A dynamic management system for a power amplifier

Abstract

The present invention relates to the field of wireless communication technologies, and particularly to a dynamic management system for a power amplifier. The data acquisition and processing module collects real-time characteristic data at preset time nodes, and uses the fuzzy clustering algorithm to denoise the data, removing the noise data to improve the data quality; the denoised real-time characteristic data is imported into the prediction model, and the working state of the power amplifier is predicted through the state transition probability matrix and neural network technologies; if the working state is an unstable state, the regulation module combines the historical regulation log and knowledge graph technology to match the optimal regulation scheme and implement the regulation process to ensure that the power amplifier quickly resumes stable operation. This system can monitor the operation state of the power amplifier in real time, accurately predict the instability risk, and effectively avoid performance degradation or faults through intelligent regulation, improving the stability and reliability of the operation of the power amplifier.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and in particular to a dynamic management system for a power amplifier. Background Art

[0002] As a key component in modern electronic systems, the performance of a power amplifier directly affects the stability and efficiency of the entire system. However, during operation, a power amplifier is susceptible to various factors, such as input signal fluctuations, load changes, and environmental temperature variations. These factors may cause the operating state of the power amplifier to become unstable, thereby affecting system performance and even leading to failures. Traditional power amplifier management methods usually rely on fixed operating parameters and manual adjustments, making it difficult to cope with complex and changing operating environments and lacking the ability to accurately monitor and predict real-time states. With the increasing performance requirements of electronic systems for power amplifiers, how to achieve dynamic management of power amplifiers and ensure their stable operation under various working conditions has become a technical problem that urgently needs to be solved.

[0003] In recent years, with the rapid development of artificial intelligence, big data, and Internet of Things technologies, data-driven dynamic management systems have gradually become an effective way to solve this problem. By constructing a power amplifier operating state prediction model and combining the acquisition and processing of real-time feature data, accurate prediction and dynamic regulation of the working state of the power amplifier can be achieved. In addition, the introduction of a communication interface module and a human-machine interaction interface enables the system to communicate efficiently with a host computer or other devices and provides an intuitive operation interface for users, further enhancing the intelligent level and user experience of the system. Therefore, developing a power amplifier dynamic management system that integrates model construction, data acquisition and processing, dynamic regulation, communication interaction, and a human-machine interface can not only significantly improve the operating efficiency and reliability of power amplifiers but also provide strong support for the intelligent management of electronic systems. Summary of the Invention

[0004] The present invention overcomes the deficiencies of the prior art and provides a dynamic management system for a power amplifier.

[0005] The technical solution adopted by the present invention to achieve the above object is as follows:

[0006] The present invention discloses a dynamic management system for a power amplifier, including:

[0007] A model construction module that constructs a power amplifier operating state prediction model based on the dynamic characteristic data and working state of the power amplifier over a preset time period;

[0008] The data acquisition and processing module collects the real-time characteristic data of the power amplifier at preset time nodes through the data acquisition and processing module, denoises the real-time characteristic data, and imports the denoised real-time characteristic data into the power amplifier operation status prediction model for prediction to obtain the working status of the power amplifier;

[0009] The regulation module, if the working status of the power amplifier is the normal status, does not perform regulation processing on the power amplifier; if the working status of the power amplifier is the unstable status, performs regulation processing on the power amplifier;

[0010] The communication interface module realizes communication with the upper computer or other systems and data transmission through wired or wireless means;

[0011] The human-computer interaction interface provides a user operation interface, including a display screen, buttons, and a touch screen.

[0012] Further, a power amplifier operation status prediction model is constructed through the dynamic characteristic data and working status of the power amplifier under a preset time period, specifically:

[0013] Obtain the operation log of the power amplifier, and obtain the dynamic characteristic data of the power amplifier under a preset time period according to the operation log; and obtain the working status of the power amplifier at each time stamp within the preset time period according to the operation log;

[0014] Introduce a Markov chain, define all possible state combinations for the characteristic data at each time stamp, including the characteristic data and working status at each time stamp, to form a state space;

[0015] Initialize a two-dimensional state transition matrix, where the rows and columns correspond to the states in the state space respectively, and all elements are set to zero initially;

[0016] Read the data from the operation log of the power amplifier in chronological order, and extract the state combinations of each time stamp and its previous time stamp;

[0017] For each time stamp, analyze whether the state combination of its previous time stamp exists in the state space. If it exists, increase the transfer count by one at the corresponding position in the state transition matrix; and so on, traverse the entire state transition matrix to calculate the total transfer times of each state combination;

[0018] Divide the transfer times between each state combination by its total transfer times to obtain the transfer probability of each state combination, and generate a state transition probability matrix according to the transfer probability of each state combination;

[0019] Construct a prediction model based on a neural network, and import the state transition probability matrix into the prediction model for encoding learning. After the prediction accuracy is greater than the preset accuracy, output the power amplifier operating state prediction model.

[0020] Further, perform denoising processing on the real-time feature data, specifically:

[0021] Introduce a fuzzy clustering algorithm, initialize the clustering center and the fuzzy membership function to adapt to the distribution characteristics of the data;

[0022] For each real-time feature data, use the fuzzy membership function to calculate its membership degree with each initial fuzzy center, and obtain the membership degrees between each real-time feature data and each initial fuzzy center;

[0023] Assign each real-time feature data to the initial fuzzy center with the highest membership degree to obtain several fuzzy data clusters;

[0024] After the assignment is completed, calculate the mean value of the real-time feature data belonging to each fuzzy data cluster, and use the mean value of the real-time feature data belonging to each fuzzy data cluster as the new fuzzy center;

[0025] Calculate the Euclidean distance between the new fuzzy center and the initial fuzzy center of each fuzzy data cluster; if the Euclidean distances between the new fuzzy centers and the initial fuzzy centers of all fuzzy data clusters are less than the preset distance threshold, stop the iterative assignment;

[0026] If there is at least one case where the Euclidean distance between the new fuzzy center and the initial fuzzy center of a fuzzy data cluster is greater than the preset distance threshold, re-initialize the clustering center and the membership function parameters of the fuzzy clustering model, and re-iterate the assignment of the real-time feature data until the Euclidean distances between the new fuzzy centers and the initial fuzzy centers of all fuzzy data clusters are less than the preset distance threshold, then stop the iterative assignment;

[0027] After receiving the stop iterative assignment instruction, calculate the Mahalanobis distance between the real-time feature data belonging to each fuzzy data cluster and its new fuzzy center;

[0028] If the Mahalanobis distance between the real-time feature data belonging to a certain fuzzy data cluster and its new fuzzy center is greater than the preset Mahalanobis distance value, mark this real-time feature data as noise data; repeat this step until all the real-time feature data belonging to each fuzzy data cluster have been discriminated;

[0029] Completely delete all the real-time feature data marked as noise data to obtain the denoised real-time feature data.

[0030] Further, import the denoised real-time feature data into the power amplifier operating state prediction model for prediction to obtain the operating state of the power amplifier, specifically as follows:

[0031] Obtain the denoised real-time feature data, and import the denoised real-time feature data into the power amplifier operating state prediction model for prediction;

[0032] Through prediction, obtain the state transition probability of the power amplifier at the current preset time node; and compare the state transition probability of the power amplifier at the current preset time node with a preset probability threshold;

[0033] When the state transition probability of the power amplifier at the current preset time node is not greater than the preset probability threshold, update the operating state of the power amplifier to the normal state;

[0034] When the state transition probability of the power amplifier at the current preset time node is greater than the preset probability threshold, update the operating state of the power amplifier to the unstable state.

[0035] Further, if the operating state of the power amplifier is the unstable state, perform regulation processing on the power amplifier, specifically as follows:

[0036] If the operating state of the power amplifier is the unstable state, obtain various real-time feature data of the power amplifier at the real-time time node;

[0037] Calculate the difference between various real-time feature data and corresponding preset feature data; compare the difference between various real-time feature data and corresponding preset feature data with a preset difference threshold;

[0038] If the difference between a certain real-time feature data and the corresponding preset feature data is not greater than the preset difference threshold, mark the corresponding real-time feature data as normal feature data;

[0039] If the difference between a certain real-time feature data and the corresponding preset feature data is greater than the preset difference threshold, mark the corresponding real-time feature data as drift feature data, and obtain the drift amplitude of the drift feature data to obtain the drift feature data and its drift amplitude of the power amplifier at the real-time time node; where the drift amplitude is the difference between this type of real-time feature data and the corresponding preset feature data.

[0040] Further, if the operating state of the power amplifier is the unstable state, the regulation processing on the power amplifier further includes the following steps:

[0041] Obtain the regulation log of the power amplifier, and obtain the corresponding historical regulation schemes when various characteristic data drift events occur in the power amplifier according to the regulation log; wherein, the characteristic data drift events include event records of various drift characteristic data and their corresponding drift amplitudes;

[0042] And obtain the regulation success rates of various historical regulation schemes, screen out the historical regulation scheme with the highest regulation success rate as the optimal regulation scheme for the corresponding characteristic data drift event, and obtain the optimal regulation schemes corresponding to various characteristic data drift events;

[0043] Construct a knowledge graph, and import the optimal regulation schemes corresponding to various characteristic data drift events into the knowledge graph;

[0044] Import the drift characteristic data and its drift amplitude of the power amplifier at the real-time time node into the knowledge graph for matching, match the corresponding optimal regulation scheme, and perform adjustment processing on the power amplifier according to the matched optimal regulation scheme.

[0045] The real-time characteristic data includes channel gain, bandwidth, operating voltage, operating current, temperature, humidity, signal amplitude, signal frequency, and signal distortion.

[0046] The present invention solves the technical defects existing in the background technology, and the present invention has the following beneficial effects:

[0047] Through the model construction module, use the dynamic characteristic data and operating state of the power amplifier under a preset time period to construct an operating state prediction model; through the data acquisition and processing module, collect real-time characteristic data at a preset time node, and use the fuzzy clustering algorithm to denoise the data, removing noise data to improve data quality; import the denoised real-time characteristic data into the prediction model, and predict the operating state of the power amplifier through the state transition probability matrix and neural network technology; if the operating state is an unstable state, then through the regulation module, combine the historical regulation log and knowledge graph technology, match the optimal regulation scheme and implement the regulation processing to ensure that the power amplifier quickly resumes stable operation. This system can monitor the operating state of the power amplifier in real time, accurately predict the instability risk, and effectively avoid performance degradation or failures through intelligent regulation, improve the stability and reliability of the power amplifier operation, and reduce the failure rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained according to these drawings.

[0049] Figure 1 This is the system block diagram of the dynamic management system of this power amplifier;

[0050] Figure 2 This is the first method flowchart of the dynamic management system of this power amplifier;

[0051] Figure 3 This is the second method flowchart of the dynamic management system of this power amplifier. Detailed implementation manners

[0052] In order to be able to more clearly understand the above objects, features and advantages of the present invention, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.

[0053] Many specific details are set forth in the following description in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0054] As Figure 1 shown, the present invention discloses a dynamic management system for a power amplifier, including:

[0055] A model construction module 10 constructs a power amplifier operation state prediction model through the dynamic characteristic data and working state of the power amplifier under a preset time period;

[0056] A data acquisition and processing module 20 acquires the real-time characteristic data of the power amplifier at a preset time node through the data acquisition and processing module, performs denoising processing on the real-time characteristic data, and imports the denoised real-time characteristic data into the power amplifier operation state prediction model for prediction to obtain the working state of the power amplifier;

[0057] A regulation module 30 does not perform regulation processing on the power amplifier if the working state of the power amplifier is a normal state; if the working state of the power amplifier is an unstable state, regulation processing is performed on the power amplifier;

[0058] A communication interface module 40 realizes communication with a host computer or other systems, and performs data transmission in a wired or wireless manner;

[0059] A human-computer interaction interface 50 provides a user operation interface, including a display screen, buttons and a touch screen.

[0060] Furthermore, constructing a power amplifier operation state prediction model through the dynamic characteristic data and working state of the power amplifier under a preset time period is specifically:

[0061] Obtain the operation log of the power amplifier, and obtain the dynamic characteristic data of the power amplifier under a preset time period according to the operation log; and obtain the working state of the power amplifier at each time stamp within the preset time period according to the operation log;

[0062] Introduce a Markov chain, and define all possible state combinations for the characteristic data at each time stamp, including the characteristic data and the working state at each time stamp, to form a state space;

[0063] Initialize a two-dimensional state transition matrix, where the rows and columns respectively correspond to the states in the state space, and all elements are set to zero initially;

[0064] Read the data from the operation log of the power amplifier in chronological order, and extract the state combinations of each time stamp and its previous time stamp;

[0065] For each time stamp, analyze whether the state combination of its previous time stamp exists in the state space. If it exists, increase the transfer count by one at the corresponding position in the state transition matrix; and so on, traverse the entire state transition matrix to calculate the total transfer times of each state combination;

[0066] Divide the transfer times between each state combination by its total transfer times to obtain the transfer probability of each state combination, and generate a state transition probability matrix according to the transfer probability of each state combination;

[0067] Construct a prediction model based on a neural network, and import the state transition probability matrix into the prediction model for encoding learning. After the prediction accuracy is greater than the preset accuracy, output the operation state prediction model of the power amplifier.

[0068] It should be noted that in this step, by comprehensively applying the Markov chain and the neural network, the information in the operation log of the power amplifier is fully exploited, and an operation state prediction model that can accurately reflect its dynamic operation characteristics is constructed. This model can effectively predict the future working state of the power amplifier based on the association and transfer rules of the dynamic characteristic data and the working state of the power amplifier within a preset time period, thereby providing an important and accurate basis for subsequent dynamic management of the power amplifier, such as timely discovering potential faults and optimizing control strategies, which helps to improve the working stability and reliability of the power amplifier and ensure the normal operation of related systems.

[0069] Furthermore, perform denoising processing on the real-time characteristic data, specifically:

[0070] Introduce a fuzzy clustering algorithm, and initialize the clustering center and the fuzzy membership function to adapt to the distribution characteristics of the data;

[0071] For each piece of real-time feature data, use the fuzzy membership function to calculate its membership degree with each initial fuzzy center, and obtain the membership degrees between each piece of real-time feature data and each initial fuzzy center;

[0072] Assign each piece of real-time feature data into the initial fuzzy center with the highest membership degree respectively, and obtain several fuzzy data clusters;

[0073] After the assignment is completed, calculate the mean value of the real-time feature data belonging to each fuzzy data cluster, and use the mean value of the real-time feature data belonging to each fuzzy data cluster as the new fuzzy center;

[0074] Calculate the Euclidean distance between the new fuzzy center of each fuzzy data cluster and the initial fuzzy center; if the Euclidean distances between the new fuzzy centers of all fuzzy data clusters and the initial fuzzy centers are all less than the preset distance threshold, stop the iterative assignment;

[0075] If there is at least one case where the Euclidean distance between the new fuzzy center of a fuzzy data cluster and the initial fuzzy center is greater than the preset distance threshold, re-initialize the clustering center and membership function parameters of the fuzzy clustering model, and re-iterate the assignment of the real-time feature data until the Euclidean distances between the new fuzzy centers of all fuzzy data clusters and the initial fuzzy centers are all less than the preset distance threshold, then stop the iterative assignment;

[0076] After receiving the instruction to stop the iterative assignment, calculate the Mahalanobis distance between the real-time feature data belonging to each fuzzy data cluster and its new fuzzy center;

[0077] If the Mahalanobis distance between the real-time feature data belonging to a certain fuzzy data cluster and its new fuzzy center is greater than the preset Mahalanobis distance value, mark this real-time feature data as noise data; repeat this step until all the real-time feature data belonging to each fuzzy data cluster are judged;

[0078] Completely delete all the real-time feature data marked as noise data to obtain the denoised real-time feature data.

[0079] It should be noted that, first of all, by initializing the clustering center and the fuzzy membership function, the real-time feature data are assigned to different fuzzy clusters, and the clustering center is iteratively optimized to ensure the rationality of the data distribution; secondly, by calculating the Mahalanobis distance, the noise data are accurately identified and removed, thus significantly improving the quality and reliability of the data. This method can not only effectively remove the outliers and noise in the data, but also retain the core features of the data, providing high-quality data support for the subsequent prediction and dynamic regulation of the power amplifier operating state. At the same time, the introduction of the fuzzy clustering algorithm makes the denoising process have high robustness and adaptability, and can cope with the complex and changeable data distribution, laying a solid foundation for the intelligent management of the power amplifier.

[0080] Further, import the denoised real-time feature data into the power amplifier operating state prediction model for prediction to obtain the operating state of the power amplifier, such as Figure 2 shown, specifically:

[0081] S102. Obtain the denoised real-time feature data, and import the denoised real-time feature data into the power amplifier operating state prediction model for prediction;

[0082] S104. Through prediction, obtain the state transition probability of the power amplifier at the current preset time node; and compare the state transition probability of the power amplifier at the current preset time node with a preset probability threshold;

[0083] S106. When the state transition probability of the power amplifier at the current preset time node is not greater than the preset probability threshold, update the operating state of the power amplifier to the normal state;

[0084] S108. When the state transition probability of the power amplifier at the current preset time node is greater than the preset probability threshold, update the operating state of the power amplifier to the unstable state.

[0085] It should be noted that, first, input the high-quality denoised data into the prediction model to obtain the state transition probability at the current time node; second, through comparison with the preset probability threshold, quickly judge whether the operating state of the power amplifier is normal or unstable. This method can monitor the operating state of the power amplifier in real time and give an early warning in time when the unstable state appears, providing a decision-making basis for dynamic regulation. At the same time, the prediction method based on the state transition probability has high accuracy and reliability, can effectively avoid misjudgment and missed judgment, and improve the stability and safety of the operation of the power amplifier.

[0086] Further, if the operating state of the power amplifier is the unstable state, then perform regulation processing on the power amplifier, such as Figure 3 shown, specifically:

[0087] S202. If the operating state of the power amplifier is the unstable state, then obtain various real-time feature data of the power amplifier at the real-time time node;

[0088] S204. Calculate the difference between various real-time feature data and corresponding preset feature data; compare the difference between various real-time feature data and corresponding preset feature data with a preset difference threshold;

[0089] S206. If the difference between a certain real-time feature data and the corresponding preset feature data is not greater than the preset difference threshold, then mark the corresponding real-time feature data as normal feature data;

[0090] S208. If the difference between a certain real-time characteristic data and the corresponding preset characteristic data is greater than the preset difference threshold, mark the corresponding real-time characteristic data as drift characteristic data, and obtain the drift amplitude of the drift characteristic data, so as to obtain the drift characteristic data and its drift amplitude of the power amplifier at the real-time time node; wherein, the drift amplitude is the difference between this kind of real-time characteristic data and the corresponding preset characteristic data.

[0091] It should be noted that by calculating the difference between the real-time characteristic data and the preset characteristic data and comparing it with the preset threshold, normal characteristic data and drift characteristic data can be effectively distinguished; by obtaining the drift amplitude of the drift characteristic data, the severity of the instability state can be quantified, providing data support for targeted regulation. This method can monitor the operating state of the power amplifier in real time, quickly locate the root cause of the problem when it is unstable, improve the accuracy and efficiency of regulation, thereby effectively avoiding the performance degradation or failure of the power amplifier caused by the unstable state, and ensuring the stable operation and performance optimization of the system.

[0092] Further, if the working state of the power amplifier is an unstable state, the regulation process for the power amplifier further includes the following steps:

[0093] Obtain the regulation log of the power amplifier, and obtain the historical regulation schemes corresponding to various characteristic data drift events of the power amplifier according to the regulation log; wherein, the characteristic data drift events include event records of various drift characteristic data and their corresponding drift amplitudes;

[0094] And obtain the regulation success rates of various historical regulation schemes, screen out the historical regulation scheme with the highest regulation success rate as the optimal regulation scheme for the corresponding characteristic data drift event, and obtain the optimal regulation schemes corresponding to various characteristic data drift events;

[0095] Construct a knowledge graph, and import the optimal regulation schemes corresponding to various characteristic data drift events into the knowledge graph;

[0096] Import the drift characteristic data and its drift amplitude of the power amplifier at the real-time time node into the knowledge graph for matching, match the corresponding optimal regulation scheme, and adjust the power amplifier according to the matched optimal regulation scheme.

[0097] The real-time characteristic data includes channel gain, bandwidth, operating voltage, operating current, temperature, humidity, signal amplitude, signal frequency, and signal distortion.

[0098] Through the above method, when the power amplifier is in an unstable operating state, based on historical regulation logs and knowledge graph technology, the optimal regulation scheme can be quickly matched and implemented. Specifically, first, by analyzing historical regulation logs, the optimal regulation schemes for different characteristic data drift events are screened out and constructed into a knowledge graph; second, by matching the real-time drift characteristic data and its drift amplitude with the knowledge graph, the optimal regulation scheme can be accurately located and the regulation process can be quickly implemented. This method can not only significantly improve the efficiency and success rate of regulation, but also avoid regulation failures caused by insufficient experience or misjudgment, ensuring that the power amplifier quickly resumes stable operation in an unstable state. At the same time, the introduction of the knowledge graph makes the regulation process highly intelligent and self-adaptive, capable of dealing with complex and changeable drift events, providing technical support for the dynamic management of power amplifiers.

[0099] In addition, during the actual operation process, this system also includes:

[0100] During the operation of the power amplifier, the gain flatness of the power amplifier is continuously obtained at multiple time points;

[0101] Introduce the locally weighted regression algorithm, and define the window size and weight function of the locally weighted regression. The window size determines the range of data points used in each regression calculation, and the weight function is used to assign different weights to each data point;

[0102] For each time point, select the data points within its adjacent window, and calculate the weighted values of the data points according to the weight function;

[0103] Use the weighted least squares method to perform local regression fitting on the data points within the window to obtain the local regression result at this time point; repeat this step and traverse all time points to obtain the local regression values at each time point;

[0104] Connect the local regression values at all time points to form a gain flatness curve graph of the power amplifier within a preset time period;

[0105] Obtain the working task requirements of the power amplifier, and obtain the gain flatness threshold range of the power amplifier within a preset time period according to the working task requirements;

[0106] Define a domain in the gain flatness curve graph according to the gain flatness threshold range of the power amplifier within a preset time period;

[0107] Obtain the curve ratio of the gain flatness curve inside and outside the domain in the gain flatness curve graph to obtain the in-domain occupancy ratio of the gain flatness curve;

[0108] If the in-domain occupancy degree of the gain flatness curve is not greater than the preset occupancy degree, it indicates that the gain flatness performance of the power amplifier is unstable within the preset time period and fails to fully meet the requirements of the work task, and then a warning message is generated.

[0109] It should be noted that during the actual operation process, this system continuously obtains the gain flatness of the power amplifier at multiple time points and introduces a locally weighted regression algorithm to process and analyze the data. First, define the window size and weight function of the locally weighted regression to determine the range of data points used and the weight distribution for each regression calculation. Then, for each time point, select the data points within the adjacent window and calculate the weighted values according to the weight function. Use the weighted least squares method to perform local regression fitting on the data points within the window to obtain the local regression results for each time point. Repeat this step for all time points to obtain the local regression values for each time point. Connect these local regression values to form the gain flatness curve graph of the power amplifier within the preset time period. According to the requirements of the work task, obtain the gain flatness threshold range of the power amplifier within the preset time period and define the domain in the curve graph. By calculating the curve ratio of the gain flatness curve inside and outside the domain, the in-domain occupancy degree of the gain flatness curve is obtained. If the in-domain occupancy degree is not greater than the preset occupancy degree, it indicates that the gain flatness performance of the power amplifier is unstable within the preset time period and fails to fully meet the requirements of the work task, and the system will generate a warning message. In this way, this system can monitor the gain flatness of the power amplifier in real time, discover and warn potential unstable situations in a timely manner, so as to ensure the stable operation of the power amplifier and its performance meets the requirements of the work task.

[0110] In several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed with each other can be through some interfaces. The indirect coupling or communication connection of the devices or units can be electrical, mechanical, or other forms.

[0111] The units described above as separate components may or may not be physically separated. The components shown as units may or may not be physical units; they can be located in one place or distributed to multiple network units; some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0112] In addition, each functional unit in the embodiments of the present invention may all be integrated into one processing unit, or each unit may be separately regarded as one unit, or two or more units may be integrated into one unit; the above integrated unit may be implemented in the form of hardware, or in the form of a combination of hardware and software functional units.

[0113] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps including the above method embodiments; and the foregoing storage medium includes: removable storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks and other various media that can store program codes.

[0114] Alternatively, if the above integrated unit of the present invention is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the embodiments of the present invention, in essence or the part that contributes to the prior art, may be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the methods of the various embodiments of the present invention. And the foregoing storage medium includes: removable storage devices, ROM, RAM, magnetic disks, or optical disks and other various media that can store program codes.

[0115] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention.

Claims

1. A dynamic management system for a power amplifier, characterized in that: include: A model building module, which builds a power amplifier operation state prediction model through dynamic characteristic data and working state of the power amplifier in a preset time period; A data acquisition and processing module collects real-time characteristic data of the power amplifier at a preset time node through the data acquisition and processing module, performs denoising on the real-time characteristic data, imports the denoised real-time characteristic data into a power amplifier operation state prediction model for prediction, and obtains the working state of the power amplifier; A control module, if the working state of the power amplifier is a normal state, then no control processing is performed on the power amplifier; if the working state of the power amplifier is an unstable state, then the power amplifier is controlled; Communication interface module, which realizes communication with the host computer or other systems, and transmits data by wired or wireless means; Human-computer interaction interface, providing a user operation interface, including display screen, buttons and touch screen; The real-time feature data is subjected to denoising processing, specifically: Introduce fuzzy clustering algorithm, initialize cluster centers and fuzzy membership functions to adapt to the distribution characteristics of data; For each real-time feature data, the fuzzy membership function is used to calculate its membership with each initial fuzzy center, and the membership between each real-time feature data and each initial fuzzy center is obtained; Allocate each real-time feature data to the initial fuzzy center with the highest membership degree, and obtain several fuzzy data clusters; After the allocation is completed, the mean of the real-time feature data in each fuzzy data cluster is calculated, and the mean of the real-time feature data in each fuzzy data cluster is used as the new fuzzy center; Calculating the Euclidean distance between the new fuzzy center of each fuzzy data cluster and the initial fuzzy center; if the Euclidean distance between the new fuzzy center of each fuzzy data cluster and the initial fuzzy center is less than a preset distance threshold, stopping the iterative allocation; If there is a situation where the Euclidean distance between the new fuzzy center of at least one fuzzy data cluster and the initial fuzzy center is greater than the preset distance threshold, the cluster center and membership function parameters of the fuzzy clustering model are reinitialized, and the real-time feature data are iteratively allocated again until the Euclidean distance between the new fuzzy center of all fuzzy data clusters and the initial fuzzy center is less than the preset distance threshold, then the iterative allocation is stopped; After receiving the instruction to stop iterative allocation, the Mahalanobis distance between the real-time feature data in each fuzzy data cluster and its new fuzzy center is calculated; If the Mahalanobis distance between the real-time feature data in a fuzzy data cluster and its new fuzzy center is greater than the preset Mahalanobis distance value, the real-time feature data is calibrated as noise data; repeat this step until all the real-time feature data in each fuzzy data cluster are judged; All real-time feature data marked as noise data are completely deleted to obtain denoised real-time feature data.

2. A dynamic management system for a power amplifier according to claim 1, characterized in that: The power amplifier operation state prediction model is constructed based on the dynamic characteristic data and working state of the power amplifier in a preset time period, specifically: Obtaining an operation log of the power amplifier, and obtaining dynamic characteristic data of the power amplifier in a preset time period according to the operation log; and obtaining the working state of the power amplifier at each time stamp in the preset time period according to the operation log; The Markov chain is introduced to define all possible state combinations of the feature data at each timestamp, including the feature data and working state at each timestamp, to form a state space; Initialize a two-dimensional state transfer matrix, where the rows and columns correspond to the states in the state space, and all elements are set to zero initially; Read data from the operation log of the power amplifier in chronological order, and extract the state combination of each timestamp and its previous timestamp; For each timestamp, analyze whether the state combination of the previous timestamp exists in the state space. If so, increase the transfer count at the corresponding position in the state transfer matrix. Similarly, traverse the entire state transfer matrix and calculate the total number of transfers for each state combination; Divide the number of transitions between each state combination by its total number of transitions to obtain the transition probability of each state combination, and generate a state transition probability matrix based on the transition probability of each state combination; A prediction model is constructed based on a neural network, and the state transition probability matrix is ​​imported into the prediction model for coding learning, and a power amplifier operation state prediction model is output after the prediction accuracy is greater than a preset accuracy.

3. A dynamic management system for a power amplifier according to claim 1, characterized in that: The real-time characteristic data after denoising is imported into the power amplifier operation state prediction model to predict and obtain the working state of the power amplifier, specifically: Acquire the real-time characteristic data after denoising, and import the real-time characteristic data after denoising into the power amplifier operation state prediction model for prediction; Obtaining the state transition probability of the power amplifier at the current preset time node through prediction; and comparing the state transition probability of the power amplifier at the current preset time node with a preset probability threshold; When the state transition probability of the power amplifier at the current preset time node is not greater than the preset probability threshold, updating the working state of the power amplifier to a normal state; When the state transition probability of the power amplifier at the current preset time node is greater than a preset probability threshold, the working state of the power amplifier is updated to an unstable state.

4. A dynamic management system for a power amplifier according to claim 1, characterized in that: If the working state of the power amplifier is an unstable state, the power amplifier is regulated, specifically: If the working state of the power amplifier is an unstable state, various real-time characteristic data of the power amplifier at the real-time time node are obtained; Calculate the difference between various real-time feature data and corresponding preset feature data; Compare the difference between various real-time feature data and corresponding preset feature data with a preset difference threshold; If the difference between a certain type of real-time feature data and the corresponding preset feature data is not greater than the preset difference threshold, the corresponding real-time feature data is marked as normal feature data; If the difference between a certain type of real-time characteristic data and the corresponding preset characteristic data is greater than a preset difference threshold, the corresponding real-time characteristic data is marked as drift characteristic data, and the drift amplitude of the drift characteristic data is obtained to obtain the drift characteristic data and the drift amplitude of the power amplifier at the real-time time node; wherein the drift amplitude is the difference between the real-time characteristic data and the corresponding preset characteristic data.

5. A dynamic management system for a power amplifier according to claim 4, characterized in that: If the working state of the power amplifier is an unstable state, the power amplifier is regulated, and the following steps are also included: Obtaining a control log of the power amplifier, and obtaining historical control schemes corresponding to various characteristic data drift events occurring in the power amplifier according to the control log; wherein the characteristic data drift event includes event records of various drift characteristic data and their corresponding drift amplitudes; and obtaining the control success rates of various historical control schemes, selecting the historical control scheme with the highest control success rate as the optimal control scheme for the corresponding characteristic data drift event, and obtaining the optimal control scheme corresponding to various characteristic data drift events; Construct a knowledge graph, and import the optimal control scheme corresponding to various feature data drift events into the knowledge graph; The drift characteristic data and drift amplitude of the power amplifier at the real-time time node are imported into the knowledge graph for matching, and a corresponding optimal control scheme is obtained by matching, and the power amplifier is adjusted according to the matched optimal control scheme.

6. A dynamic management system for a power amplifier according to claim 4, characterized in that: The real-time characteristic data includes channel gain, bandwidth, operating voltage, operating current, temperature, humidity, signal amplitude, signal frequency and signal distortion.

Images