedance, and expected performance metrics of the RF matching network to be designed (such as maximum power transfer, minimum error vector magnitude, or highest receiver sensitivity).

[0042] For example, in a 2.4GHz Wi-Fi RF front-end design, the operating frequency is 2.4GHz, and the load impedance is the antenna impedance (e.g., 50Ω). The source impedance is the output impedance of the power amplifier (e.g., 10 + j5). The goal is to achieve maximum power transfer.

[0043] The target impedance can be a complex impedance value determined based on the current matching requirements, serving as the objective for the matching network design. When maximum power transfer is desired, according to the maximum power transfer theorem, the target impedance can be set as the conjugate of the source impedance.

[0044] For example, if the source impedance is 10 + j5 The target impedance is then 10-j5. When it is desired to optimize other performance indicators (such as minimum error vector magnitude or maximum receiver sensitivity), the target impedance can be determined by back-calculation from measured data.

[0045] Furthermore, to illustrate in detail how to determine the target impedance required for the target matching network based on measured data, step S10 also includes: steps S101~S105: Step S101: Obtain measured data according to the current matching requirements. The measured data includes several measured impedance values ​​and corresponding performance index values.

[0046] It should be noted that when engineers cannot directly know the source impedance, or when the performance target they want to optimize is not maximum power delivery (e.g., optimizing error vector amplitude or receiver sensitivity), they can infer the target impedance from measured data. Measured data can be pairs of data obtained by engineers in advance through testing on an actual circuit board using a known combination of components.

[0047] Specifically, engineers can randomly select several combinations of materials (e.g., 5 to 10 combinations), solder these combinations onto a circuit board, use a network analyzer to measure the complex impedance at the port, and simultaneously use a spectrum analyzer or comprehensive tester to measure performance indicators (such as output power, error vector amplitude, or receiver sensitivity) under these conditions. Each test will yield a data point containing the real part of the impedance, the imaginary part of the impedance, and the performance indicator value.

[0048] For example, the engineer tested five combinations of materials and obtained the following measured data: the impedance of the first combination was 45 + j5. The power is 18.2dBm; the impedance of the second group is 48+j2. The power is 19.1 dBm; the impedance of the third group is 52-j3. The power is 18.7dBm; the impedance of the fourth group is 38+j8. The power is 17.5dBm; the impedance of the fifth group is 50+j0. The power is 18.9 dBm. Therefore, the above measured data can be used as input for subsequent interpolation and fitting.

[0049] In practice, the device can receive measured data input by engineers through a user interface, or read measured data from a specified data file, and parse and store it according to the format of the real part of impedance, the imaginary part of impedance, and performance index values.

[0050] Step S102: Construct a discrete data point set using each of the measured impedance values ​​and the corresponding performance index values.

[0051] It should be understood that the device can treat each set of measured data obtained above as a discrete point in three-dimensional space. The three coordinate axes of this three-dimensional space are: the real part of the impedance (R-axis), the imaginary part of the impedance (X-axis), and the performance index value (P-axis).

[0052] For example, the above five sets of measured data can be constructed into a discrete data point set: point A has coordinates (45, 5, 18.2), point B has coordinates (48, 2, 19.1), point C has coordinates (52, -3, 18.7), point D has coordinates (38, 8, 17.5), and point E has coordinates (50, 0, 18.9). These discrete points are distributed at different locations on the impedance plane, and their heights (performance index values) are different.

[0053] Understandably, due to the limited number of measured data points (usually a few to a dozen) and their uneven distribution, it is impossible to directly determine the location of the globally optimal performance point. Therefore, interpolation can be used to extend these discrete points into a continuous surface.

[0054] Step S103: Use radial basis functions to interpolate and fit the discrete data point set to generate a continuous performance surface.

[0055] It should be noted that the Radial Basis Function (RBF) is a method suitable for interpolation of discrete points with non-uniform distribution. RBF constructs a local influence region centered on each discrete data point using a radially symmetric basis function (such as a Gaussian function), and then uses the weighted superposition of all basis functions as the predicted value on the entire surface.

[0056] In practical implementation, the device can select a Gaussian kernel as the radial basis function and set a smoothing parameter (such as smooth=1.0) to prevent overfitting, make the fitted surface smoother, and avoid drastic fluctuations in the surface due to measurement errors of individual data points.

[0057] For example, taking the aforementioned five discrete points as nodes, the device can use these five discrete points as nodes to generate a continuous performance surface covering the entire impedance plane through radial basis function interpolation. Each point on this performance surface (corresponding to an impedance value) then corresponds to a predicted performance index value. Because the performance surface is continuous, the device can evaluate the performance of areas that have not been measured.

[0058] In practical implementation, the device can call the radial basis function interpolation function in a scientific computing library (such as scipy.interpolate.Rbf), passing in the coordinates and values ​​of discrete data points, and specifying the Gaussian kernel function and smoothing parameters to obtain a callable interpolation function object. This interpolation function object can then accept the coordinates of any impedance point as input and return the predicted performance index value.

[0059] Step S104: Search for the maximum value of the performance index on the performance surface, and take the impedance value corresponding to the maximum value of the performance index as the target impedance.

[0060] It should be understood that after generating a continuous performance surface, the device can find the maximum value of the performance metric (i.e., the optimal performance point) on this surface. Since the performance surface can contain multiple local maxima, the device employs a two-step search strategy to ensure that the global maximum is found: The first step is a coarse grid search. The device first generates a dense grid (e.g., 200*200 grid points) within the boundary of the measured data, calls the radial basis function interpolation function to calculate the prediction performance value of each grid point, and finds the grid point with the largest prediction performance value as the initial guess location.

[0061] The second step is precise peak finding. The device then uses the initially guessed location as a starting point and employs a gradient optimization algorithm (such as the L-BFGS-B algorithm) for iterative optimization. This gradient optimization algorithm can utilize the gradient information of the performance surface to gradually move towards the local maximum, eventually converging to the precise coordinates of the maximum point.

[0062] For example, after the above search, the device might find its maximum performance at the impedance coordinates (47.2, 1.3), with a predicted power of 19.6 dBm. The device would then use this impedance value (47.2 + j1.3). This is used as the target impedance for subsequent optimization processes.

[0063] In addition, the equipment is equipped with a result verification mechanism: if the maximum value obtained by optimization is significantly beyond the reasonable range (for example, the power value exceeds the theoretical maximum possible value), the optimization result can be abandoned and the maximum value point in the original measured data can be used as the target impedance to avoid abnormal values ​​caused by extrapolation.

[0064] Step S105: In response to the user's selection operation on the candidate matching configuration, determine at least one target matching configuration.

[0065] It should be noted that the device can receive engineers' selections of matching network topologies through a user interface. Engineers can then choose one or more topologies for optimization based on application requirements.

[0066] Target matching configuration refers to the topology of the matching network, including L-type, At least one of L-type or T-type. Furthermore, based on different matching network design requirements, this target matching configuration can be extended to, for example, multi-level LC networks or matching structures with resistors and capacitors, etc., and this embodiment does not limit this. The target matching configuration provided by the device below includes L-type, Let's take the type and T as examples to illustrate.

[0067] Among them, the L-type matching network consists of two components, has a simple structure, and is suitable for scenarios with relatively small impedance transformation. The type-1 matching network consists of three components (two parallel capacitors and one series inductor) and is suitable for wideband matching; the type-2 matching network consists of three components (two series inductors and one parallel capacitor) and is suitable for scenarios requiring high Q values.

[0068] For example, engineers can check "L-shaped" on the interface. The device offers three topologies: L-type, T-type, and L-type. In response to this selection operation, the device will switch to L-type, T-type, L ... Both the Type 1 and Type 2 configurations are determined as the target matching configurations. When the engineer only selects "..." When the "type" is used, the equipment will only The type was determined to be the target matching configuration.

[0069] It should also be noted that when multiple target matching configurations are determined, the device can independently execute the optimization process for each target matching configuration in subsequent steps. Finally, the optimal results of each configuration are combined and sorted by impedance error for output, allowing engineers to select. For ease of explanation, this embodiment will be described below using the example of an engineer selecting only one type of target matching configuration.

[0070] This implementation addresses the issue that in existing debugging processes, engineers can only view multiple sets of impedance-performance data as discrete points, making it difficult to determine the globally optimal region. This implementation uses radial basis function interpolation to automatically mark the impedance target corresponding to the optimal performance point on the fitted surface. This not only provides a clear direction for matching optimization but also offers engineers a visual debugging reference. Furthermore, this implementation allows engineers to independently select the target matching network configuration, thus meeting diverse matching network design needs.

[0071] Step S20: Construct the current search space based on the target matching configuration and the material index space, and run the differential evolution algorithm in the current search space with the goal of minimizing impedance error, and search for the optimal nominal value combination. The impedance error is the distance between the input impedance corresponding to the continuous value in the current search space and the target impedance.

[0072] It should be noted that the current search space can be constructed in a continuous parameter space. The device determines the dimension of this continuous parameter space based on the target matching configuration, and this dimension is equal to the number of elements in the target matching configuration. For example, an L-shaped configuration contains 2 elements, and the continuous parameter space is a 2-dimensional plane; The type or T-type configuration contains 3 elements, and the continuous parameter space is a 3-dimensional cube.

[0073] Furthermore, the equipment can use the minimum and maximum nominal values ​​of each type of component in the material index space as the boundaries of this continuous parameter space to construct the current search space. Each point in the current search space corresponds to a continuous value, and each continuous value is used to represent different combinations of component nominal values.

[0074] It should be understood that differential evolution is a population-based global optimization algorithm. The device can initialize a population in the current search space, where each individual corresponds to a continuous value vector, representing a set of component nominal values.

[0075] Specifically, the device can run a differential evolution algorithm with the goal of minimizing impedance error. Through iterative execution of mutation, crossover, and selection operations, the population gradually converges towards a region with smaller impedance error. Here, impedance error refers to the distance between the input impedance corresponding to the continuous value in the current search space and the target impedance, which is measured using complex Euclidean distance.

[0076] During the execution of the differential evolution algorithm, the device can calculate the input impedance based on the combination of nominal values ​​of the components represented by each individual component, combined with the target matching configuration, operating frequency, and load impedance, following a recursive approach from the load end to the source end: traversing each component in the matching network, calculating its complex impedance according to the component type (resistor, inductor, capacitor), and progressively recursively calculating based on the connection method of the components in the topology (series or parallel), ultimately obtaining the input impedance seen from the source end. Then, combining the input impedance and the target impedance, the impedance error is calculated.

[0077] Step S30: Map the optimal nominal value combination to an integer index in the material index space, and enumerate all candidate material combinations in the neighborhood of the integer index.

[0078] It should be noted that after the differential evolution algorithm finishes running, the device can obtain the optimal nominal value combination. Since this optimal nominal value combination is represented as a continuous value vector, the values ​​of each component may not have an exact corresponding nominal value in the material library. For example, for an L-type matching network, the device may output the optimal nominal value combination as [1.67pF, 11.3nH], where 1.67pF may not have an exact corresponding capacitance value in the material library.

[0079] The equipment can then map this optimal nominal value combination to the material index space. Specifically, for each component nominal value component in the optimal nominal value combination, the equipment matches the corresponding integer index in the material index space according to the principle of the closest nominal value.

[0080] For example, the capacitance value of 1.67pF differs from 1.5pF (index 2) in the material library by 0.17pF and from 1.8pF (index 3) by 0.13pF, so index 3 (1.8pF) is chosen as the mapping result; the inductance value of 11.3nH differs from 10nH (index 0) in the material library by 1.3nH and from 12nH (index 1) by 0.7nH, so index 1 (12nH) ​​is chosen as the mapping result. Thus, the device can obtain an integer index vector, such as [3,1], which uniquely corresponds to a set of procurable material combinations (1.8pF and 12nH).

[0081] It should be understood that after obtaining the integer index corresponding to each component, the device can construct the search interval for each index with each integer index as the center and according to the preset neighborhood radius.

[0082] The neighborhood radius can be set or dynamically adjusted according to the sparsity of the material library. For example, if the neighborhood radius is set to r=2, the index search interval for capacitors is [1,2,3,4,5] (corresponding to 1.2pF, 1.5pF, 1.8pF, 2.2pF, 2.7pF), and the index search interval for inductors is [0,1,2,3] (corresponding to 10nH, 12nH, 15nH, 18nH).

[0083] Next, the device can take the Cartesian product of each index search interval to obtain all candidate material combinations. In this example, the number of candidate material combinations is 5 × 4 = 20.

[0084] Finally, the equipment can read the corresponding actual nominal value from the material library for each candidate material combination, preparing for impedance error calculation. Therefore, all the above candidate material combinations are considered procureable and do not require subsequent manual approximation replacement.

[0085] Step S40: Determine the impedance error corresponding to each candidate material combination, and select the candidate material combination with the smallest impedance error as the current optimal discrete solution.

[0086] It should be noted that the device can calculate the corresponding input impedance for each candidate material combination obtained through the aforementioned enumeration. When calculating the input impedance, the device can calculate the input impedance based on the actual nominal value corresponding to the candidate material combination, combined with the target matching configuration, operating frequency, and load impedance, following a recursive approach from the load end to the source end. This calculation method is the same as the steps for calculating individual input impedances described above, and will not be repeated here.

[0087] It should be understood that after obtaining the input impedance of each candidate material combination, the equipment can also calculate the complex Euclidean distance between each input impedance and the target impedance, which is used as the impedance error corresponding to that candidate material combination.

[0088] Finally, the equipment iterates through all candidate material combinations, compares the impedance errors of each combination, and selects the one with the smallest impedance error as the current optimal discrete solution. This current optimal discrete solution can be represented as an integer index vector (e.g., [3,1]), which uniquely corresponds to a set of material combinations that can be actually procured.

[0089] Step S50: After converting the current optimal discrete solution back to a continuous value, return to the step of running the differential evolution algorithm until the termination condition is met, and output the optimal material combination corresponding to the current optimal discrete solution. The optimal material combination is used to design the target matching network.

[0090] It should be noted that the device can convert the current optimal discrete solution back to continuous values, that is, convert the integer index vector of the current optimal discrete solution into the corresponding nominal value vector to obtain individuals in continuous value form. For example, the integer index vector [3,1] can be converted into the nominal value vector [1.8pF,12nH].

[0091] Next, the device can inject the individual with the continuous value into the current population of the differential evolution algorithm, replacing the individual with the worst fitness in the population, while keeping the population size unchanged. Subsequently, the device returns to the aforementioned step S20 and re-executes the differential evolution algorithm and its subsequent steps in the current search space, forming a closed-loop iteration.

[0092] It should be understood that the device can repeatedly execute steps S20 to S50 until a termination condition is met. This termination condition may include: the impedance error of the current optimal discrete solution is less than a preset threshold (e.g., 0.5). If the optimal impedance error is not improved in multiple iterations, or if the preset maximum number of iterations (e.g., 100 iterations) is reached, the device stops iterating and prepares to output the optimal material combination.

[0093] Understandably, when outputting the optimal material combination, the equipment can support the following output formats: BOM (CSV or Excel format), component value table, or circuit design diagram. The BOM format includes component reference designators, component types, nominal values, package information, material index, and expected impedance tolerance, facilitating direct use by engineers for material procurement and circuit soldering. The component value table format only contains component types and nominal values, suitable for quick viewing. The circuit design diagram format labels the nominal values ​​of each component on the matching network topology diagram, suitable for design document archiving.

[0094] Furthermore, the optimal material combination output by the equipment can be more than just one: based on user-preset output quantity requirements (e.g., the three combinations with the smallest output impedance error), the equipment can select several combinations with the smallest impedance error from multiple discrete solutions that meet the termination condition recorded during the iteration process as the output results. These output results are sorted in ascending order of impedance error, allowing engineers to select according to actual needs (such as cost, inventory, size, etc.). For example, the equipment can output three material combinations: the first combination has an impedance error of 0.36... The impedance error of the second group is 0.42. The impedance error of the third group is 0.51. ).

[0095] If the user selected multiple target matching configurations in the aforementioned step S105 (e.g., simultaneously selected L-type, ...), (Both L-type and T-type) The equipment can independently execute a complete optimization process for each target matching configuration, thereby obtaining the optimal material combination and its impedance error under that matching configuration. After merging the optimal results of all matching configurations, the equipment can also globally sort them according to impedance error from smallest to largest, and output a sorted list of results. For example, the impedance error of the optimal combination for L-type is 0.35Ω. The impedance error of the optimal combination of type L is 0.42Ω, and the impedance error of the optimal combination of type T is 0.51Ω. Therefore, the output order is L-type, Type, T-type.

[0096] Furthermore, when the user-preset output quantity requirement (e.g., requiring five material combinations) cannot be met in the current optimization results, the device can also employ a substitute output strategy. Specifically, the device first outputs all material combinations that meet the termination condition (e.g., impedance error is less than a preset threshold). If the number of material combinations meeting the condition is still insufficient to meet the user's requirement, the device selects the optimal material combination from the next priority target matching configuration as a substitute output. The order of this next priority level is preset by the user or determined according to impedance error sorting, for example, set to L-type > Type > T-type.

[0097] For example, the user requests five material combinations, but currently only three meet the termination condition (from L-type and...). If the optimal combination of all matching configurations is not sufficient to meet the user's requirements, the equipment can output all the material combinations obtained and prompt the user to adjust the preset threshold or increase the types of components in the material library.

[0098] In its implementation, the device stops iterating when the termination condition is met, and finally outputs a list of optimal material combinations. Each material combination in this list can be directly used to design the target matching network without the need for manual approximation. The corresponding data for each material combination can include component type, nominal value, package information, material index, and expected impedance error. For example: capacitor C1 = 1.8pF (0402 package, index 3), inductor L1 = 12nH (0402 package, index 1), expected impedance error 0.36Ω.

[0099] This embodiment utilizes techniques such as acquiring a material index space, mapping the optimal nominal value combination to an integer index within that space, and enumerating candidate material combinations within the neighborhood of that integer index. This allows the optimization process to be directly constrained to the range of actually procurable materials. The output optimal material combination can be directly used to design the target matching network without manual approximation, thus eliminating the disconnect between design and material selection. Furthermore, by converting the current optimal discrete solution back to a continuous value and then iterating, a closed-loop feedback loop is formed between the outer continuous search and the inner discrete search. The optimization result continuously approaches the global optimum, significantly improving the efficiency and accuracy of matching network design compared to manual traversal or unidirectional optimization methods.

[0100] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 , Figure 2 This is a flowchart illustrating the second embodiment of the radio frequency matching network design method based on material library constraints in this application.

[0101] In this embodiment, to specifically illustrate how to construct the current search space and perform the outer search in a two-layer search based on the current search space, step S20 specifically includes: steps S201~S204: Step S201: Determine the dimension of the continuous parameter space according to the target matching configuration, wherein the dimension is equal to the number of elements in the target matching configuration.

[0102] It should be noted that the continuous parameter space can be the search space used to run the differential evolution algorithm, and its dimension is determined by the number of elements in the target matching configuration. Different matching network topologies contain different numbers of elements: the L-type matching network consists of two elements, therefore its continuous parameter space is two-dimensional. The type-1 matching network consists of three components (two parallel capacitors and one series inductor), and its continuous parameter space is three-dimensional; the type-2 matching network consists of three components (two series inductors and one parallel capacitor), and its continuous parameter space is also three-dimensional.

[0103] For example, if the engineer chooses an L-shaped matching network as the target matching configuration, the device determines the dimension of the continuous parameter space to be 2; if the engineer chooses... If the matching network is of type 1 or type 2, then the device determines the dimension of the continuous parameter space to be 3.

[0104] Furthermore, when engineers select multiple target matching configurations, the device can independently execute the steps of this embodiment for each configuration, with each configuration corresponding to a continuous parameter space of its own dimension.

[0105] Step S202: Using the minimum and maximum nominal values ​​of each type of element in the material index space as the boundaries of the continuous parameter space, construct the current search space. Each continuous value in the current search space is used to represent a combination of different element nominal values.

[0106] It should be noted that after determining the dimensions of the continuous parameter space, the equipment can determine the value range of each dimension in the continuous parameter space, and use the minimum nominal value and maximum nominal value of each type of element in the material index space as the boundary.

[0107] Specifically, the device traverses the material index space. For capacitor-type components, it searches for all nominal capacitance values, takes the minimum value as the lower bound of the capacitance dimension, and takes the maximum value as the upper bound of the capacitance dimension. The same processing is performed for inductor-type and resistor-type dimensions.

[0108] For example, in the material index space, the minimum nominal value of a capacitor is 0.5pF and the maximum nominal value is 100pF, so the value range for the capacitor dimension is [0.5, 100]; the minimum nominal value of an inductor is 1nH and the maximum nominal value is 50nH, so the value range for the inductance dimension is [1, 50]; the minimum nominal value of a resistor is 10... The maximum nominal value is 1000 The resistance dimension can be set to [10, 1000].

[0109] For an L-type matching network (e.g., a capacitor and an inductor), the constructed current search space can be a two-dimensional rectangular region: x-coordinate C [0.5, 100], ordinate L [1,50].

[0110] for The matching network (e.g., capacitor C1, inductor L, capacitor C2) constructs a three-dimensional cube in the current search space: C1 [0.5,100], L [1,50], C2 [0.5,100].

[0111] It should also be noted that each point in the current search space, i.e., a D-dimensional continuous value vector, represents a set of continuous combinations of component nominal values. For example, the point (1.5, 12) in the two-dimensional space represents the combination of a capacitor of 1.5pF and an inductor of 12nH.

[0112] Step S203: Obtain actual test data according to the current matching requirements, and determine the initial individual based on the actual test data.

[0113] It should be noted that after constructing the current search space, the device can initialize the population of the differential evolution algorithm. In this embodiment, an initialization strategy based on measured data can be adopted.

[0114] It should be understood that the equipment can acquire measured data based on current matching requirements. Measured data consists of pairs of data obtained by engineers on actual circuit boards using a known combination of components. These pairs include several sets of measured impedance values ​​and corresponding performance index values. Each set of measured data corresponds to a set of actual components used, which consists of specific component nominal values.

[0115] Specifically, the equipment can convert the material combination corresponding to each set of measured data into a continuous value vector, and then use the continuous value vectors corresponding to all measured data as candidates for the initial individual.

[0116] For example, if an engineer uses a combination of materials in a test: a capacitor of 1.5pF and an inductor of 12nH, the device will convert that combination into a continuous value vector [1.5, 12].

[0117] It should be noted that since the measured data comes from material combinations that performed well in actual tests, the initial individuals determined based on the measured data have high fitness (i.e., small impedance error) and can guide the algorithm to converge quickly to the optimal region.

[0118] Step S204: Construct an initial population based on the initial individuals, and run a differential evolution algorithm in the current search space based on the initial population, with the goal of minimizing the impedance error, to search for the optimal combination of nominal values.

[0119] It should be understood that the device can first construct an initial population that meets a preset population size. This initial population can include initial individuals determined based on measured data, as well as individuals randomly generated within the current search space. If the number of initial individuals provided by the measured data is less than the preset population size, the device can randomly generate the remaining number of individuals within the boundaries of the current search space in a uniform distribution.

[0120] Understandably, once the initial population is constructed, the device can run the differential evolution algorithm in the current search space based on this initial population. The differential evolution algorithm is a population-based global optimization algorithm that can gradually converge the population towards the optimal solution through iterative execution of three operations: mutation, crossover, and selection.

[0121] It should be noted that the mutation operation can be performed as follows: for each target individual in the population, the device randomly selects three other distinct individuals and generates a mutation vector. The formula for calculating the mutation vector is: Mutation vector = Individual A + Scaling factor * (Individual B - Individual C). Here, the scaling factor is a preset constant, typically between 0.5 and 1, used to control the amplification degree of the difference.

[0122] The crossover operation can be performed as follows: the device mixes the target individual with the mutation vector to generate an experimental vector. For each dimension, the device generates a random number. If the random number is less than a preset crossover probability (e.g., set to 0.9), the value for that dimension is taken from the mutation vector; otherwise, the value for that dimension is taken from the target individual. The crossover operation helps increase population diversity and prevents the algorithm from getting trapped in local optima too early.

[0123] The selection operation can be as follows: the device calculates the fitness of the target individual and the trial vector separately. The fitness function is the impedance error, which is the complex Euclidean distance between the individual's input impedance and the target impedance. If the impedance error of the trial vector is less than or equal to the impedance error of the target individual, the trial vector enters the next generation of the population; otherwise, the target individual is retained in the next generation. The selection operation helps ensure that the population as a whole evolves in the direction of smaller impedance errors.

[0124] In practical implementation, the device can repeatedly perform the above mutation, crossover, and selection operations until a preset maximum number of iterations (e.g., 200 generations) is reached or the impedance error is less than a preset threshold (e.g., 0.5). After the algorithm finishes running, the device outputs the individual with the smallest impedance error in the current population as the optimal nominal value combination. This optimal nominal value combination is represented in the form of a continuous value vector, for example, [1.67pF, 11.3nH].

[0125] Furthermore, to illustrate in detail how to perform an inner search based on an outer search, step S30 specifically includes: steps S301~S303: Step S301: For each component nominal value component in the optimal nominal value combination, match the corresponding integer index in the material index space according to the principle of closest nominal value.

[0126] It should be noted that, since the optimal combination of nominal values ​​obtained above is represented in the form of a continuous value vector, the values ​​of each component may not have an exact corresponding nominal value in the material library. Therefore, the device can map this continuous value vector to an integer index vector to perform a discrete search in the material index space.

[0127] Specifically, the equipment can find the nominal value closest to the value of each component in the optimal nominal value combination in the material index space and take the integer index corresponding to that nominal value.

[0128] For example, the optimal nominal value combination obtained by the device is [1.67pF, 11.3nH]. For the capacitance component of 1.67pF, the device searches in the capacitance material index space: index 2 corresponds to 1.5pF (difference of 0.17pF), and index 3 corresponds to 1.8pF (difference of 0.13pF). Therefore, the 1.8pF corresponding to index 3 is closer, and the device selects index 3. For the inductance component of 11.3nH, the device searches in the inductance material index space: index 0 corresponds to 10nH (difference of 1.3nH), and index 1 corresponds to 12nH (difference of 0.7nH). Therefore, the 12nH corresponding to index 1 is closer, and the device selects index 1. The mapping result is an integer index vector [3, 1].

[0129] It should also be noted that when there are two nominal values ​​that are equidistant from the target value, the device can select according to preset rules, such as prioritizing the larger nominal value or prioritizing the smaller nominal value.

[0130] Step S302: Using each integer index as the center, construct the search interval for each index according to the preset neighborhood radius.

[0131] It should be noted that obtaining the integer index corresponding to each component... Then, the device can index each integer. A search index is constructed centered on this central index. This search index is centered on this central index. Subtract neighborhood radius To the central index Add neighborhood radius It consists of all integer indices within the range. That is, it is represented as [ , ]. Among them, the neighborhood radius This is a preset value that can be adjusted based on the sparsity of the material library. The initial value can be set to 2, indicating a search center. There are 2 indexes on each side, for a total of 5 indexes.

[0132] Understandably, if the nominal values ​​of this type of component in the material library are relatively sparse (i.e., the nominal values ​​corresponding to adjacent indices differ significantly), the device can appropriately increase the radius to ensure that the search covers sufficient candidate combinations; if the nominal values ​​of this type of component in the material library are relatively dense, the device can appropriately decrease the radius to improve search efficiency.

[0133] It should also be noted that the device can perform boundary processing on the index search range, eliminating values ​​that exceed the valid index range, thereby ensuring that the valid range of the index starts from 0 and ends at the maximum index value of this type of element.

[0134] Step S303: Take the Cartesian product of each index search interval to obtain all candidate material combinations, and the candidate material combinations correspond to component nominal value combinations.

[0135] It should be noted that the device can take the Cartesian product of the search intervals of each component index to generate all possible combinations of integer indices. Each combination of integer indices uniquely corresponds to a set of component nominal value combinations.

[0136] For example, consider an L-shaped matching network with two component locations (denoted as Component 1 and Component 2). Each component location can independently select its component type; for instance, Component 1 can be either a capacitor or an inductor, and Component 2 can also be either a capacitor or an inductor. If the index search range for Component 1 contains n indices, each corresponding to a different nominal value for either capacitor or inductor, and the index search range for Component 2 also contains n indices, each corresponding to a different nominal value for either capacitor or inductor, then the Cartesian product is n*n=n. 2 A combination of integer indices.

[0137] Understandably, since each of the above candidate material combinations belongs to a valid index combination in the material index space, and each combination corresponds to a set of nominal values ​​of components that can actually be purchased from the material library, the equipment can store the above integer index combinations as candidate material combinations for subsequent impedance error calculation, and then select the candidate material combination from them as the current optimal discrete solution.

[0138] Furthermore, when selecting the optimal discrete solution, an error threshold can be preset. After determining the candidate material combination with the smallest impedance error, it can be further determined whether this minimum impedance error is less than the aforementioned error threshold. If it is less, it can be returned to the aforementioned outer search stage as the current optimal discrete solution to achieve a closed-loop feedback of the two-layer search; if it is greater, it can be adjusted according to a preset step size (e.g., set to step size = 1) within the maximum radius. Within the constraints, gradually expand the aforementioned preset neighborhood radius ( To construct a new index search interval, repeat step S303 above until a candidate material combination with a minimum impedance error less than the above error threshold is obtained as the current optimal discrete solution.

[0139] This embodiment determines the dimension of the continuous parameter space based on the target matching configuration and constructs the current search space using the minimum and maximum nominal values ​​of each type of element in the material index space as boundaries. This ensures that the search range of the differential evolution algorithm strictly matches the coverage of the actual material library, avoiding invalid searches. By determining the initial individuals and constructing the initial population based on measured data, the differential evolution algorithm starts iterating from a relatively optimal initial solution, accelerating the convergence speed. By mapping the optimal nominal value combination to integer indices, constructing a neighborhood search interval, and enumerating candidate material combinations by taking the Cartesian product, it ensures that the inner search can fully cover feasible material combinations near the optimal solution, providing a complete candidate set for subsequent selection of the optimal discrete solution.

[0140] Based on the first and second embodiments of this application, in the third embodiment of this application, the content that is the same as or similar to that in embodiments one and two above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 3 , Figure 3 This is a flowchart illustrating the third embodiment of the radio frequency matching network design method based on material library constraints in this application.

[0141] In this embodiment, to specifically illustrate how impedance error is calculated in the aforementioned outer layer search and inner layer search, the impedance error calculation steps specifically include: steps A01~A04: Step A01: Determine several measured impedance values ​​and corresponding theoretical calculation values ​​based on the measured data, and obtain fitting parameters based on each measured impedance value and corresponding theoretical calculation value.

[0142] It should be noted that in the actual design of RF circuits, the PCB traces introduce non-negligible phase shifts and impedance transformations, causing a deviation between the theoretically calculated input impedance and the actual impedance measured at the port. To eliminate this deviation, this embodiment uses transmission line parameter fitting to calibrate the theoretical impedance.

[0143] It should be understood that the equipment first acquires measured data. Measured data is data obtained by engineers through testing on actual circuit boards using several known combinations of components. The data obtained from each test may include: measuring the complex impedance at the port using a network analyzer (i.e., the measured impedance value), while simultaneously recording the nominal values ​​of the components in that combination.

[0144] It should be understood that for each set of measured data, the equipment can calculate the theoretical input impedance based on the nominal values ​​of the components in its material combination, combined with the target matching configuration, operating frequency, and load impedance, following a recursive approach from the load end to the source end. This theoretical calculation can disregard the influence of PCB traces and is based solely on the complex impedance model of the components themselves. Thus, the equipment obtains several sets of data pairs, each containing a measured impedance value and a corresponding theoretically calculated value.

[0145] In practical implementation, the device can establish a transmission line model to describe the impact of PCB traces. This transmission line model treats the PCB traces as equivalent to a uniform transmission line, whose characteristics are determined by two key parameters: characteristic impedance. (Real number, unit is) ) and electrical length (Radians). The mathematical expression for this model is:

[0146] In the formula, This is the measured impedance value at the port. These are the corresponding theoretical calculation values.

[0147] The device can then utilize the aforementioned data pairs to decouple and fit the transmission line model using the least squares method or other optimization algorithms. An error function is defined, which is the sum of the squares of the differences between the measured impedance values ​​and the model predictions for each set. The device then uses an optimization algorithm to find the value that minimizes this error function. and Take a value. Then the solution obtained is... and These are the fitting parameters, used for subsequent impedance calibration.

[0148] Step A02: Based on the target matching configuration, calculate the theoretical input impedance corresponding to the combination of nominal values ​​of each component.

[0149] It should be noted that during the outer layer search, each point, i.e., each continuous value, in the current search space corresponds to a combination of component nominal values; during the inner layer search, each candidate material combination also corresponds to a combination of component nominal values. Therefore, during the dual-layer search process, the device can first calculate the corresponding theoretical input impedance based on the combination of component nominal values. This theoretical input impedance does not consider the influence of PCB traces.

[0150] Specifically, the device can be combined with target matching configurations (such as L-type, ... The theoretical input impedance corresponding to the combination of nominal component values ​​is calculated by considering the type (type or T), operating frequency, and load impedance, following a recursive approach from the load end to the source end.

[0151] The recursive process is as follows: Initialize the current total impedance as the load impedance; traverse each component in order from the load end to the source end, and calculate its complex impedance according to the component type (resistor, inductor, capacitor); if the components are connected in series, add the current total impedance to the component impedance; if the components are connected in parallel, use the parallel equivalent value of the current total impedance and the component impedance as the new total impedance; after the traversal is completed, the current total impedance is the theoretical input impedance.

[0152] Step A03: The theoretical input impedance is calibrated using the fitting parameters to obtain the input impedance.

[0153] It should be noted that the device can utilize the fitting parameters (characteristic impedance) calculated above. and electrical length The theoretical input impedance described above is then calibrated. The calibration process uses the same transmission line model as described above. Since it is known at this point... and This allows us to calculate the actual port impedance, i.e., the input impedance, after the PCB trace transformation.

[0154] In practice, the device can sequentially calibrate the theoretical input impedance of all candidate material combinations to obtain the actual port impedance, i.e., the input impedance, for each candidate material combination. Because this input impedance has been calibrated using fitting parameters, it is closer to the impedance value measured by engineers at the actual PCB port, thereby eliminating errors introduced by the transmission line.

[0155] Step A04: Calculate the complex Euclidean distance between the input impedance and the target impedance as the impedance error.

[0156] It should be understood that after obtaining the input impedances, the device can calculate the complex Euclidean distance between the input impedance and the target impedance as the impedance error. The formula for calculating the impedance error is:

[0157] In the formula, For impedance error, and These are the real and imaginary parts of the input impedance, respectively. and These are the real and imaginary parts of the target impedance, respectively.

[0158] In the specific implementation, during the outer layer search process, after calculating the impedance error of all component nominal value combinations, the device can select the component nominal value combination with the smallest impedance error as the optimal nominal value combination, which serves as the search basis for the inner layer search. In the inner layer search, after the device calculates the impedance error corresponding to each candidate material combination (corresponding to the component nominal value combination), it can select the candidate material combination with the smallest impedance error as the current optimal discrete solution, so as to return as a new individual to participate in the next round of outer layer search.

[0159] This embodiment determines the fitting parameters based on measured data and calibrates the theoretical input impedance using a transmission line model. This eliminates the impact of phase shift and impedance transformation introduced by PCB traces on impedance calculation, making the calculated input impedance closer to the actual port measurement value, thereby improving the accuracy of the matching network design. Simultaneously, by using the complex Euclidean distance between the calibrated impedance and the target impedance as the impedance error, a unified optimization objective is provided for the outer-layer differential evolution algorithm and the inner-layer discrete search, ensuring the consistency between the optimization results and the actual circuit performance.

[0160] Furthermore, this can be referenced here. Figure 4 This document provides a complete explanation of the application process. Figure 4 This is a schematic diagram illustrating the entire process of the radio frequency matching network design method based on material library constraints in this application.

[0161] Depend on Figure 4 It is known that before performing matching, the equipment first performs material data extraction and standardization operations, converting the nominal values ​​of capacitors, inductors, and resistors in the bill of materials into standard units and arranging them in ascending order, assigning a unique integer index to each nominal value, thereby constructing a material index space. When starting matching, the equipment determines the target impedance based on the current matching requirements. This can be obtained through theoretical conjugate matching or by back-calculation using radial basis function interpolation based on measured data, and the user selects the target matching configuration, i.e., L-type, ... At least one of the type or T-type topologies.

[0162] After completing the preliminary preparations, the two-layer search architecture is adopted. The outer layer performs a coarse search in the continuous parameter space, determines the spatial dimension based on the selected matching configuration, constructs the search space with the minimum and maximum nominal values ​​of the elements in the material index space as boundaries, runs the differential evolution algorithm, and through the iteration of mutation, crossover and selection operations, with the goal of minimizing impedance error, quickly locates the approximate region of the global optimum and obtains the optimal combination of nominal values.

[0163] The inner layer performs a discrete local search for material library awareness in the material index space, maps the optimal continuous solution output by the outer layer to the nearest integer index, constructs a search interval with each integer index as the center and a preset neighborhood radius, enumerates all candidate material combinations by taking the Cartesian product of each interval, evaluates each candidate material combination by calculating impedance error, and selects the candidate material combination with the smallest impedance error as the current optimal discrete solution.

[0164] After obtaining the current optimal discrete solution, the device determines whether the termination condition is met. This termination condition includes impedance error being less than a preset threshold, no improvement after multiple consecutive generations, or reaching the maximum number of iterations. If not met, the current optimal discrete solution is converted back to a continuous value and injected into the outer population to replace the worst individual before continuing iteration. If met, the optimal material combination is output, and it can be exported as a BOM, component value table, or circuit design diagram for use in actual design target matching networks.

[0165] In addition, when a user selects multiple matching configurations, the above process is executed for each matching configuration. After merging the results, they are sorted and output according to impedance error. If the number of combinations that meet the conditions is insufficient, the best combination of the next priority configuration is used as a substitute for the output.

[0166] This application embeds discrete constraints of the material library into the optimization process, employing a two-layer search architecture of outer continuous differential evolution and inner discrete neighborhood enumeration, and establishing a feedback iteration mechanism. This ensures that the output optimal material combination directly corresponds to the nominal values ​​of available components, eliminating the need for manual approximation and resolving the disconnect between design and material selection. Simultaneously, radial basis function interpolation is used to back-calculate the target impedance and transmission line parameters, improving design accuracy. Compared to manual traversal or unidirectional optimization, this application significantly improves debugging efficiency, achieving an automated closed loop from material extraction to result output.

[0167] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the radio frequency matching network design method based on material library constraints in this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0168] Furthermore, this application also provides a radio frequency matching network design device based on material library constraints, such as... Figure 5 As shown, Figure 5 This is a schematic diagram of the module structure of the RF matching network design device based on material library constraints in this application. Then, by... Figure 5 It is known that the device includes: The matching preparation module 501 is used to obtain the material index space and determine the target impedance and target matching configuration according to the current matching requirements. The outer search module 502 is used to construct the current search space based on the target matching configuration and the material index space, and run the differential evolution algorithm in the current search space to search for the optimal nominal value combination with the goal of minimizing impedance error. The impedance error is the distance between the input impedance corresponding to the continuous value in the current search space and the target impedance. The inner search module 503 is used to map the optimal nominal value combination to an integer index in the material index space, and enumerate all candidate material combinations in the neighborhood of the integer index; The inner search module 503 is also used to determine the impedance error corresponding to each candidate material combination, and select the candidate material combination with the smallest impedance error as the current optimal discrete solution. The result output module 504 is used to convert the current optimal discrete solution back to a continuous value and return to the step of running the differential evolution algorithm until the termination condition is met, and then output the optimal material combination corresponding to the current optimal discrete solution.

[0169] This application employs a technique of obtaining a material index space and mapping the optimal nominal value combination to an integer index in the material index space, and enumerating candidate material combinations within the neighborhood of the integer index. This allows the optimization process to be directly constrained to the actual procureable material library. The output optimal material combination can be directly used to design the target matching network without manual approximation, thus eliminating the disconnect between design and material selection. At the same time, by using a technique of constructing the current search space based on the target matching configuration and the material index space, running a differential evolution algorithm in the search space to obtain the optimal continuous solution, and converting the current optimal discrete solution back to a continuous value before returning to the iteration, a closed-loop feedback is formed between the outer continuous search and the inner discrete search. The optimization result continuously approaches the global optimum, significantly improving the efficiency and accuracy of the matching network design compared to manual traversal or unidirectional optimization methods.

[0170] This application also provides a library-based RF matching network design device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to execute the library-based RF matching network design method in the first embodiment described above.

[0171] The following is for reference. Figure 6 , Figure 6 This is a schematic diagram of the structure of the radio frequency matching network design device based on library constraints according to this application. The radio frequency matching network design device based on library constraints in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, personal digital assistants (PDAs), tablet computers (PADs), portable media players (PMPs), etc., as well as fixed terminals such as digital TVs, desktop computers, etc. Figure 6 The illustrated radio frequency matching network design device based on material library constraints is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0172] like Figure 6 As shown, the RF matching network design device based on library constraints may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 1002 or a program loaded from storage device 1003 into random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the RF matching network design device based on library constraints. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the library-constrained RF matching network design device to communicate wirelessly or wiredly with other devices to exchange data. Although the figure shows a library-constrained RF matching network design device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented alternatively.

[0173] The RF matching network design device based on library constraints provided in this application employs the RF matching network design method based on library constraints in the above embodiments, and can solve the technical problems of the RF matching network design method based on library constraints. Compared with the prior art, the beneficial effects of the RF matching network design device based on library constraints provided in this application are the same as the beneficial effects of the RF matching network design method based on library constraints provided in the above embodiments, and other technical features in the RF matching network design device based on library constraints are the same as the features disclosed in the method of the previous embodiment, and will not be repeated here.

[0174] This application also provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the RF matching network design method based on library constraints in the above embodiments.

[0175] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0176] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described RF matching network design method based on library constraints, thereby solving the technical problems of the RF matching network design method based on library constraints. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the RF matching network design method based on library constraints provided in the above embodiments, and will not be repeated here.

[0177] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other elements in the process, method, article, or system that includes that element.

[0178] The above embodiment numbers are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments. They are only some embodiments of this application and do not limit the scope of this application. All equivalent structural transformations made under the technical concept of this application and using the content of this application specification and drawings, or direct / indirect applications in other related technical fields, are included within the protection scope of this application.

Claims

1. A method for designing radio frequency matching networks based on material library constraints, characterized in that, The method includes: Obtain the material index space and determine the target impedance and target matching configuration based on the current matching requirements; Based on the target matching configuration and the material index space, a current search space is constructed, and a differential evolution algorithm is run in the current search space with the goal of minimizing impedance error. The optimal nominal value combination is obtained by searching. The impedance error is the distance between the input impedance corresponding to the continuous value in the current search space and the target impedance. The optimal nominal value combination is mapped to an integer index in the material index space, and all candidate material combinations are enumerated in the neighborhood of the integer index. Determine the impedance error corresponding to each candidate material combination, and select the candidate material combination with the smallest impedance error as the current optimal discrete solution; After converting the current optimal discrete solution back to a continuous value, the process returns to the step of running the differential evolution algorithm until the termination condition is met. Then, the optimal material combination corresponding to the current optimal discrete solution is output, and the optimal material combination is used to design the target matching network.

2. The method as described in claim 1, characterized in that, Before the step of obtaining the material index space, the method further includes: Obtain the current bill of materials, which includes each component type and its corresponding nominal values. Each component type includes capacitors, inductors, and resistors. After converting the nominal values ​​of each component type to standard units, arrange them in ascending order to obtain the sorting result; Based on the arrangement result, assign corresponding integer indices to each of the transformed nominal values ​​to establish a material index space.

3. The method as described in claim 1, characterized in that, The steps of determining the target impedance and target matching configuration based on the current matching requirements include: Based on the current matching requirements, the measured data is obtained, including several measured impedance values ​​and corresponding performance index values. A discrete data point set is constructed using the measured impedance values ​​and the corresponding performance index values. The discrete data point set is interpolated and fitted using radial basis functions to generate a continuous performance surface; Search for the maximum value of the performance index on the performance surface, and take the impedance value corresponding to the maximum value of the performance index as the target impedance; In response to the user's selection of candidate matching configurations, at least one target matching configuration is determined.

4. The method as described in claim 1, characterized in that, The step of constructing the current search space based on the target matching configuration and the material index space includes: The dimension of the continuous parameter space is determined based on the target matching configuration, and the dimension is equal to the number of elements in the target matching configuration; The minimum and maximum nominal values ​​of each type of element in the material index space are used as the boundaries of the continuous parameter space to construct the current search space. Each continuous value in the current search space is used to represent a combination of different element nominal values.

5. The method as described in claim 4, characterized in that, The step of running the differential evolution algorithm in the current search space, with the goal of minimizing impedance error, to search for the optimal combination of nominal values ​​includes: Based on the current matching requirements, obtain actual test data, and determine the initial individual based on the actual test data; An initial population is constructed based on the initial individuals, and a differential evolution algorithm is run in the current search space based on the initial population to search for the optimal combination of nominal values ​​with the goal of minimizing the impedance error. Wherein, the impedance error is the complex Euclidean distance between the input impedance of different individuals and the target impedance obtained by running the differential evolution algorithm, and each individual corresponds to a different continuous value in the current search space.

6. The method as described in claim 1, characterized in that, The step of mapping the optimal nominal value combination to an integer index in the material index space, and enumerating all candidate material combinations in the neighborhood of the integer index, includes: For each component nominal value in the optimal nominal value combination, the corresponding integer index is matched in the material index space according to the principle of the closest nominal value. With each integer index as the center, construct the search interval for each index according to the preset neighborhood radius; Take the Cartesian product of each of the index search intervals to obtain all candidate material combinations, and each candidate material combination corresponds to a component nominal value combination.

7. The method as described in claim 1 or 4, characterized in that, The candidate material combination corresponds to a component nominal value combination, and the steps for calculating the impedance error include: Based on the measured data, several measured impedance values ​​and corresponding theoretical calculation values ​​are determined, and fitting parameters are obtained based on each of the measured impedance values ​​and corresponding theoretical calculation values. Based on the target matching configuration, calculate the theoretical input impedance corresponding to each combination of nominal values ​​of the components; The theoretical input impedance is calibrated using the fitting parameters to obtain the input impedance; The complex Euclidean distance between the input impedance and the target impedance is calculated as the impedance error.

8. A design device for a radio frequency matching network based on a material library constraint, characterized in that, The device includes: The matching preparation module is used to obtain the material index space and determine the target impedance and target matching configuration based on the current matching requirements. The outer search module is used to construct the current search space based on the target matching configuration and the material index space, and run the differential evolution algorithm in the current search space to search for the optimal nominal value combination with the goal of minimizing impedance error. The impedance error is the distance between the input impedance corresponding to the continuous value in the current search space and the target impedance. The inner search module is used to map the optimal nominal value combination to an integer index in the material index space, and enumerate all candidate material combinations in the neighborhood of the integer index; The inner search module is also used to determine the impedance error corresponding to each candidate material combination, and select the candidate material combination with the smallest impedance error as the current optimal discrete solution. The result output module is used to convert the current optimal discrete solution back to a continuous value and then return to the step of running the differential evolution algorithm until the termination condition is met, and then output the optimal material combination corresponding to the current optimal discrete solution.

9. A radio frequency matching network design device based on material library constraints, characterized in that, The device includes: a memory, a processor, and a library-constrained radio frequency matching network design program stored in the memory and executable on the processor, wherein the library-constrained radio frequency matching network design program, when executed by the processor, implements the library-constrained radio frequency matching network design method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium stores a library-based RF matching network design program, which, when executed by a processor, implements the library-based RF matching network design method as described in any one of claims 1 to 7.

Images