Method and system for dynamic configuration of parameters of a multi-mode radio module

Abstract

The application provides a kind of multi-mode wireless module parameter dynamic configuration method and system, comprising: obtaining the operation data of multi-mode wireless module, and carrying out real-time knowledge extraction, constructs dynamic knowledge graph;The dynamic knowledge graph is inferred by graph neural network, and an initial parameter configuration scheme is generated;Obtain the real-time demand parameter of multi-mode wireless module, input the real-time demand parameter into the first correction model, and obtain the first correction value;Obtain the real-time environmental parameter of the environment where the multi-mode wireless module is located, input the real-time environmental parameter into the second correction model, and obtain the second correction value;The initial parameter configuration scheme is corrected based on the first correction value, the second correction value, and the final parameter configuration scheme is obtained.In the present application, the defects that it is difficult to capture the dynamic correlation between parameters in real time and cannot quickly respond to real-time environmental parameters and user demand changes to optimize the parameter configuration scheme are overcome.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for dynamic parameter configuration of a multi-mode wireless module. Background Technology

[0002] Multi-mode wireless modules (such as DTU modules integrating Wi-Fi and BLE) are increasingly used in industrial automation, smart devices, and other scenarios. These modules need to dynamically adjust configuration parameters (such as operating mode, transmit power, serial port baud rate, etc.) based on real-time operating status, user needs, and environmental conditions to balance performance, power consumption, and reliability. However, existing parameter configuration methods suffer from the following technical bottlenecks:

[0003] Traditional methods rely on manually set rules or single machine learning models to process data. However, the operational data of multi-mode modules (such as serial communication status, wireless signal strength, and power consumption curves) has heterogeneous characteristics with multiple sources, and there are complex dependencies between parameters (such as the coupling effect between operating mode and channel selection, and between transmit power and temperature). Existing technologies struggle to capture the dynamic correlation between parameters in real time, resulting in configuration schemes that cannot fully reflect the actual operating status of the module.

[0004] The module's operating environment (such as temperature and electromagnetic interference in industrial settings) and user requirements (such as sudden high-speed transmission or low-power modes) are dynamic and changeable. Traditional methods typically use fixed thresholds or offline trained models for parameter adjustment, which cannot quickly respond to real-time environmental parameters (such as sudden temperature changes or increased channel interference) and changes in user requirements (such as prioritizing transmission stability), resulting in lagging configuration schemes and affecting equipment performance. Summary of the Invention

[0005] The main objective of this invention is to provide a method and system for dynamic parameter configuration of a multi-mode wireless module, aiming to overcome the shortcomings of current methods in that they are difficult to capture the dynamic correlation between parameters in real time and cannot quickly respond to changes in real-time environmental parameters and user needs to optimize parameter configuration schemes.

[0006] To achieve the above objectives, the present invention provides a method for dynamic parameter configuration of a multi-mode wireless module, comprising the following steps:

[0007] Acquire operational data from multi-mode wireless modules and perform real-time knowledge extraction to construct a dynamic knowledge graph;

[0008] Graph neural network inference is performed on the dynamic knowledge graph to generate an initial parameter configuration scheme;

[0009] Obtain the real-time demand parameters of the multi-mode wireless module, input the real-time demand parameters into the first correction model, and obtain the first correction value of the initial parameter configuration scheme;

[0010] The real-time environmental parameters of the environment in which the multi-mode wireless module is located are obtained, and the real-time environmental parameters are input into the second correction model to obtain the second correction value of the initial parameter configuration scheme.

[0011] The initial parameter configuration scheme is corrected based on the first correction value and the second correction value to obtain the final parameter configuration scheme.

[0012] Furthermore, the operating data of the multi-mode wireless module includes serial communication data and power consumption data.

[0013] Furthermore, real-time knowledge extraction is performed to construct a dynamic knowledge graph, including:

[0014] Named entity recognition technology is used to extract key entities from the runtime data; the key entities include configuration parameters, real-time status, and performance indicators.

[0015] The relationships between the key entities are identified using a relationship extraction algorithm;

[0016] A three-layer knowledge graph is constructed, comprising a parameter layer, a state layer, and an indicator layer. The parameter layer stores configuration parameters and their value ranges; the state layer stores real-time states and their parameters; and the indicator layer stores performance indicators and their parameters.

[0017] Based on the aforementioned relationships, the relationships between key entities in the three-layer knowledge graph are labeled to obtain the dynamic knowledge graph.

[0018] Furthermore, graph neural network inference is performed on the dynamic knowledge graph to generate an initial parameter configuration scheme, including:

[0019] The dynamic knowledge graph is subjected to graph embedding processing to obtain a temporally enhanced node vector representation;

[0020] The node vector representation is subjected to hierarchical graph neural network inference processing to obtain the parameter correlation matrix, constraint filtering results and performance prediction vector;

[0021] The optimal parameter solution set is obtained by performing multi-objective optimization search on the parameter correlation matrix, constraint filtering results and performance prediction vector.

[0022] The optimal parameter solution set is verified using historical data and anomaly detection processing is performed to obtain the initial parameter configuration scheme.

[0023] Further, the real-time demand parameters are input into the first correction model to obtain a first correction value for the initial parameter configuration scheme, including:

[0024] The real-time demand parameters are preprocessed by normalization to obtain a normalized demand vector;

[0025] The normalized demand vector is analyzed to obtain a multi-objective weight allocation vector.

[0026] A parameter sensitivity analysis is performed on the multi-objective weight allocation vector and the initial parameter configuration scheme to obtain the parameter adjustment priority matrix;

[0027] An incremental parameter search is performed on the parameter adjustment priority matrix to obtain a set of candidate correction parameters;

[0028] The candidate correction parameter set is fused to obtain the first correction value.

[0029] Further, the real-time environmental parameters are input into the second correction model to obtain a second correction value for the initial parameter configuration scheme, including:

[0030] Spatiotemporal feature extraction processing is performed on the real-time environmental parameters to obtain an environmental state feature vector;

[0031] An environmental risk assessment is performed on the environmental state feature vector to obtain a parameter constraint adjustment matrix;

[0032] The initial parameter configuration scheme and parameter constraint adjustment matrix are subjected to conflict detection processing to obtain a list of environmentally sensitive parameters;

[0033] Adaptive compensation calculations are performed on the list of environmentally sensitive parameters to obtain a set of environmental correction parameters;

[0034] The environmental correction parameter set is optimized to obtain the second correction value.

[0035] This invention also provides a dynamic parameter configuration system for a multi-mode wireless module, comprising:

[0036] The building unit is used to acquire the operating data of the multi-mode wireless module, perform real-time knowledge extraction, and build a dynamic knowledge graph.

[0037] The reasoning unit is used to perform graph neural network reasoning on the dynamic knowledge graph and generate an initial parameter configuration scheme.

[0038] The first correction unit is used to obtain the real-time demand parameters of the multi-mode wireless module, input the real-time demand parameters into the first correction model, and obtain the first correction value of the initial parameter configuration scheme.

[0039] The second correction unit is used to obtain real-time environmental parameters of the environment in which the multi-mode wireless module is located, and input the real-time environmental parameters into the second correction model to obtain the second correction value of the initial parameter configuration scheme.

[0040] The configuration unit is used to correct the initial parameter configuration scheme based on the first correction value and the second correction value to obtain the final parameter configuration scheme.

[0041] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of any of the methods described above.

[0042] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the methods described above.

[0043] The present invention provides a method and system for dynamic parameter configuration of a multi-mode wireless module, comprising: acquiring operational data of the multi-mode wireless module and performing real-time knowledge extraction to construct a dynamic knowledge graph; performing graph neural network inference on the dynamic knowledge graph to generate an initial parameter configuration scheme; acquiring real-time demand parameters of the multi-mode wireless module and inputting the real-time demand parameters into a first correction model to obtain a first correction value for the initial parameter configuration scheme; acquiring real-time environmental parameters of the environment in which the multi-mode wireless module is located and inputting the real-time environmental parameters into a second correction model to obtain a second correction value for the initial parameter configuration scheme; and correcting the initial parameter configuration scheme based on the first correction value and the second correction value to obtain a final parameter configuration scheme. In this invention, a dynamic knowledge graph reflecting parameter relationships is constructed through real-time knowledge extraction, an initial scheme is generated using graph neural network inference, and the initial scheme is optimized through a dual correction model by combining real-time demands and environmental parameters. This overcomes the limitations of traditional methods in data association analysis and improves the environmental adaptability, demand response speed, and multi-objective optimization capability of the configuration scheme. It overcomes the shortcomings of current methods, such as difficulty in capturing the dynamic relationship between parameters in real time and inability to quickly respond to changes in real-time environmental parameters and user needs to optimize parameter configuration schemes. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the steps of a method for dynamically configuring parameters of a multi-mode wireless module in one embodiment of the present invention;

[0045] Figure 2 This is a block diagram of the parameter dynamic configuration system of a multi-mode wireless module in one embodiment of the present invention;

[0046] Figure 3 This is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0047] The implementation, functional features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0049] Reference Figure 1 One embodiment of the present invention provides a method for dynamic parameter configuration of a multi-mode wireless module, comprising the following steps:

[0050] Step S1: Obtain the operating data of the multi-mode wireless module and perform real-time knowledge extraction to construct a dynamic knowledge graph;

[0051] Step S2: Perform graph neural network reasoning on the dynamic knowledge graph to generate an initial parameter configuration scheme;

[0052] Step S3: Obtain the real-time demand parameters of the multi-mode wireless module, and input the real-time demand parameters into the first correction model to obtain the first correction value of the initial parameter configuration scheme.

[0053] Step S4: Obtain the real-time environmental parameters of the environment in which the multi-mode wireless module is located, and input the real-time environmental parameters into the second correction model to obtain the second correction value of the initial parameter configuration scheme;

[0054] Step S5: Correct the initial parameter configuration scheme based on the first correction value and the second correction value to obtain the final parameter configuration scheme.

[0055] In one embodiment, as described in step S1 above, firstly, operational data is obtained from the multi-mode wireless module, including serial communication parameters (such as baud rate and parity settings), wireless signal indicators (such as Wi-Fi signal strength RSSI and Bluetooth connection status), power consumption information, system status information (such as operating temperature and power supply voltage), and historical configuration records. The above data comes from a wide range of sources and has different formats and characteristics, constituting multi-source heterogeneous data.

[0056] Next, real-time knowledge extraction is performed on the above data. Natural language processing techniques such as named entity recognition and relation extraction are used to extract key entities (such as "transmit power" and "operating mode") and the relationships between entities (such as "high temperature will cause a decrease in transmit power" and "AP mode limits the maximum number of connections"). At the same time, timestamps are added to the extracted knowledge to record the order of parameter changes and performance changes, for example, "at 14:20, the operating mode was switched to STA, and at 14:22, the throughput increased by 10%".

[0057] Finally, a dynamic knowledge graph is constructed based on the extracted knowledge. The graph comprises a parameter layer (storing configuration parameters and their value ranges), a state layer (recording real-time state data), and a metric layer (reflecting performance metrics). Nodes in each layer are connected by directed edges, forming a network of relationships between parameters, states, and performance. A graph database (such as Neo4j) is used to update the graph in real time, with an update frequency not exceeding 100ms, ensuring that the knowledge graph always reflects the latest operating state of the module and provides structured and dynamic knowledge support for subsequent reasoning.

[0058] As described in step S2 above, deep reasoning is performed using a graph neural network to generate an initial configuration scheme that meets basic performance requirements. First, graph embedding techniques (such as Node2Vec) are used to map nodes (parameters, states, metrics) in the knowledge graph into low-dimensional vectors, preserving node attributes while capturing the relational features between nodes. Simultaneously, a position-aware graph attention mechanism is employed to enhance the weights of temporally dependent edges (such as the latency impact of parameter changes on performance) and causal edges (such as the correlation between baud rate and bit error rate), highlighting the impact of key relationships on reasoning.

[0059] Subsequently, inference is performed using a hierarchical graph neural network: In the parameter correlation layer, a graph convolutional network (GCN) is used to calculate the correlation between parameters and identify strongly correlated parameter groups (such as "working mode" and "channel selection" needing to be configured together); in the constraint inference layer, invalid configurations that do not meet the conditions are filtered out based on hardware limitations (such as the power supply voltage range of 3.3V±0.15V) and communication protocol specifications (such as Wi-Fi security protocol requirements) (such as prohibiting STA mode without AP hotspot configuration); in the performance prediction layer, a graph neural network is used to predict the performance indicators corresponding to different parameter combinations (such as predicting the throughput under "AP mode + channel 6").

[0060] Next, candidate parameter combinations are generated based on the inference results, and Monte Carlo Tree Search (MCTS) is used to optimize multi-objective functions (such as maximizing throughput and reliability, and minimizing power consumption). Different objectives are balanced by adjusting the weight coefficients (initially set to a balanced mode, such as [0.4, 0.3, 0.3]). Finally, historical data is used to verify the feasibility of candidate schemes, eliminating configurations that lead to excessively high bit error rates (e.g., >5%) or excessive power consumption (e.g., >500mW), thereby generating an initial parameter configuration scheme, such as "STA mode + baud rate 9600bps + transmit power 15dBm".

[0061] As described in step S3 above, real-time demand parameters are obtained. These parameters come from user-defined settings (such as high-speed transmission mode, low-power priority) or task instructions issued by the upper-layer system. The demand intent is parsed through natural language processing or preset tags, and the demand is transformed into a multi-objective weight vector. For example, when the demand is "high-speed transmission mode", the weight vector can be set to [0.6, 0.2, 0.2] to highlight the priority of the throughput target.

[0062] Next, a sensitivity analysis is performed on the parameters in the initial configuration scheme. The gradient values ​​of each parameter with respect to performance indicators are calculated, and a parameter-indicator sensitivity matrix is ​​constructed to clarify the degree of impact of each parameter change on different performance indicators. For example, the sensitivity of baud rate to throughput is 0.8, indicating that increasing the baud rate can significantly increase throughput; the sensitivity of transmit power to power consumption is -0.5, meaning that reducing transmit power can effectively reduce power consumption. Combining the weight vector, an adjustment priority matrix is ​​generated to determine the order in which parameters should be adjusted, such as prioritizing the baud rate parameter that has a greater impact on key objectives.

[0063] Then, a small-step search is performed within the parameter constraint space (e.g., adjusting the baud rate by ±10%) to predict performance changes after parameter adjustment through local linear approximation, generating a candidate set of correction parameters. Finally, a reinforcement learning model (based on Q-value evaluation of historical configurations), a causal inference model (analyzing the causal effects of parameter adjustments), and an expert rule engine (e.g., security protocols must comply with the WPA2-PSK standard) are integrated to comprehensively evaluate the candidate parameter set and generate the first correction value. For example, increasing the baud rate by 20% may increase throughput by 15%, but power consumption may also increase by 8%. This result is the specific manifestation of the first correction value, used to correct the initial configuration scheme.

[0064] As described in step S4 above, the configuration scheme is further optimized to improve environmental adaptability based on the real-time environmental conditions of the wireless module. First, real-time environmental parameters are collected, including temperature, humidity, electromagnetic interference intensity, and metal obstruction rate around the antenna. Electromagnetic signals are analyzed using a sliding window Fourier transform to extract interference frequency band characteristics; a spatial interpolation algorithm is used to generate a thermal map of metal obstruction in the antenna layout to quantify the impact of the environment on signal transmission.

[0065] Next, a risk assessment is performed based on the collected environmental parameters. An environmental risk assessment model is constructed, and response strategies are set for different environmental conditions. For example, when the temperature exceeds 80°C, it is determined to be a high-temperature risk, and the upper limit of the transmission power is forcibly reduced (e.g., from 17dBm to 15dBm) to avoid overheating; when strong electromagnetic interference is detected in the 5GHz band, the channel is automatically switched to an idle channel in the 2.4GHz band (e.g., from channel 11 to channel 1). This generates a parameter constraint adjustment matrix to limit the value range of sensitive parameters and ensure that the configuration scheme adapts to environmental requirements.

[0066] Then, the initial configuration scheme is compared with the constraint adjustment matrix to detect conflicts. For example, if the transmit power in the initial scheme is 17dBm, which exceeds the upper limit set by the constraint matrix in a high-temperature environment, adjustments are required. By designing an adaptive compensation algorithm, parameters are corrected according to environmental changes, such as reducing the transmit power by 2dBm for every 10°C increase, or switching channels in real time according to interference frequency bands.

[0067] Finally, the stability of the compensated scheme is verified by using an environmental simulator (such as a high-temperature chamber to simulate a high-temperature environment and an electromagnetic interference generator to simulate signal interference) to ensure that the module can still operate reliably after the configuration adjustment, thereby generating a second correction value, such as a 2dBm reduction in transmit power and a channel switch to 1, to provide an environmental dimension optimization basis for the final scheme.

[0068] As described in step S5 above, the aim is to integrate optimization information from both demand and environment dimensions to generate a globally optimal final configuration scheme. First, a weighted summation or priority arbitration strategy is used to fuse the first correction value (demand-oriented) and the second correction value (environment-oriented). For example, when the user demand is "high-speed transmission," if environmental conditions permit, the temperature limit can be appropriately relaxed to prioritize throughput requirements; however, when the ambient temperature reaches above 90°C, a low-power mode is forcibly entered to ensure the module's safe operation, in which case environmental priority takes precedence over demand.

[0069] Next, the feasibility of the merged parameter combination is verified. This involves checking whether the configuration meets hardware constraints (e.g., module power supply current does not exceed 500mA) and communication protocol specifications (e.g., the Bluetooth network configuration command format is correct). Simultaneously, it's verified whether configuration adjustments will cause system anomalies (e.g., a continuously high NET indicator light indicates network failure). If problems are found, the adjusted correction values ​​are returned or the parameter combination is re-evaluated.

[0070] Finally, the final configuration scheme is generated, including specific parameter combinations (such as "STA mode + baud rate 115200bps + transmit power 15dBm + channel 1"), performance prediction results (such as expected throughput of 60Mbps and power consumption of 220mW), and risk warning information (such as recommending additional heat dissipation measures in high-temperature environments). This scheme comprehensively considers the module's operating status, user needs, and environmental conditions, achieving multi-dimensional dynamic optimization to ensure stable and efficient operation of the wireless module in complex scenarios.

[0071] In this embodiment, the operating data of the multi-mode wireless module includes serial communication data and power consumption data.

[0072] In one embodiment, real-time knowledge extraction is performed to construct a dynamic knowledge graph, including:

[0073] Named entity recognition technology is used to extract key entities from the runtime data; the key entities include configuration parameters, real-time status, and performance indicators.

[0074] The relationships between the key entities are identified using a relationship extraction algorithm;

[0075] A three-layer knowledge graph is constructed, comprising a parameter layer, a state layer, and an indicator layer. The parameter layer stores configuration parameters and their value ranges; the state layer stores real-time states and their parameters; and the indicator layer stores performance indicators and their parameters.

[0076] Based on the aforementioned relationships, the relationships between key entities in the three-layer knowledge graph are labeled to obtain the dynamic knowledge graph.

[0077] In one embodiment, the operational data of the multi-mode wireless module is first processed. This operational data encompasses various aspects such as serial communication records, wireless signal status, and system operating parameters, with different formats and types. To extract useful information from this complex data, named entity recognition technology is employed—a technique capable of accurately locating specific names within text. Key entities are categorized into three types: configuration parameters, real-time status, and performance metrics. Configuration parameters include manually set or modified parameters such as operating mode (e.g., AP mode, STA mode), transmit power, and serial port baud rate. Real-time status reflects the module's current instantaneous operating condition, such as current temperature, wireless signal strength (RSSI), and the number of connected devices. Performance metrics measure the module's performance, such as data transmission throughput, bit error rate during transmission, and power consumption. Through named entity recognition technology, the system can quickly and accurately identify these key entities from the operational data, laying the foundation for subsequent knowledge graph construction.

[0078] After identifying key entities, it's necessary to further clarify the inherent relationships between them. This is where relationship extraction algorithms come into play. Several relationship types are predefined, including causal relationships (where a change in one entity triggers a change in another, e.g., an increase in module operating temperature may lead to a decrease in transmission power); configuration dependencies (where certain configuration parameters are mutually restrictive, such as setting the operating mode to AP mode limiting the maximum number of connected devices); temporal relationships (describing the chronological order of entity changes, e.g., modifying the serial port baud rate first, followed by a change in the data transmission error rate); and numerical mapping relationships (indicating a numerical correspondence between two entities, e.g., a specific transmission power corresponding to a certain range of wireless signal coverage distance). The relationship extraction algorithm analyzes the contextual information of key entities in the data, as well as their frequency and patterns of occurrence, to determine the specific relationship type between entities. Furthermore, the reliability of some relationships is evaluated by counting the number of times they co-occur in the data, ensuring that the identified associations are genuine and valid.

[0079] Based on the extracted key entities and identified relationships, a knowledge graph is constructed. This knowledge graph adopts a hierarchical structure, divided into a parameter layer, a state layer, and an indicator layer. The parameter layer primarily stores configuration parameters and their value ranges. Each configuration parameter has a specific name, such as "transmit power," and specifies its adjustable value range. For example, transmit power is typically adjustable within a certain decibel-milliwatt (dBm) range. Default values ​​are also recorded for initial or reset purposes. The state layer stores real-time status and related parameters. Each real-time status node includes a status name, such as "current temperature," along with a specific value and acquisition time, such as "current temperature = 75℃, acquisition time: May 20, 2025, 14:30:00." This information clearly indicates the module's operating status at a given moment. The indicator layer stores performance indicators and their parameters. Besides the indicator name and calculation method, thresholds are set, such as setting the normal threshold for the bit error rate to less than 1%, to facilitate subsequent evaluation of whether the module's performance meets standards. These three layers are not independent of each other, but are closely connected through the relationships between entities. For example, changes in the configuration parameters of the parameter layer will affect the real-time state of the state layer, which in turn will affect the performance indicators of the indicator layer.

[0080] After completing the hierarchical construction of the knowledge graph, the nodes in the knowledge graph are labeled with relationships based on the previously identified key entity connections. During labeling, directed edges are used to represent the direction of the relationship. For example, a directed edge from the "Transmit Power" node to the "Radio Signal Strength (RSSI)" node indicates that changes in transmit power affect RSSI, and the edge is marked as a causal relationship. Each edge also includes attributes such as weight and confidence level. The weight reflects the magnitude of the relationship's influence, while the confidence level reflects the reliability of the relationship. For example, this causal relationship has a weight of 0.8 and a confidence level of 95%, indicating that transmit power has a significant impact on RSSI and that the relationship is highly reliable. Furthermore, the knowledge graph has the ability to be dynamically updated. When new operational data is detected and new relationships are identified, this information will be added to the graph in a timely manner. For relationships that are no longer valid over time, they will be deleted or marked as invalid according to the set rules. This ensures that the knowledge graph can always accurately reflect the latest status of key entities and their relationships during the operation of the multi-mode wireless module, becoming a dynamic and vibrant knowledge network that provides strong support for the reasoning and optimization of subsequent module parameter configuration.

[0081] In this embodiment, graph neural network inference is performed on the dynamic knowledge graph to generate an initial parameter configuration scheme, including:

[0082] The dynamic knowledge graph is subjected to graph embedding processing to obtain a temporally enhanced node vector representation;

[0083] The node vector representation is subjected to hierarchical graph neural network inference processing to obtain the parameter correlation matrix, constraint filtering results and performance prediction vector;

[0084] The optimal parameter solution set is obtained by performing multi-objective optimization search on the parameter correlation matrix, constraint filtering results and performance prediction vector.

[0085] The optimal parameter solution set is verified using historical data and anomaly detection processing is performed to obtain the initial parameter configuration scheme.

[0086] In this embodiment, firstly, the dynamic knowledge graph graphically stores the complex relationships between various parameters, states, and indicators during the operation of the multi-mode wireless module. However, this structured data format is difficult to use directly for algorithmic computation. Therefore, graph embedding technology is used to transform each node in the knowledge graph—that is, entities such as configuration parameters, real-time states, and performance indicators—into low-dimensional vectors that can be processed by a computer. During the transformation process, a temporal enhancement mechanism can be introduced to better capture the temporal order and causal logic of parameter changes. For example, information such as the switching time of the module's operating mode and the time of performance indicator changes after the transmission power adjustment are incorporated into the node vectors. In this way, the knowledge that originally existed in the form of nodes and edges in the graph becomes a numerical vector containing time-dimensional features. This not only preserves the relationships between entities but also highlights the temporal characteristics of data changes, providing a richer information foundation for subsequent reasoning and computation.

[0087] After obtaining the time-enhanced node vector representation, a hierarchical graph neural network is used for in-depth analysis. In the first parameter association layer, the graph neural network identifies the close relationships between different configuration parameters by calculating the similarity and correlation between node vectors. For example, operating mode and channel selection often need to be set collaboratively. These relationships are ultimately presented in the form of a parameter association matrix, clearly showing which parameters are strongly correlated and which are relatively independent. Next, in the constraint inference layer, based on the hardware limitations of the wireless module, such as the power supply voltage range, the upper limit of transmit power, and communication protocol specifications, such as Wi-Fi security protocol standards, all possible parameter combinations in the parameter association matrix are screened one by one, filtering out invalid configurations that do not meet the requirements. For example, setting to STA mode is prohibited when no AP hotspot is configured, resulting in constraint filtering results. Finally, in the performance prediction layer, the graph neural network predicts the possible performance of the module under different parameter combinations, such as throughput, power consumption, and bit error rate, based on the information processed in the previous two layers and the correspondence between parameter configurations and performance indicators in historical data. These prediction results are summarized into a performance prediction vector, providing a quantitative basis for subsequent parameter optimization.

[0088] Based on the parameter correlation matrix, constraint filtering results, and performance prediction vector obtained in the previous step, a multi-objective optimization search begins. The parameter configuration of a wireless module typically needs to simultaneously consider multiple objectives, such as maximizing data transmission throughput to meet the demands of rapid data exchange; minimizing power consumption during operation to extend device battery life; and ensuring connection reliability to reduce the bit error rate. These objectives are incorporated into the optimization considerations, and different parameter configuration schemes are continuously explored and evaluated within the effective parameter combination space that satisfies hardware and protocol constraints using algorithms such as Monte Carlo tree search. During the search process, the performance prediction vector is used to determine the performance of each scheme on each objective, and a pre-set trade-off strategy is used to comprehensively evaluate the merits of each scheme. For example, for a scheme that significantly improves throughput but consumes too much power, and a scheme that consumes less power but has only moderate throughput, the optimal parameter combination that balances multiple objectives is found based on pre-set weights or dynamically adjusted priorities, ultimately forming an optimal parameter solution set. Each parameter combination in these solution sets represents a better configuration choice under the current conditions.

[0089] After obtaining the optimal parameter solution set, historical data is used for verification and anomaly detection to ensure that these parameter combinations are feasible and reliable in actual operation. First, each parameter combination in the optimal solution set is compared with a large amount of actual data recorded during past module operation to check whether these configurations have ever experienced performance anomalies or equipment failures in historical environments. For example, the combination of a high transmit power and a specific operating mode is checked to see if it has caused module overheating and crashes in historical records. Simultaneously, anomaly detection algorithms are used to predict potential anomalies that parameter combinations may cause, analyzing whether these configurations violate implicit constraints that are not explicitly stated, such as certain parameter combinations potentially leading to a sharp increase in wireless signal interference. Through historical data verification and anomaly detection, problematic or high-risk parameter combinations are removed from the solution set, and the final selected parameter combinations constitute the initial parameter configuration scheme. This scheme comprehensively considers knowledge graph relational reasoning, multi-objective optimization, and the feasibility of actual operation, laying a solid foundation for further optimization of the configuration based on real-time needs and the environment.

[0090] In one embodiment, the real-time demand parameters are input into a first correction model to obtain a first correction value for the initial parameter configuration scheme, including:

[0091] The real-time demand parameters are preprocessed by normalization to obtain a normalized demand vector;

[0092] The normalized demand vector is analyzed to obtain a multi-objective weight allocation vector.

[0093] A parameter sensitivity analysis is performed on the multi-objective weight allocation vector and the initial parameter configuration scheme to obtain the parameter adjustment priority matrix;

[0094] An incremental parameter search is performed on the parameter adjustment priority matrix to obtain a set of candidate correction parameters;

[0095] The candidate correction parameter set is fused to obtain the first correction value.

[0096] In this embodiment, after obtaining the real-time demand parameters, since these parameters come from different sources and have different dimensions and value ranges (e.g., some parameters are in percentage form, while others are specific numerical values), they need to be normalized preprocessed to facilitate subsequent processing and comparison. The normalization process is similar to converting items of different specifications into a standard unit, transforming various real-time demand parameters into a unified numerical range, typically between 0 and 1. After this processing, all demand parameters are integrated into a normalized demand vector. Each element in this vector represents a standardized demand parameter, eliminating the differences in dimensions and scale between the original parameters, allowing subsequent analysis and calculation to treat each demand parameter more fairly and accurately.

[0097] After obtaining the normalized requirement vector, it's crucial to deeply understand the true intent behind these requirements, that is, to clarify the user's or upper-layer system's emphasis on the performance of the multi-mode wireless module. For example, when "high-speed transmission mode" appears in the requirement parameters, it should be interpreted as meaning the user values ​​data transmission throughput more; while "low power priority" indicates the user wants the module to minimize energy consumption during operation. Based on preset rules or a trained model, these requirement intentions are translated into weighted allocations for different performance objectives. For instance, if it's determined that the user prioritizes throughput, then in the multi-objective weight allocation vector, the weight corresponding to throughput will be relatively high, while the weights of other objectives such as power consumption and reliability will be correspondingly reduced. In this way, abstract requirement intentions are transformed into concrete, computationally usable multi-objective weight allocation vectors, providing clear directional guidance for subsequent parameter adjustments.

[0098] With the multi-objective weight allocation vector in place, and combined with the initial parameter configuration scheme, sensitivity analysis is performed on each parameter. The purpose of parameter sensitivity analysis is to determine the degree of influence of each parameter on different performance objectives. For example, for the throughput performance objective, changes in the baud rate parameter may have a significant impact, while for the power consumption objective, adjustments to the transmit power parameter may be more critical. By analyzing the impact of small changes in each parameter in different directions on various performance objectives, a parameter adjustment priority matrix is ​​constructed. In this matrix, each element represents the importance ranking of a parameter to a specific performance objective, with parameters of higher importance appearing earlier in the matrix. This clearly identifies which parameters should be adjusted first and how adjusting these parameters will affect the achievement of different performance objectives, thus providing a scientific basis for subsequent parameter optimization.

[0099] Based on the parameter adjustment priority matrix, an incremental parameter search begins to find parameter configurations that better meet real-time requirements. Incremental search means that the system does not make large-scale adjustments to parameters all at once, but rather explores the parameter space step by step. Starting from the initial parameter configuration, parameters with higher priority are increased or decreased slightly according to the priority order determined in the parameter adjustment priority matrix. After each adjustment, the impact on the performance target is evaluated and recorded. By continuously performing these small-step parameter adjustments and evaluating their effects, a better parameter configuration is gradually searched in the parameter space. Finally, all evaluated and well-performing parameter combinations are collected to form a candidate correction parameter set. Each parameter combination in this set represents a configuration option that may better meet real-time requirements than the initial scheme.

[0100] After obtaining the candidate correction parameter set, the final first correction value needs to be determined from these candidate schemes. Since each scheme in the candidate parameter set may have advantages and disadvantages for different performance objectives—for example, some schemes have high throughput but high power consumption, while others have low power consumption but slightly lower reliability—these factors need to be comprehensively considered for fusion processing. During the fusion process, a multi-objective weight allocation vector will be referenced, giving greater emphasis to performance objectives with higher weights. Simultaneously, practical factors such as the feasibility and stability of parameter adjustments, as well as compatibility with hardware and protocols, will be considered. Through comprehensive evaluation and trade-offs, the parameter combination that best balances all aspects of the requirements is selected from the candidate correction parameter set. The difference between this parameter combination and the initial parameter configuration scheme is the final first correction value. This first correction value will be used to correct the initial parameter configuration scheme, making the module configuration more in line with the user's real-time needs.

[0101] In one embodiment, the real-time environmental parameters are input into a second correction model to obtain a second correction value for the initial parameter configuration scheme, including:

[0102] Spatiotemporal feature extraction processing is performed on the real-time environmental parameters to obtain an environmental state feature vector;

[0103] An environmental risk assessment is performed on the environmental state feature vector to obtain a parameter constraint adjustment matrix;

[0104] The initial parameter configuration scheme and parameter constraint adjustment matrix are subjected to conflict detection processing to obtain a list of environmentally sensitive parameters;

[0105] Adaptive compensation calculations are performed on the list of environmentally sensitive parameters to obtain a set of environmental correction parameters;

[0106] The environmental correction parameter set is optimized to obtain the second correction value.

[0107] In this embodiment, after acquiring the real-time environmental parameters of the environment in which the multi-mode wireless module is located, these parameters contain rich temporal and spatial information, requiring further mining and extraction of key features. For example, the trend of ambient temperature changes over a period of time, the distribution of electromagnetic interference intensity across different frequency bands and locations, and the differences in the metal obstruction rate around the antenna with spatial location. The system uses methods such as sliding window analysis and spatial interpolation to extract feature information reflecting dynamic environmental changes and spatial characteristics from these real-time environmental parameters. This feature information is then integrated and encoded to form an environmental state feature vector. This vector comprehensively and concisely summarizes the key states of the current environment, providing an intuitive and effective data foundation for subsequent analysis and decision-making.

[0108] With the environmental state feature vector, a deep analysis of the current environmental state is conducted based on a pre-defined environmental risk assessment model to determine the potential impact of environmental factors on the operation of the wireless module. For example, excessively high ambient temperatures may cause the module to overheat, leading to performance degradation or even damage; excessive electromagnetic interference may severely affect the transmission quality of wireless signals. Based on different environmental state characteristics, corresponding risk levels are assessed, and it is determined which module configuration parameters need adjustment to mitigate risk. These assessment results are then compiled into a parameter constraint adjustment matrix. This matrix clearly defines the value range, adjustment direction, and adjustment magnitude of each configuration parameter under specific environmental conditions. For example, when the temperature exceeds a certain threshold, the matrix limits the maximum transmit power to prevent module failure due to overheating, thus providing clear rules and a basis for subsequent parameter adjustments.

[0109] After obtaining the parameter constraint adjustment matrix, the system compares and analyzes it with the initial parameter configuration scheme to identify parameters in the initial scheme that are incompatible with the current environmental conditions and may cause operational problems. Each parameter in the initial parameter configuration scheme is checked one by one to determine if it meets the constraints specified in the parameter constraint adjustment matrix. If a parameter exceeds the value range set in the matrix, or its configuration poses a risk in the current environment, then this parameter is identified as an environmentally sensitive parameter. For example, in an environment with strong electromagnetic interference, a specific communication frequency band used in the initial scheme may be severely interfered with; in this case, the frequency band parameter is considered an environmentally sensitive parameter. All these environmentally sensitive parameters are filtered out to form an environmentally sensitive parameter list, so that these parameters can be adjusted and optimized in a targeted manner later.

[0110] For parameters in the environmentally sensitive parameter list, adaptive compensation calculations are performed based on the current environmental state and the requirements of the parameter constraint adjustment matrix. Adaptive compensation calculation means that the system adjusts parameters using appropriate algorithms and strategies according to different environmental conditions and parameter characteristics. For example, for transmit power, which is listed as an environmentally sensitive parameter due to excessive temperature, its value is automatically reduced according to certain rules to decrease module heat generation; for communication frequency band parameters affected by electromagnetic interference, a relatively idle and less-interfered frequency band is automatically selected for replacement based on the distribution of interfering frequency bands. By performing such adaptive adjustments to each parameter in the environmentally sensitive parameter list, a new set of parameter values ​​is obtained, which constitute the environmental correction parameter set. The parameter configuration in this set aims to enable the module to better adapt to current environmental conditions and reduce operational risks caused by environmental factors.

[0111] After obtaining the environmental correction parameter set, it is further optimized to ensure that the final parameter configuration not only adapts to the environment but also achieves optimal performance and stability. During optimization, multiple factors are considered, such as the impact of parameter adjustments on different performance indicators of the module, the compatibility between adjusted parameters, and the feasibility and complexity of the adjustment operation. For example, if adjusting a parameter reduces environmental risk but severely impacts the module's data transmission rate, the adjustment strategy will be re-evaluated to find a better balance. By continuously evaluating, adjusting, and verifying the parameters in the environmental correction parameter set, a set of optimal parameter adjustment values ​​is finally determined. The difference between this set of adjustment values ​​and the initial parameter configuration scheme is the second correction value. This second correction value will be used to correct the initial parameter configuration scheme, enabling the wireless module to operate stably and efficiently in the current environment.

[0112] In one embodiment, after obtaining the final parameter configuration scheme, the process includes:

[0113] The parameter configuration scheme is processed by extracting parameters and adding each character from each parameter to a matrix to obtain a parameter character matrix.

[0114] Obtain the identification information of the multi-mode wireless module, extract the character attributes from the identification information, and determine the corresponding adjustment rules based on the character attributes;

[0115] Based on the adjustment rules, the parameter character matrix is ​​adjusted to obtain an adjusted character matrix;

[0116] The adjusted character matrix is ​​divided into a first matrix with three rows and two columns and a second matrix with two rows and two columns. The first matrix and the second matrix are then multiplied to obtain the target matrix.

[0117] The characters in the target matrix are combined into a communication key, which serves as the communication key between the multi-mode wireless module and external devices.

[0118] In this embodiment, after determining the final parameter configuration scheme, a parameter extraction operation is first performed on the scheme to further process and utilize this parameter information. The parameter configuration scheme includes multiple parameters such as operating mode, transmit power, and serial port baud rate, each existing in character or number form. The characters in these parameters are separated one by one and added sequentially to a matrix. This matrix stores the character information of the parameter configuration scheme in a structured manner, providing a unified data format for subsequent processing.

[0119] After constructing the parameter character matrix, the next step is to obtain the identification information of the multi-mode wireless module itself. This identification information typically includes the module's model number, production number, serial number, etc. Each character has specific attributes, such as its type (number, letter, special symbol), its position in the identification information, and its frequency of occurrence. The characters in the identification information are analyzed in depth to extract these character attributes. Then, based on a pre-defined rule base, the corresponding adjustment rules are determined according to the extracted character attributes.

[0120] After determining the adjustment rules, the parameter character matrix is ​​manipulated according to these rules. For each character in the parameter character matrix, a corresponding transformation is performed based on the corresponding adjustment rule. For example, if the adjustment rule requires converting letters to uppercase, then all lowercase letters in the matrix will be converted; if the rule specifies a particular replacement operation for numbers, then numbers that meet the conditions will be replaced with the specified values. By adjusting each character in the parameter character matrix one by one, a new matrix, the adjusted character matrix, is finally obtained. The characters in this matrix are the result of rule processing, and compared with the original parameter character matrix, both the form and content of the characters have changed to meet specific needs and security requirements.

[0121] After obtaining the adjusted character matrix, it undergoes matrix segmentation to further generate the communication key. According to pre-defined rules, the adjusted character matrix is ​​divided into a first matrix of three rows and two columns and a second matrix of two rows and two columns. This segmentation method is determined based on the algorithm and logic for generating the communication key, aiming to group character information appropriately. After segmentation, a specific multiplication calculation is performed on the two matrices. In this embodiment, the matrix multiplication can be performed directly, or the characters at corresponding positions in the two matrices can be combined or transformed according to a preset algorithm. Through a series of calculation steps, a new matrix, the target matrix, is finally obtained. This target matrix integrates the information from the two sub-matrices and generates new character combination forms through calculation.

[0122] After obtaining the target matrix through the preceding steps, the characters in the target matrix are combined sequentially. This combination process also follows specific rules; characters can be extracted sequentially from left to right or top to bottom, or arranged according to the order specified by a certain encryption algorithm. These characters are combined to form a complete string, which is the communication key. This communication key serves as a crucial basis for authentication and data encryption when the multi-mode wireless module communicates with external devices. In subsequent communication processes, only when the external device uses the same communication key for authentication and data processing can it successfully establish a secure and reliable connection with the module, thereby achieving correct data transmission and interaction, and ensuring the security and accuracy of communication.

[0123] In one embodiment, after obtaining the final parameter configuration scheme, the process includes:

[0124] The parameter configuration scheme is processed by extracting parameters and adding each character from each parameter to a matrix to obtain a parameter character matrix.

[0125] Obtain the identification information of the multi-mode wireless module, extract the numeric characters from the identification information, and generate the corresponding curve based on the numeric characters;

[0126] The curves are superimposed onto the parameter character matrix according to preset rules, and the parameter character matrix is ​​divided based on the curves to obtain multiple sub-matrices;

[0127] The curves are superimposed on each sub-matrix according to preset rules, and matrix elements in each sub-matrix that satisfy a preset distance relationship with the curves are obtained as target elements;

[0128] The target elements are combined sequentially to obtain an element combination, which serves as the communication key between the multi-mode wireless module and external devices.

[0129] In this embodiment, after determining the final parameter configuration scheme, the parameters in the scheme are first decomposed and analyzed. The parameter configuration scheme typically includes multiple specific parameters such as operating mode, transmit power, and communication frequency, each composed of different characters. These characters are separated one by one and filled into a pre-defined matrix structure in a specific order. In this way, the originally scattered parameter information is integrated into a structured parameter character matrix, providing a unified data format and operational basis for subsequent processing.

[0130] After constructing the parameter character matrix, the system acquires the identification information of the multi-mode wireless module. This information typically includes the module's model, serial number, and production batch. All numeric characters are extracted from this identification information; for example, 1, 2, 3, 4, and 5 are extracted from the serial number "ABC12345". Next, a corresponding curve is generated based on these numeric characters. Specifically, each numeric character is mapped to a specific coordinate point, and these coordinate points are connected sequentially according to the order in which the numbers appear in the identification information, forming a continuous curve. The shape and trend of this curve are determined by the specific values ​​and arrangement order of the numeric characters; therefore, the curve generated for each module is unique, containing the module's identity characteristics.

[0131] After obtaining the curve, it is superimposed onto the parameter character matrix according to preset rules. This superimposition process is similar to drawing a path on an image; the curve will cover a portion of the elements in the parameter character matrix. The preset rules include the curve's starting position, direction, and extension method within the matrix. After superimposition, the parameter character matrix is ​​segmented based on the curve's position and direction within the matrix. For example, the curve may divide the matrix into multiple regions (vertical, horizontal, or irregular), each region forming a submatrix. In this way, the originally uniform parameter character matrix is ​​divided into multiple submatrices with specific relationships, providing more refined operational units for subsequent target element extraction.

[0132] After obtaining multiple sub-matrices, the curve is superimposed onto each sub-matrix again according to preset rules. Unlike before, this superposition is performed individually for each sub-matrix, with the curve occupying a specific position within each sub-matrix. Next, matrix elements in each sub-matrix that satisfy a preset distance relationship with the curve are searched. The preset distance relationship can be elements directly adjacent to the curve, elements within a certain range from the curve, or elements satisfying a specific geometric relationship. For example, the element closest to the curve, or elements within a specific range on either side of the curve, are selected as target elements. In this way, elements meeting the conditions are filtered from each sub-matrix; these elements will serve as key data for generating the communication key.

[0133] After obtaining the target elements from each submatrix, these target elements are combined sequentially. The combination order is usually related to the division order of the submatrix and the direction of the curve in each submatrix. For example, it can be in the order of top to bottom and left to right of the submatrix, connecting the target elements in each submatrix one by one. Through this combination method, the target elements scattered in each submatrix are integrated into a complete element combination. This element combination contains the dual characteristics of parameter character matrix and module identification information, possessing high uniqueness and security. Therefore, it can be used as the authentication and data encryption key when the multi-mode wireless module communicates with external devices. In subsequent communication processes, only communication keys generated using the same method can achieve secure connection and data transmission between devices, thereby ensuring the reliability and security of communication.

[0134] Reference Figure 2 In another embodiment of the present invention, a dynamic parameter configuration system for a multi-mode wireless module is also provided, comprising:

[0135] The building unit is used to acquire the operating data of the multi-mode wireless module, perform real-time knowledge extraction, and build a dynamic knowledge graph.

[0136] The reasoning unit is used to perform graph neural network reasoning on the dynamic knowledge graph and generate an initial parameter configuration scheme.

[0137] The first correction unit is used to obtain the real-time demand parameters of the multi-mode wireless module, input the real-time demand parameters into the first correction model, and obtain the first correction value of the initial parameter configuration scheme.

[0138] The second correction unit is used to obtain real-time environmental parameters of the environment in which the multi-mode wireless module is located, and input the real-time environmental parameters into the second correction model to obtain the second correction value of the initial parameter configuration scheme.

[0139] The configuration unit is used to correct the initial parameter configuration scheme based on the first correction value and the second correction value to obtain the final parameter configuration scheme.

[0140] In this embodiment, the specific implementation of each unit in the above system embodiment is described in the above method embodiment, and will not be repeated here.

[0141] Reference Figure 3 This invention also provides a computer device, which can be a server, and its internal structure can be as follows: Figure 3 As shown, the computer device includes a processor, memory, display screen, input device, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores the data corresponding to this embodiment. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements the above-described method.

[0142] Those skilled in the art will understand that Figure 3 The structures shown are merely block diagrams of some structures related to the present invention and do not constitute a limitation on the computer devices on which the present invention is applied.

[0143] An embodiment of the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method. It is understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0144] In summary, the method and system for dynamic parameter configuration of a multi-mode wireless module provided in this embodiment of the invention include: acquiring the operating data of the multi-mode wireless module and performing real-time knowledge extraction to construct a dynamic knowledge graph; performing graph neural network inference on the dynamic knowledge graph to generate an initial parameter configuration scheme; acquiring the real-time demand parameters of the multi-mode wireless module and inputting the real-time demand parameters into a first correction model to obtain a first correction value for the initial parameter configuration scheme; acquiring the real-time environmental parameters of the environment in which the multi-mode wireless module is located and inputting the real-time environmental parameters into a second correction model to obtain a second correction value for the initial parameter configuration scheme; and correcting the initial parameter configuration scheme based on the first correction value and the second correction value to obtain a final parameter configuration scheme. In this invention, a dynamic knowledge graph reflecting parameter correlation is constructed through real-time knowledge extraction, an initial scheme is generated using graph neural network inference, and the initial scheme is optimized through a dual correction model by combining real-time demand and environmental parameters. This overcomes the limitations of data correlation analysis in traditional methods and improves the environmental adaptability, demand response speed, and multi-objective optimization capability of the configuration scheme. It overcomes the shortcomings of current methods, such as difficulty in capturing the dynamic relationship between parameters in real time and inability to quickly respond to changes in real-time environmental parameters and user needs to optimize parameter configuration schemes.

[0145] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the present invention and embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM, etc.

[0146] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0147] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for dynamically configuring parameters of a multi-mode wireless module, characterized in that, Includes the following steps: The process involves acquiring operational data from a multi-mode wireless module and performing real-time knowledge extraction to construct a dynamic knowledge graph. This includes: using named entity recognition technology to extract key entities from the operational data; key entities include configuration parameters, real-time status, and performance metrics; identifying the relationships between the key entities using a relation extraction algorithm; constructing a three-layer knowledge graph comprising a parameter layer, a status layer, and a metric layer; storing configuration parameters and their value ranges in the parameter layer; storing real-time status and its parameters in the status layer; and storing performance metrics and their parameters in the metric layer; and based on these relationships, labeling the relationships between key entities in the three-layer knowledge graph to obtain the dynamic knowledge graph. Graph neural network inference is performed on the dynamic knowledge graph to generate an initial parameter configuration scheme; The process involves: acquiring real-time demand parameters of a multi-mode wireless module and inputting these parameters into a first correction model to obtain a first correction value for the initial parameter configuration scheme; including: performing normalization preprocessing on the real-time demand parameters to obtain a normalized demand vector; parsing the normalized demand vector to obtain a multi-objective weight allocation vector; performing parameter sensitivity analysis on the multi-objective weight allocation vector and the initial parameter configuration scheme to obtain a parameter adjustment priority matrix; performing incremental parameter search on the parameter adjustment priority matrix to obtain a candidate correction parameter set; and fusing the candidate correction parameter set to obtain the first correction value. The process involves: acquiring real-time environmental parameters of the environment in which the multi-mode wireless module is located; inputting these real-time environmental parameters into a second correction model to obtain a second correction value for the initial parameter configuration scheme; including: performing spatiotemporal feature extraction processing on the real-time environmental parameters to obtain an environmental state feature vector; performing environmental risk assessment on the environmental state feature vector to obtain a parameter constraint adjustment matrix; performing conflict detection processing on the initial parameter configuration scheme and the parameter constraint adjustment matrix to obtain a list of environmentally sensitive parameters; performing adaptive compensation calculation on the list of environmentally sensitive parameters to obtain an environmental correction parameter set; and optimizing the environmental correction parameter set to obtain the second correction value. The initial parameter configuration scheme is corrected based on the first correction value and the second correction value to obtain the final parameter configuration scheme. This includes: using a weighted summation or priority arbitration strategy to fuse the first correction value and the second correction value, and performing a feasibility verification on the fused parameter combination; if there are problems, the correction value is adjusted or the parameter combination is re-evaluated; finally, the final configuration scheme is generated, which includes specific parameter combinations, performance prediction results, and risk warning information.

2. The method for dynamic parameter configuration of a multi-mode wireless module according to claim 1, characterized in that, The operating data of the multi-mode wireless module includes serial communication data and power consumption data.

3. The method for dynamic parameter configuration of a multi-mode wireless module according to claim 1, characterized in that, Graph neural network inference is performed on the dynamic knowledge graph to generate an initial parameter configuration scheme, including: The dynamic knowledge graph is subjected to graph embedding processing to obtain a temporally enhanced node vector representation; The node vector representation is subjected to hierarchical graph neural network inference processing to obtain the parameter correlation matrix, constraint filtering results and performance prediction vector; The optimal parameter solution set is obtained by performing multi-objective optimization search on the parameter correlation matrix, constraint filtering results and performance prediction vector. The optimal parameter solution set is verified using historical data and anomaly detection processing is performed to obtain the initial parameter configuration scheme.

4. A dynamic parameter configuration system for a multi-mode wireless module, characterized in that, include: A construction unit is used to acquire operational data of a multi-mode wireless module and perform real-time knowledge extraction to construct a dynamic knowledge graph. This includes: using named entity recognition technology to extract key entities from the operational data; key entities include configuration parameters, real-time status, and performance indicators; identifying the relationships between the key entities using a relation extraction algorithm; constructing a three-layer knowledge graph comprising a parameter layer, a status layer, and an indicator layer; storing configuration parameters and their value ranges in the parameter layer; storing real-time status and its parameters in the status layer; and storing performance indicators and their parameters in the indicator layer; and labeling the relationships between key entities in the three-layer knowledge graph based on the relationships to obtain the dynamic knowledge graph. The reasoning unit is used to perform graph neural network reasoning on the dynamic knowledge graph and generate an initial parameter configuration scheme. The first correction unit is used to acquire real-time demand parameters of the multi-mode wireless module, input the real-time demand parameters into the first correction model, and obtain a first correction value for the initial parameter configuration scheme; including: performing normalization preprocessing on the real-time demand parameters to obtain a normalized demand vector; performing demand intent parsing on the normalized demand vector to obtain a multi-objective weight allocation vector; performing parameter sensitivity analysis on the multi-objective weight allocation vector and the initial parameter configuration scheme to obtain a parameter adjustment priority matrix; performing incremental parameter search on the parameter adjustment priority matrix to obtain a candidate correction parameter set; and performing fusion processing on the candidate correction parameter set to obtain the first correction value. The second correction unit is used to acquire real-time environmental parameters of the environment in which the multi-mode wireless module is located, and input the real-time environmental parameters into the second correction model to obtain a second correction value for the initial parameter configuration scheme; including: performing spatiotemporal feature extraction processing on the real-time environmental parameters to obtain an environmental state feature vector; performing environmental risk assessment on the environmental state feature vector to obtain a parameter constraint adjustment matrix; performing conflict detection processing on the initial parameter configuration scheme and the parameter constraint adjustment matrix to obtain a list of environmentally sensitive parameters; performing adaptive compensation calculation on the list of environmentally sensitive parameters to obtain an environmental correction parameter set; and optimizing the environmental correction parameter set to obtain the second correction value; The configuration unit is used to correct the initial parameter configuration scheme based on the first correction value and the second correction value to obtain the final parameter configuration scheme; including: using a weighted summation or priority arbitration strategy to fuse the first correction value and the second correction value, and performing a feasibility verification on the fused parameter combination; if there is a problem, returning to adjust the correction value or re-evaluate the parameter combination; finally, generating the final configuration scheme, which includes specific parameter combinations, performance prediction results and risk warning information.

Images