Intelligent debugging method and system for 5G base station filters
By building the coupling matrix and neural network model of the filter and optimizing the filter parameters, the problem of strong dependence of traditional debugging methods is solved, and efficient and stable filter automatic debugging is achieved.
Patent Information
- Application Number
- CN202410797417.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-20
- Publication Date
- 2024-12-17
- Estimated Expiration
- 2044-06-20
AI Technical Summary
Traditional filter debugging methods are highly dependent on manual labor, resulting in low debugging efficiency and unstable quality.
By constructing the coupling matrix between the resonators in the filter, obtaining initial parameters and expected performance indicators, using neural networks to establish the logical relationship between S parameter data and screw state, optimizing the coupling matrix and performing adaptive adjustment and control, and achieving automated debugging.
Replace manual debugging, improve filter debugging efficiency, improve debugging quality, and achieve automation and accuracy.
Smart Images
Figure CN118632271B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of filter debugging technology, and in particular to an intelligent debugging method and system for 5G base station filters. Background Art
[0002] In electronic systems, digital filters are widely used. Their quality directly determines the quality of the product. Therefore, filter debugging is an important step in signal processing. In traditional filter debugging methods, the debugging strategy is based on design, that is, starting from the principle of the filter. Commonly used methods include frequency domain methods and time domain methods. However, traditional filter debugging methods are highly dependent on manual labor, resulting in low debugging efficiency and unstable debugging quality. Summary of the invention
[0003] The present application provides an intelligent debugging method and system for 5G base station filters, which are used to solve the technical problems in the prior art that traditional filter debugging methods are highly dependent on manual labor, resulting in low debugging efficiency and unstable debugging quality.
[0004] The first aspect of the present application provides an intelligent debugging method for a 5G base station filter, the method comprising: constructing a coupling matrix for the coupling relationship between resonators in the 5G base station filter; obtaining initial parameters and expected performance indicators of the base station filter; assigning values to the coupling matrix based on the initial parameters to construct an initial coupling matrix; obtaining a filter S parameter data set, performing data set learning based on a neural network, establishing a logical relationship between S parameter data and screw state, and constructing an optimal parameter analysis model; performing parameter optimization through the optimal parameter analysis model according to the expected performance indicators to obtain optimal parameters and screw state information; optimizing the coupling matrix of the initial coupling matrix based on the optimal parameters, and outputting an optimized target coupling matrix; configuring an adjustment step according to the relationship between frequency, wavelength, and size according to the optimized target coupling matrix and the screw state information, and performing adaptive adjustment control.
[0005] In a second aspect of the present application, an intelligent debugging system for a 5G base station filter is provided, the system comprising: a coupling matrix construction module, the coupling matrix construction module is used to construct a coupling matrix for the coupling relationship between resonators in the 5G base station filter; an expected performance index acquisition module, the expected performance index acquisition module is used to obtain initial parameters and expected performance indicators of the base station filter; an initial coupling matrix construction module, the initial coupling matrix construction module is used to assign the coupling matrix based on the initial parameters to construct an initial coupling matrix; an optimal parameter analysis model construction module, the optimal parameter analysis model construction module is used to obtain a filter S parameter data set, based on a neural network Conduct data set learning, establish a logical relationship between S parameter data and screw state, and construct an optimization parameter analysis model; a parameter optimization module, the parameter optimization module is used to perform parameter optimization through the optimization parameter analysis model according to the expected performance indicators, and obtain optimization parameters and screw state information; a coupling matrix optimization module, the coupling matrix optimization module is used to optimize the initial coupling matrix based on the optimization parameters, and output the optimization target coupling matrix; an adaptive adjustment control module, the adaptive adjustment control module is used to configure the adjustment step according to the optimization target coupling matrix, the screw state information, and the relationship between frequency, wavelength, and size, and perform adaptive adjustment control.
[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0007] The intelligent debugging method for 5G base station filters provided in the present application relates to the field of filter debugging technology. By constructing a coupling matrix based on the coupling relationship between resonators in the 5G base station filter, and obtaining the initial parameters and expected performance indicators of the base station filter, the coupling matrix is assigned, and the optimization parameters and screw state information are obtained according to the optimization parameter analysis model. The coupling matrix is optimized based on the optimization parameters, and the optimization target coupling matrix is output. According to the optimization target coupling matrix and the screw state information, the adjustment step size is configured to perform adaptive adjustment control, which solves the technical problem that the traditional filter debugging method in the prior art is highly dependent on manual labor, resulting in low debugging efficiency and unstable debugging quality, and achieves the technical effect of replacing manual debugging, improving the efficiency of filter job debugging, and improving the debugging quality. BRIEF DESCRIPTION OF THE DRAWINGS
[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0009] Figure 1 A flow chart of an intelligent debugging method for a 5G base station filter provided in an embodiment of the present application;
[0010] Figure 2 A schematic diagram of a process for constructing a coupling matrix in an intelligent debugging method for a 5G base station filter provided in an embodiment of the present application;
[0011] Figure 3 A schematic diagram of a process for obtaining a multi-station data source in an intelligent debugging method for a 5G base station filter provided in an embodiment of the present application;
[0012] Figure 4 A schematic diagram of the structure of an intelligent debugging system for a 5G base station filter provided in an embodiment of the present application.
[0013] Explanation of the reference numerals: coupling matrix construction module 11 , expected performance index acquisition module 12 , initial coupling matrix construction module 13 , optimal parameter analysis model construction module 14 , parameter optimization module 15 , coupling matrix optimization module 16 , adaptive adjustment control module 17 . DETAILED DESCRIPTION
[0014] The present application provides an intelligent debugging method for 5G base station filters, which is used to solve the technical problem that the traditional filter debugging method in the prior art is highly dependent on manual labor, resulting in low debugging efficiency and unstable debugging quality.
[0015] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.
[0016] It should be noted that the terms "first", "second", etc. in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Example
[0017] like Figure 1As shown, the present application provides an intelligent debugging method for a 5G base station filter, the method comprising:
[0018] P10: Construct a coupling matrix based on the coupling relationship between resonators in 5G base station filters;
[0019] Further, such as Figure 2 As shown, step P10 of the embodiment of the present application also includes:
[0020] P11: Detect and analyze the electromagnetic coupling relationship and coupling strength between the resonators in the 5G base station filter to obtain the coupling relationship and coupling strength between the resonators, wherein the coupling relationship includes mutual coupling and self-coupling;
[0021] P12: Based on the coupling relationship and coupling strength between the resonators, construct the coupling matrix, where the coupling matrix is ,in, Characterizes the mutual coupling value between resonator i and resonator j, It represents the self-coupling value of resonator i, i, j are [1, n], and n is a positive integer.
[0022] Specifically, a coupling matrix is constructed for the coupling relationship, that is, the interaction relationship, between the resonators in the 5G base station filter. First, the electromagnetic coupling relationship and coupling strength between the resonators in the 5G base station filter are detected and analyzed by electromagnetic simulation software, etc., to obtain the coupling relationship and coupling strength between the resonators. Among them, the coupling relationship refers to the way and nature of the interaction between the resonators through the electromagnetic field. For example, two resonators may interfere with or enhance each other due to the action of the electromagnetic field. This interaction is the electromagnetic coupling relationship. The coupling strength can measure the degree of electromagnetic coupling between the resonators. The greater the coupling strength, the stronger the interaction between the resonators. And the coupling relationship includes mutual coupling and self-coupling. The mutual coupling is the coupling between different resonators, and the self-coupling is the coupling between different parts of the same resonator.
[0023] Furthermore, based on the coupling relationship and coupling strength between the resonators, the coupling matrix is constructed: ,in, Characterizes the mutual coupling value between resonator i and resonator j, The coupling matrix represents the self-coupling value of resonator i, i, j are [1, n], and n is a positive integer. The coupling matrix can provide a mathematical basis for subsequent parameter optimization and filter optimization.
[0024] P20: Obtain the initial parameters and expected performance indicators of the base station filter;
[0025] P30: Assign the coupling matrix based on the initial parameters to construct an initial coupling matrix;
[0026] Optionally, according to the base station filter type, the initial parameters of the base station filter are obtained, including center frequency, bandwidth, roll-off factor, insertion loss, stopband attenuation, etc., and according to the application requirements and system specifications of the filter, the expected performance indicators are determined, including determining passband fluctuations, stopband suppression, transition bandwidth, etc. Further, the above coupling matrix is assigned using the initial parameters to construct an initial coupling matrix.
[0027] P40: Obtain the filter S parameter data set, conduct data set learning based on the neural network, establish the logical relationship between the S parameter data and the screw state, and construct an analytical model for optimizing parameters;
[0028] It should be understood that by collecting experimental data, simulation results or historical records, the filter S parameter data set is obtained. The S parameters, also known as scattering parameters, are complex parameters used to describe the behavior of microwave networks at radio frequency and microwave frequencies. They usually include S11 reflection coefficient, S21 forward transmission coefficient, S12 reverse transmission coefficient and S22 output reflection coefficient. For the filter, the S parameter data set contains S parameter measurements at different frequencies, different screw positions or different filter parameter settings.
[0029] Furthermore, the S parameter data set is learned based on a neural network, and a suitable neural network structure is selected, such as a multilayer perceptron, a convolutional neural network, etc., and the S parameter data is used as input, and the screw state or related parameters are used as output for training to establish a logical relationship between the S parameter data and the screw state, and to construct an optimal parameter analysis model. The optimal parameter analysis model can be used to perform parameter optimization and quickly find filter parameter settings that meet performance requirements.
[0030] P50: According to the expected performance index, parameter optimization is performed through the optimization parameter analysis model to obtain optimization parameters and screw state information;
[0031] Optionally, the expected performance index is used as input, and parameter optimization is performed through the optimization parameter analysis model. After the parameter optimization process, a set of optimized parameter settings are output as optimization parameters. The optimization parameters can make the performance of the filter close to or reach the expected performance index. At the same time, since the model establishes a logical relationship between the S parameter data and the screw state, the corresponding screw state information can be given while outputting the optimization parameters, including the screw position, angle or other adjustment parameters that may affect the filter performance.
[0032] P60: Optimizing the coupling matrix of the initial coupling matrix based on the optimization parameters, and outputting an optimized target coupling matrix;
[0033] Specifically, the initial coupling matrix is optimized using the optimization parameters, and the coupling coefficients in the coupling matrix are adjusted according to the optimization parameters. This can be done for a specific resonator pair or can be a global adjustment to change the frequency response of the entire filter. After multiple iterations and tests to ensure that the final coupling matrix can produce the desired filter performance, an optimized target coupling matrix is obtained, which describes the coupling relationship between the internal resonators of the optimized filter and can make the filter performance meet or approach the desired performance indicators.
[0034] P70: According to the optimization target coupling matrix, the screw state information, and the relationship between frequency, wavelength, and size, the adjustment step size is configured to perform adaptive adjustment control.
[0035] Furthermore, step P70 of the embodiment of the present application further includes:
[0036] P71: Set the plane where one end surface of the filter cavity is located as the first plane, establish a debugging device to connect with the screw, and move it in a direction perpendicular to the first plane to obtain screw depth parameter adjustment data, where the screw depth parameter adjustment data is monitoring data of screw depth adjustment, including adjustment size and resonant frequency;
[0037] P72: According to the screw depth parameter adjustment data, the relationship between the adjustment size and the resonant frequency is fitted to obtain the relationship between the frequency and the size;
[0038] P73: Get the wavelength and frequency relationship of the filter;
[0039] P74: According to the optimization target coupling matrix, the target debugging parameters are obtained, and based on the target debugging parameters, the size is adjusted to optimize according to the relationship between the frequency and the size, and the relationship between the wavelength and the frequency, and the adjustment step length of the screw depth is configured;
[0040] P75: Based on the adjustment step of the screw depth, the screw depth parameter is debugged and controlled by the debugging device;
[0041] P76: When the debugging device is adjusting the screw depth, real-time monitoring data is captured;
[0042] P77: Evaluate the adjustment parameters based on the real-time monitoring data, and adaptively adjust the adjustment step size according to the adjustment evaluation results.
[0043] Optionally, based on the optimized target coupling matrix and the screw state information, the adjustment step is configured according to the relationship between the frequency, wavelength and size, and the filter is adaptively adjusted and controlled. Specifically, the plane where one end face of the filter cavity is located is set as the first plane, and the debugging device is connected to the screw to ensure that the depth of the screw can be accurately controlled. Further, the debugging device is moved in a direction perpendicular to the first plane to change the depth of the screw, and the size of each adjustment and the corresponding resonant frequency are recorded to obtain the screw depth parameter adjustment data, that is, the monitoring data of the screw depth adjustment, including the adjustment size and resonant frequency. The original technical solution of the absolute parallelism between the debugging device and the filter is modified, thereby eliminating the accuracy error of the gantry in the prior art, avoiding the problem of difficulty in keeping level with the plane where the filter is located and difficulty in ensuring the accuracy on the z-axis due to the accuracy error of the gantry, thereby ensuring the accuracy of the debugging system in terms of horizontal accuracy, and improving the accuracy of the scheduling result on the z-axis, that is, the depth parameter debugging.
[0044] Furthermore, the screw depth parameter adjustment data is used to obtain the mathematical relationship between the screw adjustment dimension and the resonant frequency through data analysis or curve fitting methods, and based on the design principle of the filter, the basic relationship between the wavelength and frequency of the filter is obtained.
[0045] Furthermore, parameters directly related to screw adjustment are extracted from the optimization target coupling matrix as target debugging parameters, and based on the target debugging parameters, the adjustment size is optimized according to the relationship between frequency and size, and the relationship between wavelength and frequency through calculation or search algorithm, the optimal screw adjustment size is determined, and then the adjustment step of the screw depth is configured.
[0046] Furthermore, according to the adjustment step of the screw depth, the screw depth is precisely controlled by a debugging device to achieve the desired filter performance, and during the screw depth adjustment process, the filter performance data, such as resonant frequency, bandwidth, etc., are captured and recorded in real time to obtain real-time monitoring data. Furthermore, adjustment parameter evaluation is performed based on the real-time monitoring data, and the impact of the current adjustment step on the filter performance is evaluated by analyzing the real-time monitoring data, and the adjustment step of the screw depth is dynamically adjusted according to the adjustment evaluation results to improve the adjustment efficiency and accuracy. If the current step length leads to improved performance, the step length may continue to be used. If the performance decreases or the change is not obvious, it may be necessary to reduce the step length or change the adjustment direction. Through adaptive adjustment control, efficient and precise adjustment of filter performance is achieved.
[0047] Furthermore, step P77 of the embodiment of the present application also includes:
[0048] P77-1: Construct an adaptive activation function, in which the activation logic relationship between the adaptive activation instruction and the adjustment parameter evaluation result is configured;
[0049] P77-2: When the adjustment parameter evaluation result satisfies the activation logic relationship of the adaptive activation function, the adaptive activation instruction is generated, the adjustment size optimization space is called, the adjustment step length is optimized according to the difference between the adjustment monitoring data and the target debugging parameter and the approach speed, and the adjustment step length is adjusted using the optimization result;
[0050] P77-3: Based on the adaptive activation function and adjusting the size optimization space, iterate according to the real-time adjustment parameter evaluation results until the target debugging parameters are reached.
[0051] Optionally, by collecting sample adjustment data for mathematical statistical analysis, an adaptive activation function is constructed, and the adaptive activation function is a logic judgment mechanism for determining when to activate the adaptive adjustment process of the adjustment step size, which can be determined based on the relationship between the adjustment parameter evaluation result and the preset activation condition. And an activation logic relationship between the adaptive activation instruction and the adjustment parameter evaluation result is configured, for example, when the adjustment parameter evaluation result, such as the performance improvement rate, the error reduction speed, etc., reaches a certain threshold or meets a specific trend, the adaptive adjustment process is activated.
[0052] Furthermore, when the evaluation results of the adjustment parameters monitored in real time meet the activation logic relationship of the adaptive activation function, an adaptive activation instruction is generated, and the adjustment step length is optimized in the adjustment size optimization space according to the difference and the approach speed between the adjustment monitoring data and the target debugging parameters. For example, the optimal adjustment step length is found through optimization algorithms such as gradient descent and genetic algorithm, and the current screw depth adjustment step length is adjusted according to the optimal step length found by the optimization algorithm.
[0053] Furthermore, after each adjustment of the screw depth, new real-time monitoring data is obtained and the adjustment parameters are evaluated. Then, based on the evaluation results and the logic of the adaptive activation function, it is decided whether to continue iteratively adjusting the step length, and so on and so forth until the target debugging parameters are reached, thereby achieving adaptive adjustment of the screw depth adjustment step length.
[0054] Furthermore, the embodiment of the present application further includes step P80, and step P80 further includes:
[0055] P81: Build a minimized debugging channel, where the minimized debugging channel has a single piece of test data;
[0056] P82: Get multi-station data source;
[0057] P83: Based on the multi-station data source, multiple minimized debugging channels are configured, loaded to the edge computing end for parallel calculation, and debugging data of each station is obtained.
[0058] Specifically, in the process of filter debugging, in order to improve efficiency and reduce resource usage, a minimized debugging channel can be constructed, that is, an easy-to-process data structure can be constructed for massive data analysis, and the minimized debugging channel has a single test data, that is, the minimized debugging channel focuses on the debugging of a single function or parameter and only transmits necessary test data. Because the data source may come from different workstations in the future, the selective debugging channel with minimized data packets not only avoids the possible bottleneck of high data volume in transmission rate, but also solves the problem of distributed computing of massive data in the cloud.
[0059] Furthermore, in an actual production or debugging environment, there may be multiple workstations for debugging or testing at the same time, and each workstation will generate a data source related to its debugging task, so it is necessary to obtain the multi-station data source from each workstation in real time or periodically. Furthermore, the multi-station data source is configured with multiple minimized debugging channels, and the multi-station data source is loaded to the edge computing end for parallel computing to obtain the debugging data of each workstation. Among them, the edge computing end is usually deployed close to the data source, can quickly receive, process and forward data, and can perform parallel computing to realize the simultaneous processing and analysis of debugging data of multiple workstations, thereby improving debugging efficiency.
[0060] Further, such as Figure 3 As shown, step P82 of the embodiment of the present application also includes:
[0061] P82-1: Decomposing each S parameter test index according to the multi-station data source to obtain each test index data, and performing data frequency analysis on each test index data to determine the resolution bandwidth;
[0062] P82-2: constructing a reporting data packet of each test index data with the resolution bandwidth as the minimum granularity;
[0063] P82-3: Obtain the debugging status of the workstation product, generate a status tag, and add the status tag to the reporting data packet, where the status tag corresponds to the debugging status.
[0064] It should be understood that, based on the multi-station data source, the S parameter test index of each station product is decomposed to obtain each specific test index data, that is, any one of the S parameters, such as S21, S11, S12. Further, the data frequency analysis is performed on the test index data to determine the frequency distribution, main frequency components, etc. of the data, and then determine the resolution bandwidth. The resolution bandwidth represents the minimum granularity of data processing, that is, the minimum frequency range that the system can distinguish and process. By selecting an appropriate resolution bandwidth, the accuracy and efficiency of data processing can be ensured.
[0065] Furthermore, a reporting data packet of each test index data is constructed with the resolution bandwidth as the minimum granularity, and the reporting data packet includes the test index data related to a specific workstation product, and frequency information of the data, such as frequency range, resolution bandwidth, etc.
[0066] Furthermore, the debugging status of the workstation product is obtained, such as normal, abnormal, waiting for debugging, etc., and a corresponding status tag is generated based on the obtained debugging status, and the status tag is added to the reporting data packet. The status tag corresponds to the debugging status and can be used to identify the current status of the product.
[0067] Furthermore, before configuring a plurality of the minimized debugging channels based on the multi-station data source and loading them to the edge computing end for parallel computing, the embodiment of the present application further includes step P82a, and step P82a further includes:
[0068] P82-1a: Obtain a debugging target according to the status tags of the data sources of each workstation;
[0069] P82-2a: According to the debugging target, a data collection strategy is configured, and data is extracted from the data source of each workstation according to the data collection strategy, including configuring the data with low resolution when the debugging target is production capacity, and configuring the data with high resolution when the debugging target is quality.
[0070] Specifically, multiple minimized debugging channels are configured based on the multi-station data source and loaded to the edge computing end for parallel computing before being loaded. The debugging target corresponding to each station is obtained according to the status label of each station data source. For example, if the station status label shows insufficient production capacity, the debugging target may be to increase production capacity. If the quality is abnormal, the debugging target may be to solve quality problems.
[0071] Furthermore, according to the debugging target, the data collection strategy is configured, including the frequency, range, resolution, etc. of data collection. And data is extracted from the data sources of each workstation according to the data collection strategy. According to different debugging targets, the data resolution is configured. When the debugging target is production capacity, the main focus is on production speed and efficiency, and a lower data resolution can be configured to reduce the complexity of data processing and analysis. When the debugging target is quality, it is necessary to capture and analyze the details and changes in the data more accurately, so a higher data resolution is configured.
[0072] Furthermore, the embodiment of the present application further includes step P90, and step P90 further includes:
[0073] P91: Build Python platform, encapsulate VISA library, and establish VISA virtual instrument architecture;
[0074] P92: Based on the VISA virtual instrument architecture, decoupling the application of different test instruments and test systems at different workstations is performed, and data connections between the different test instruments and test systems are established;
[0075] P93: Identify indicators of test data based on keyword recognition technology and determine test indicators.
[0076] In a possible embodiment of the present application, before obtaining a multi-station data source, a Python platform is constructed, a VISA library is encapsulated, and a VISA virtual instrument architecture is established. Wherein, the Python platform is a widely used programming language. In this step, Python is selected as a platform to build an integrated framework for a test system, and the VISA library is a standard for configuring, programming, and troubleshooting interface instruments. Establishing a VISA virtual instrument architecture means implementing a unified interface on a Python platform for communicating with various physical test instruments, which enables the test system to process data from different instruments in a standardized manner.
[0077] Furthermore, through the VISA virtual instrument architecture, the application decoupling of different test instruments and test systems at different workstations is carried out, that is, the VISA architecture is used to realize the decoupling between the underlying resources such as instrument type, driver, interface protocol, instruction set and upper-level application, and achieve compatibility with most mainstream models. PyVISA is used to encapsulate the VISA library, and the simplicity of the scripting language is fully utilized to import the upper-level application of the instrument automatic matching link and the subsequent requirements, and the data analysis link is unified to the Python platform to realize the automatic matching of test instruments. On the basis of decoupling, the data connection between the different test instruments and the test system is established. The following table lists the main problems and technical paths solved by the automatic matching technology of test instruments:
[0078] Main Issues Traditional testing software Test instrument automatic matching technology Instrument Brand / Model Compatibility So brand and model debugging Relying on VISA virtual instrument architecture, it is compatible with most mainstream models Automatic instrument discovery Not supported Relying on SNMP technology to support automatic discovery of instruments Automatic interface identification Not supported, manual configuration of interface type and IP address is required Yes, VISA architecture inherently realizes the decoupling of USB / GPIB / LAN from upper-layer applications Difficulty in compatibility with new instrument models More difficult, requiring secondary development and debugging Easy, using the PyVISA package library, making full use of the simplicity of the original language and the decoupling of the VISA architecture from the underlying layer to achieve fast version selection
[0079] Furthermore, through keyword recognition technology, the test system can automatically identify various test indicators from the test data, that is, use keyword recognition technology to decouple the client's test requirements from the test process inside the software. This solves the problem of non-standard description of client test indicators. The test indicator may be a value or state related to filter performance, circuit parameters or other related factors, which can provide data support for subsequent testing and analysis.
[0080] In summary, the embodiments of the present application have at least the following technical effects:
[0081] This application constructs a coupling matrix based on the coupling relationship between resonators in the 5G base station filter, obtains the initial parameters and expected performance indicators of the base station filter, assigns the coupling matrix, obtains the optimal parameters and screw state information according to the optimal parameter analysis model, optimizes the coupling matrix based on the optimal parameters, outputs the optimized target coupling matrix, configures the adjustment step size according to the optimized target coupling matrix and the screw state information, and performs adaptive adjustment control.
[0082] The technical effect of replacing manual debugging, improving the efficiency of filter debugging and improving the debugging quality is achieved. Example
[0083] Based on the same inventive concept as the intelligent debugging method for 5G base station filter in the aforementioned embodiment, Figure 4 As shown, the present application provides an intelligent debugging system for 5G base station filters, and the system and method embodiments in the embodiments of the present application are based on the same inventive concept. The system includes:
[0084] A coupling matrix building module 11, wherein the coupling matrix building module 11 is used to build a coupling matrix according to the coupling relationship between resonators in a 5G base station filter;
[0085] An expected performance indicator acquisition module 12, wherein the expected performance indicator acquisition module 12 is used to acquire initial parameters and expected performance indicators of a base station filter;
[0086] An initial coupling matrix construction module 13, the initial coupling matrix construction module 13 is used to assign the coupling matrix based on the initial parameters to construct an initial coupling matrix;
[0087] An optimal parameter analysis model building module 14, wherein the optimal parameter analysis model building module 14 is used to obtain a filter S parameter data set, perform data set learning based on a neural network, establish a logical relationship between the S parameter data and the screw state, and build an optimal parameter analysis model;
[0088] A parameter optimization module 15, wherein the parameter optimization module 15 is used to perform parameter optimization through the optimization parameter analysis model according to the expected performance index to obtain optimization parameters and screw state information;
[0089] A coupling matrix optimization module 16, wherein the coupling matrix optimization module 16 is used to optimize the initial coupling matrix based on the optimization parameters and output an optimized target coupling matrix;
[0090] The adaptive adjustment control module 17 is used to configure the adjustment step size and perform adaptive adjustment control according to the relationship between the frequency, wavelength and size according to the optimization target coupling matrix, the screw state information, etc.
[0091] Furthermore, the coupling matrix construction module 11 is also used to perform the following steps:
[0092] Detect and analyze the electromagnetic coupling relationship and coupling strength between the resonators in the 5G base station filter to obtain the coupling relationship and coupling strength between the resonators, wherein the coupling relationship includes mutual coupling and self-coupling;
[0093] Based on the coupling relationship and coupling strength between the resonators, the coupling matrix is constructed, wherein the coupling matrix is ,in, Characterizes the mutual coupling value between resonator i and resonator j, It represents the self-coupling value of resonator i, i, j are [1, n], and n is a positive integer.
[0094] Furthermore, the adaptive adjustment control module 17 is also used to perform the following steps:
[0095] The plane where one end surface of the filter cavity is located is set as the first plane, the debugging device is connected to the screw, and the debugging device is moved in a direction perpendicular to the first plane to obtain screw depth parameter adjustment data, wherein the screw depth parameter adjustment data is monitoring data of screw depth adjustment, including adjustment size and resonant frequency;
[0096] According to the screw depth parameter adjustment data, the relationship between the adjustment size and the resonant frequency is fitted to obtain the relationship between the frequency and the size;
[0097] Get the wavelength and frequency relationship of the filter;
[0098] According to the optimization target coupling matrix, the target debugging parameters are obtained, and based on the target debugging parameters, the size is adjusted to optimize according to the relationship between the frequency and the size, and the relationship between the wavelength and the frequency, and the adjustment step length of the screw depth is configured;
[0099] Based on the adjustment step of the screw depth, the screw depth parameter is debugged and controlled by the debugging device;
[0100] When the debugging device performs screw depth adjustment, real-time monitoring data is captured;
[0101] An adjustment parameter evaluation is performed according to the real-time monitoring data, and the adjustment step size is adaptively adjusted according to the adjustment evaluation result.
[0102] Furthermore, the adaptive adjustment control module 17 is also used to perform the following steps:
[0103] Constructing an adaptive activation function, wherein the adaptive activation instruction is configured to adjust the activation logic relationship between the parameter evaluation results;
[0104] When the adjustment parameter evaluation result satisfies the activation logic relationship of the adaptive activation function, the adaptive activation instruction is generated, the adjustment size optimization space is called, the adjustment step length is optimized according to the difference between the adjustment monitoring data and the target debugging parameter and the approach speed, and the adjustment step length is adjusted using the optimization result;
[0105] Based on the adaptive activation function and the size optimization space, the iteration is performed according to the real-time adjustment parameter evaluation result until the target debugging parameter is reached.
[0106] Furthermore, the system also includes:
[0107] A minimized debugging channel construction module, wherein the minimized debugging channel construction module is used to construct a minimized debugging channel, wherein the minimized debugging channel has a single piece of test data;
[0108] A multi-station data source acquisition module, wherein the multi-station data source acquisition module is used to acquire a multi-station data source;
[0109] The debugging data acquisition module for each workstation is used to configure multiple minimized debugging channels based on the multi-workstation data source, load them to the edge computing end for parallel computing, and obtain the debugging data of each workstation.
[0110] Furthermore, the system also includes:
[0111] A resolution bandwidth determination module, the resolution bandwidth determination module is used to decompose each S parameter test index according to the multi-station data source, obtain each test index data, and perform data frequency analysis on each test index data to determine the resolution bandwidth;
[0112] A reporting data packet construction module, the reporting data packet construction module is used to construct a reporting data packet of each test indicator data with the resolution bandwidth as the minimum granularity;
[0113] A status label generation module is used to obtain the debugging status of the workstation product, generate a status label, and add the status label to the reporting data packet. The status label corresponds to the debugging status.
[0114] Furthermore, the system also includes:
[0115] A debugging target acquisition module, the debugging target acquisition module is used to obtain a debugging target according to the status tags of the data sources of each workstation;
[0116] A data extraction module is used to configure a data collection strategy according to the debugging target, and extract data from the data sources of each workstation according to the data collection strategy, including configuring the data with low resolution when the debugging target is production capacity and configuring the data with high resolution when the debugging target is quality.
[0117] Furthermore, the system also includes:
[0118] A virtual instrument architecture building module, wherein the virtual instrument architecture building module is used to build a Python platform, encapsulate a VISA library, and build a VISA virtual instrument architecture;
[0119] An application decoupling module, wherein the application decoupling module is used to perform application decoupling of different test instruments and a test system at different workstations based on the VISA virtual instrument architecture, and to establish data connections between the different test instruments and the test system;
[0120] The test indicator determination module is used to identify indicators of test data based on keyword recognition technology to determine test indicators.
[0121] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0122] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.
[0123] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.
Claims
1. An intelligent debugging method for 5G base station filters, characterized in that: The method comprises: Construct a coupling matrix based on the coupling relationship between resonators in 5G base station filters; Obtaining initial parameters and expected performance indicators of base station filters; Assigning the coupling matrix based on the initial parameters to construct an initial coupling matrix; Obtain the filter S parameter data set, conduct data set learning based on neural network, establish the logical relationship between S parameter data and screw state, and build an analytical model for optimal parameter search; According to the expected performance index, parameter optimization is performed through the optimization parameter analysis model to obtain optimization parameters and screw state information; Perform coupling matrix optimization on the initial coupling matrix based on the optimization parameters, and output an optimized target coupling matrix; According to the optimization target coupling matrix, the screw state information, and the relationship between frequency, wavelength, and size, an adjustment step is configured to perform adaptive adjustment control; According to the optimization target coupling matrix, the screw state information, and the relationship between the frequency, wavelength, and size, the adjustment step is configured to perform adaptive adjustment control, including: The plane where one end surface of the filter cavity is located is set as the first plane, the debugging device is connected to the screw, and the debugging device is moved in a direction perpendicular to the first plane to obtain screw depth parameter adjustment data, wherein the screw depth parameter adjustment data is monitoring data of screw depth adjustment, including adjustment size and resonant frequency; According to the screw depth parameter adjustment data, the relationship between the adjustment size and the resonant frequency is fitted to obtain the relationship between the frequency and the size; Get the wavelength and frequency relationship of the filter; According to the optimization target coupling matrix, the target debugging parameters are obtained, and based on the target debugging parameters, the size is adjusted to optimize according to the relationship between the frequency and the size, and the relationship between the wavelength and the frequency, and the adjustment step length of the screw depth is configured; Based on the adjustment step of the screw depth, the screw depth parameter is debugged and controlled by the debugging device; When the debugging device performs screw depth adjustment, real-time monitoring data is captured; Performing an adjustment parameter evaluation according to the real-time monitoring data, and adaptively adjusting the adjustment step size according to the adjustment evaluation result; The step size is adaptively adjusted according to the adjustment evaluation result, including: Constructing an adaptive activation function, wherein the adaptive activation instruction is configured to adjust the activation logic relationship between the parameter evaluation results; When the adjustment parameter evaluation result satisfies the activation logic relationship of the adaptive activation function, the adaptive activation instruction is generated, the adjustment size optimization space is called, the adjustment step length is optimized according to the difference between the adjustment monitoring data and the target debugging parameter and the approach speed, and the adjustment step length is adjusted using the optimization result; Based on the adaptive activation function and the size optimization space, the iteration is performed according to the real-time adjustment parameter evaluation result until the target debugging parameter is reached.
2. The method according to claim 1, characterized in that The constructing of the coupling matrix comprises: Detect and analyze the electromagnetic coupling relationship and coupling strength between the resonators in the 5G base station filter to obtain the coupling relationship and coupling strength between the resonators, wherein the coupling relationship includes mutual coupling and self-coupling; Based on the coupling relationship and coupling strength between the resonators, the coupling matrix is constructed, wherein the coupling matrix is Among them, m ij Characterizes the mutual coupling value between resonator i and resonator j, m ii It represents the self-coupling value of resonator i, i, j are [1, n], and n is a positive integer.
3. The method according to claim 1, characterized in that The method further comprises: Building a minimized debugging channel, wherein the minimized debugging channel has a single piece of test data; Get multi-station data sources; Based on the multi-station data source, multiple minimized debugging channels are configured and loaded to the edge computing end for parallel calculation to obtain debugging data of each station.
4. The method according to claim 3, characterized in that The obtaining of a multi-station data source includes: Decomposing each S parameter test index according to the multi-station data source to obtain each test index data, and performing data frequency analysis on each test index data to determine the resolution bandwidth; Taking the resolution bandwidth as the minimum granularity, constructing a reporting data packet for each test indicator data; The debugging status of the workstation product is obtained, a status tag is generated, and the status tag is added to the reporting data packet, wherein the status tag corresponds to the debugging status.
5. The method according to claim 4, characterized in that Configuring a plurality of the minimized debugging channels based on the multi-station data source and before loading to the edge computing end for parallel computing includes: Obtaining a debugging target according to the status tags of the data sources of each workstation; According to the debugging target, a data collection strategy is configured, and data is extracted from the data sources of each workstation according to the data collection strategy, including configuring the data with low resolution when the debugging target is production capacity and configuring the data with high resolution when the debugging target is quality.
6. The method according to claim 3, characterized in that Before obtaining the multi-station data source, the method further includes: Build Python platform, encapsulate VISA library, and establish VISA virtual instrument architecture; Based on the VISA virtual instrument architecture, decoupling the application of different test instruments and test systems at different workstations is performed, and data connections between the different test instruments and test systems are established; Based on keyword recognition technology, the test data is identified and the test indicators are determined.
7. Intelligent debugging system for 5G base station filter, characterized in that: The system comprises: A coupling matrix building module, wherein the coupling matrix building module is used to build a coupling matrix for the coupling relationship between resonators in a 5G base station filter; An expected performance index acquisition module, which is used to acquire initial parameters and expected performance indicators of the base station filter; An initial coupling matrix construction module, wherein the initial coupling matrix construction module is used to assign the coupling matrix based on the initial parameters to construct an initial coupling matrix; An optimal parameter analysis model construction module is used to obtain a filter S parameter data set, perform data set learning based on a neural network, establish a logical relationship between the S parameter data and the screw state, and construct an optimal parameter analysis model; A parameter optimization module, wherein the parameter optimization module is used to perform parameter optimization through the optimization parameter analysis model according to the expected performance index to obtain optimization parameters and screw state information; A coupling matrix optimization module, wherein the coupling matrix optimization module is used to optimize the initial coupling matrix based on the optimization parameters and output an optimized target coupling matrix; An adaptive adjustment control module, the adaptive adjustment control module is used to configure the adjustment step size according to the relationship between the frequency, wavelength and size, and perform adaptive adjustment control according to the optimization target coupling matrix, the screw state information, and the relationship between the frequency, wavelength and size; The adaptive regulation control module is also used to perform the following steps: The plane where one end surface of the filter cavity is located is set as the first plane, the debugging device is connected to the screw, and the debugging device is moved in a direction perpendicular to the first plane to obtain screw depth parameter adjustment data, wherein the screw depth parameter adjustment data is monitoring data of screw depth adjustment, including adjustment size and resonant frequency; According to the screw depth parameter adjustment data, the relationship between the adjustment size and the resonant frequency is fitted to obtain the relationship between the frequency and the size; Get the wavelength and frequency relationship of the filter; According to the optimization target coupling matrix, the target debugging parameters are obtained, and based on the target debugging parameters, the size is adjusted to optimize according to the relationship between the frequency and the size, and the relationship between the wavelength and the frequency, and the adjustment step length of the screw depth is configured; Based on the adjustment step of the screw depth, the screw depth parameter is debugged and controlled by the debugging device; When the debugging device performs screw depth adjustment, real-time monitoring data is captured; Performing an adjustment parameter evaluation according to the real-time monitoring data, and adaptively adjusting the adjustment step size according to the adjustment evaluation result; The step size is adaptively adjusted according to the adjustment evaluation result, including: Constructing an adaptive activation function, wherein the adaptive activation instruction is configured to adjust the activation logic relationship between the parameter evaluation results; When the adjustment parameter evaluation result satisfies the activation logic relationship of the adaptive activation function, the adaptive activation instruction is generated, the adjustment size optimization space is called, the adjustment step length is optimized according to the difference between the adjustment monitoring data and the target debugging parameter and the approach speed, and the adjustment step length is adjusted using the optimization result; Based on the adaptive activation function and the size optimization space, the iteration is performed according to the real-time adjustment parameter evaluation result until the target debugging parameter is reached.