Intelligent router automatic configuration method and system based on artificial intelligence

Abstract

The invention provides an intelligent router automatic configuration method and system based on artificial intelligence, and the method comprises the steps: directly collecting an original data set from various sensors and equipment ends, and carrying out the preprocessing; analyzing the environment feature data set, and generating a targeted router configuration scheme based on an analysis result by using an adaptive multi-arm machine algorithm; designing a globally optimized objective function according to the targeted router configuration scheme to obtain an optimized router configuration scheme; adaptive feedback learning is carried out on the optimized router configuration scheme based on real-time feedback of the network, and a configuration strategy is dynamically adjusted; the device configuration is dynamically adjusted based on real-time feedback and the dynamic score. According to the method, the defects of static configuration, lack of environmental perception and multi-device cooperation in the prior art are overcome, the network configuration is continuously optimized through a self-adaptive learning mechanism, and a more intelligent and dynamic network configuration scheme with self-optimization capability is provided.

Description

Technical Field

[0001] The present invention belongs to the technical field of artificial intelligence, and particularly relates to an intelligent router automatic configuration method and system based on artificial intelligence. Background Art

[0002] With the wide application of intelligent devices and Internet of Things technologies, the complexity and requirements of home and enterprise networks are constantly increasing. Nowadays, Wi-Fi routers have become one of the core devices indispensable in these network environments. However, although the existing routers and network configuration technologies on the market have made significant progress, traditional routers still have many problems when facing increasingly complex and dynamically changing network environments, mainly reflected in the following aspects:

[0003] Static configuration and limited adaptability: Most existing routers rely on static configuration and manual intervention. The configuration process is usually based on default settings or manual adjustment by users. Although some high-end routers provide basic automatic configuration functions (such as automatically selecting channels), these configurations are still relatively limited and cannot effectively cope with complex and dynamically changing network environments. For example, factors such as signal interference, dense device distribution, and network topology changes may cause unstable network performance, and the existing technologies cannot adjust the configuration in real time to adapt to these changes.

[0004] Lack of perception of complex environments: Existing routers generally rely on simple hardware settings and signal strength as the basis for performance optimization and lack in-depth perception of the physical environment. In complex wireless environments, factors such as reflection of obstacles, multipath effects, and electromagnetic interference have a significant impact on signal quality. However, the existing technologies fail to make full use of environmental perception technologies to dynamically adjust the router's configuration, resulting in poor signal coverage and network stability, especially in environments with severe signal shielding or a large number of devices.

[0005] Lack of single-device optimization and multi-device collaboration: With the popularization of smart home and Internet of Things devices, modern homes and offices usually use multiple routers, access points (APs), and devices for linkage. However, the existing technologies usually focus on the optimization of a single device and ignore how to achieve collaborative optimization among multiple devices. In an environment with multiple router deployments, problems such as interference between devices and uneven bandwidth allocation often lead to unstable network performance. The existing technologies fail to effectively support dynamic coordination and resource sharing among multiple devices, resulting in the network being difficult to maintain the best performance under high load or frequent changes.

[0006] Lack of self - learning and optimization mechanism: Existing routers often fail to self - optimize based on historical data or real - time feedback when facing changing network loads and device requirements. The configuration adjustments of many routers rely on manual intervention or simple static algorithms, lacking the ability of continuous learning and adaptation. Traditional optimization strategies cannot flexibly adapt to the dynamic changes of the environment. Especially under different time periods and different network requirements, they cannot automatically adjust parameters such as the bandwidth, channel, and power of the router to meet user needs.

[0007] Therefore, although the existing technology has achieved certain results in basic wireless network configuration, there are still many deficiencies. Especially in a dynamic network environment, it cannot achieve real - time, automatic, and global optimization. To address these issues, the existing technology requires a more intelligent and automated solution that can not only sense environmental changes but also perform self - learning and optimization based on multi - device collaboration to ensure the stability and efficiency of network performance. Summary of the Invention

[0008] The object of the present invention is to design an intelligent router automatic configuration method and system based on artificial intelligence. The present invention not only solves the defects of static configuration, lack of environmental perception, and multi - device collaboration existing in the prior art, but also continuously optimizes the network configuration through an adaptive learning mechanism, providing a more intelligent, dynamic, and self - optimizing network configuration solution.

[0009] To achieve the above object, in the first aspect of the present invention, an intelligent router automatic configuration method based on artificial intelligence is provided. The method includes:

[0010] S1. Directly collect the original data set from various sensors and device terminals, and perform pre - processing to obtain an environmental feature data set; the original data set includes signal quality, network topology, environmental signal data, and physical environment information; the environmental feature data set includes signal quality, interference between devices, and bandwidth utilization rate;

[0011] S2. Analyze the environmental feature data set, and use the adaptive multi - armed bandit algorithm to generate a targeted router configuration plan based on the analysis results; wherein, the router configuration plan includes bandwidth allocation, channel selection, and device scheduling;

[0012] S3. Design a globally optimized objective function according to the targeted router configuration plan to maximize bandwidth utilization rate, channel quality, and the effectiveness of device scheduling, and minimize interference. At the same time, construct a multi - dimensional resource competition matrix based on the interference degree, distance, and bandwidth overlap degree between devices to quantify the interference and competition between devices. At the same time, based on the globally optimized objective function, use the incremental optimization gradient descent algorithm to perform global optimization to obtain an optimized router configuration plan;

[0013] S4. Perform adaptive feedback learning on the optimized router configuration scheme based on network-based real-time feedback, and dynamically adjust the configuration strategy;

[0014] S5. Monitor the network status in real time, collect the key performance indicators of each device in the network, dynamically score the devices by combining historical data and target indicators, and dynamically adjust the device configuration based on real-time feedback and dynamic scoring to ensure that the network always operates efficiently under changing environmental conditions.

[0015] Further, the environmental signal data includes the signal strength, channel quality, interference situation, and noise received by each device;

[0016] The preprocessing is weighted moving average filtering, and the specific processing process is as follows:

[0017] Smooth the signal strength through weighted averaging, where the weights are set based on the device location and the attenuation function of the signal strength;

[0018] Extract key features from the environmental data:

[0019] The signal quality represents the signal-to-noise ratio between device i and device j, and is used to measure the communication quality between devices;

[0020] The interference between devices is the interference degree evaluation between devices in the network;

[0021] The bandwidth utilization rate is to calculate the bandwidth occupancy rate of device i;

[0022] Form a data set by extracting key features and perform principal component analysis dimensionality reduction processing to reduce redundant features and retain key influencing factors, obtaining an environmental feature data set.

[0023] Further, the analysis of the environmental feature data set is as follows:

[0024] Evaluate the environmental feature data set through an adaptive Gaussian mixture model to identify different environmental states of the network; among them, the adaptive Gaussian mixture model uses maximum likelihood estimation to model the environmental feature data set and obtains the mean μ of each category i and variance σ i , thereby obtaining the probability distribution of different environmental states.

[0025] Further, generate a targeted router configuration scheme using the adaptive multi-armed bandit algorithm based on the analysis results, and the specific processing process is as follows:

[0026] For each environmental state, a configuration scheme is correspondingly configured. The configuration scheme includes bandwidth allocation, channel selection, and device scheduling. The adaptive multi-armed bandit algorithm dynamically adjusts according to the real-time feedback of each configuration scheme to maximize network performance. The update rule of the adaptive multi-armed bandit algorithm is as follows:

[0027]

[0028] Among them, Q i (t) represents the return estimate of the current configuration scheme of state S i , R i (t) is the real-time feedback, and n i (t) is the number of times this configuration is selected.

[0029] Furthermore, the objective function of the global optimization is expressed as:

[0030]

[0031] Among them, Util(B i ) represents the bandwidth utilization rate of device i; Qual(C i ) represents the channel quality of device i; M ij represents the resource competition degree between device i and device j; α, β, γ respectively represent the weight coefficients in the objective function, controlling the impacts of bandwidth, channel quality, and interference respectively;

[0032] Among them, the resource competition degree is determined based on the interference degree I i,j between device i and device j, the distance D i,j between device i and device j, and the bandwidth overlap degree L i,j between device i and device j.

[0033] Furthermore, the optimized router configuration scheme needs to be subjected to the first regularization process, which is used to balance configuration differences and ensure the balanced allocation of network resources; the first regularization term of the first regularization is constrained based on the average value of all device configurations.

[0034] Furthermore, the real-time feedback includes network performance metrics and environmental changes;

[0035] Then, adaptive feedback learning is performed on the optimized router configuration scheme based on the real-time feedback of the network, and the configuration strategy is dynamically adjusted. The specific processing process is as follows:

[0036] Based on the difference between the current feedback R i (t) and the expected feedback R target , the feedback adjustment factor of each device configuration is determined;

[0037] Combined with the feedback adjustment factors configured for each device, the configuration strategy of each device is updated through an adaptive learning algorithm; among them, in order to prevent excessive update of configuration adjustment, a second regularization term is introduced during the update process to balance the resource allocation differences between devices and make the resource allocation more balanced;

[0038] Generate an optimized configuration plan set after adjustment.

[0039] Furthermore, the dynamic scoring is based on a real-time performance scoring function, expressed as:

[0040] P i (t) = ω i (t) · [R i (t) - R target +) 1 - ω i (t)) · P i (t - 1)

[0041] Wherein, P i (t) represents the real-time performance score of device i at time t; ω i (t) represents the dynamic weighting coefficient of device i at time t, reflecting the importance of device configuration; R i (t) represents the real-time feedback of device i; R target represents the target feedback value; P i (t - 1) represents the performance score of device i at the previous moment.

[0042] Furthermore, the feedback adjustment factors for each device configuration are adjusted based on the real-time performance scoring function; based on the adjusted feedback adjustment factors, the configuration plan set is updated within each time period.

[0043] In the second aspect of the present invention, an intelligent router automatic configuration system based on artificial intelligence is provided, and the system includes:

[0044] A signal acquisition unit, used to directly collect the original data set from various sensors and device terminals, and perform preprocessing to obtain an environmental feature data set; the original data set includes signal quality, network topology, environmental signal data, and physical environment information; the environmental feature data set includes signal quality, interference between devices, and bandwidth utilization rate;

[0045] A signal analysis unit, used to analyze the environmental feature data set, and generate a targeted router configuration plan based on the analysis result using an adaptive multi-armed bandit algorithm; wherein, the router configuration plan includes bandwidth allocation, channel selection, and device scheduling;

[0046] Configure an analysis unit to design a globally optimized objective function based on a targeted router configuration scheme, maximize bandwidth utilization, channel quality, and the effectiveness of device scheduling, and minimize interference. At the same time, construct a multi-dimensional resource competition matrix based on the interference degree, distance, and bandwidth overlap degree between devices to quantify the interference and competition between devices. Then, based on the globally optimized objective function, use the incremental optimization gradient descent algorithm to perform global optimization to obtain an optimized router configuration scheme;

[0047] Configure an adjustment unit to perform adaptive feedback learning on the optimized router configuration scheme based on the real-time feedback of the network and dynamically adjust the configuration strategy;

[0048] Configure a dynamic configuration unit to monitor the network state in real time, collect the key performance indicators of each device in the network, dynamically score the devices by combining historical data and target indicators, and dynamically adjust the device configuration based on the real-time feedback and dynamic scoring to ensure that the network always operates efficiently under changing environmental conditions.

[0049] The beneficial technical effects of the present invention are at least as follows:

[0050] (1) The present invention first introduces a variety of environmental perception sensors (such as wireless spectrum detectors, device positioning sensors, temperature and humidity sensors, etc.), enabling the router to perceive changes in the network environment in real time, including factors such as signal interference, device density, and wall reflection. The environmental data collected by these sensors is used to drive the intelligent adjustment of router configuration, thus solving the problem of the lack of sensitivity and adaptability to environmental changes in the prior art. For example, when the router perceives a signal interference source or uneven device distribution, the system will automatically adjust the channel, power, and bandwidth allocation to ensure signal quality and network stability.

[0051] (2) The present invention realizes the collaborative optimization between multiple routers, access points, and terminal devices through a distributed learning framework (such as federated learning), solving the limitation of only relying on single-device optimization in the prior art. Each device can perform local optimization according to its own network state and load conditions, and share and feedback the optimization results through the network to ensure the optimal performance of the entire network. Specifically, the system dynamically adjusts the allocation of network resources according to the real-time data of multi-device cooperation through a reinforcement learning algorithm, avoiding bandwidth contention and signal interference between devices, and realizing resource coordination and optimization within the entire network.

[0052] (3) The present invention introduces reinforcement learning and adaptive control algorithms. The system can self-learn based on historical data and real-time network feedback, and gradually optimize the router's configuration strategy. Over time, the system can automatically adapt to different network requirements and environmental changes. It can not only cope with instantaneous network load fluctuations but also perform more efficient optimization according to long-term usage patterns and device behaviors. For example, when the network load is high or the device is frequently used, the system can predict in advance and allocate bandwidth for these periods to ensure that network performance is not affected during peak periods. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The present invention will be further described with reference to the accompanying drawings. However, the embodiments in the drawings do not constitute any limitation to the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to the following drawings.

[0054] Figure 1 It is a flowchart of the automatic configuration method of the intelligent router based on artificial intelligence according to the present invention.

[0055] Figure 2 It is a framework diagram of the automatic configuration system of the intelligent router based on artificial intelligence according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0056] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are only used to explain the present invention and should not be construed as a limitation to the present invention.

[0057] In one or more embodiments, as Figure 1 shown, an automatic configuration method of an intelligent router based on artificial intelligence is disclosed. The method includes the following steps S1 - S5:

[0058] S1. Collect the original data set directly from various sensors and device terminals and perform preprocessing to obtain the environmental feature data set; the original data set includes signal quality, network topology, environmental signal data, and physical environment information; the environmental feature data set includes signal quality, interference between devices, and bandwidth utilization rate.

[0059] Specifically, the input of this step is the original data set E raw collected directly from various sensors and device terminals, and this data set includes the following types of information:

[0060] Device information: basic information of routers, access points (APs), and terminal devices, including MAC addresses, device types, connection rates, bandwidth occupancy, and current load, etc.

[0061] Network topology: The connection method and location of each device in the network (such as physical connection, distance between devices, stability of device connection, etc.).

[0062] Environmental signal data: Signal strength, channel quality, interference situation, noise level, etc. received by each device.

[0063] Physical environment information: Such as the impact of physical obstacles such as walls, buildings, and furniture on signal propagation, reflecting the relationship between the device location and the physical environment.

[0064] Furthermore, these data are collected through the built-in sensors of the router, the access point detection function, and the feedback of terminal devices. Although the data is collected in real time, there are usually situations of noise, loss, or abnormal fluctuations.

[0065] The original data usually contains noise or abnormal fluctuations caused by signal interference. To ensure the validity of the data, we first smooth the key information such as signal strength and device load. The weighted moving average filtering method is used to denoise the signal data. Specifically, for the signal strength S i,t (signal strength of device i at time t), we perform smoothing through weighted average, and the weight w k is set according to factors such as the distance of the device and signal attenuation:

[0066]

[0067] where, is the denoised signal strength of device i at time t, S i,t-k is the original signal strength at the past K moments, and w k is the weight set based on the attenuation function of the device location and signal strength. This method helps to eliminate fluctuations within a short period and provides stable signal strength data.

[0068] Next, several key features are extracted from the environmental data. These features help to quantify information such as signal quality, interference between devices, and bandwidth utilization in the environment, facilitating the formulation of subsequent optimization decisions. The key features include:

[0069] Signal quality (SNR): The signal-to-noise ratio (SNR) between device i and device j, used to measure the communication quality between devices. The formula is as follows:

[0070]

[0071] where, is the signal power between device i and device j, is the noise power between the two devices. Connections with poor signal quality will affect network performance and need special attention.

[0072] Inter-device interference: The formula for evaluating the degree of interference between devices in the network is as follows:

[0073]

[0074] Where α and β are weight coefficients, representing the contributions of signal strength and noise power to the interference degree respectively. This formula helps to quantify the interference, so as to carry out targeted optimization configuration.

[0075] Bandwidth utilization rate: Calculate the bandwidth occupancy rate of device i:

[0076]

[0077] Where B i is the bandwidth utilization rate of device i, U i is the bandwidth actually used by device i, and C i is the maximum bandwidth of device i. Devices with too high bandwidth utilization rate may become bottlenecks and need to optimize the bandwidth allocation.

[0078] Furthermore, finally, after feature extraction, we obtain a dataset E features containing multiple useful features, and perform principal component analysis (PCA) dimensionality reduction processing to reduce redundant features and retain key influencing factors. This processing will greatly improve the computational efficiency and optimization accuracy of subsequent steps.

[0079] E features = PCA(SNR i,j , I i,j , B i ) (5)

[0080] At this time, E features represents a dimensionality-reduced environmental feature dataset.

[0081] S2. Analyze the environmental feature dataset, and generate a targeted router configuration scheme based on the analysis results; wherein, the router configuration scheme includes bandwidth allocation, channel selection, and device scheduling.

[0082] Specifically, first, evaluate the environmental state data based on the adaptive Gaussian mixture model (GMM) to identify different working states S 1 , S 2 , …, S n of the network. Specifically, use the maximum likelihood estimation (MLE) to model the data, and obtain the mean μ i and variance σ i of each category, so as to obtain the probability distribution of different environmental states. The formula is as follows:

[0083]

[0084] Through clustering analysis, we identify the working state of the current network and provide a basis for subsequent configuration decisions.

[0085] Furthermore, after identifying the environmental state, we use the Adaptive Multi-Armed Bandit algorithm (AMAB) to generate targeted configuration solutions. Each environmental state S i corresponds to a configuration solution C i , including parameters such as bandwidth allocation, channel selection, and device scheduling. The AMAB algorithm dynamically adjusts according to the real-time feedback of each configuration solution to maximize network performance. The update formula of the AMAB algorithm is:

[0086]

[0087] where Q i (t) represents the reward estimate of the current configuration solution for state S i , R i (t) is the real-time feedback, and n i (t) is the number of times this configuration is selected. Over time, the system continuously optimizes the configuration solution according to the real-time feedback.

[0088] Furthermore, combining state evaluation and real-time feedback, this step will generate a series of configuration decisions. Specifically, it includes:

[0089] Bandwidth allocation: According to the bandwidth utilization rate B i,j between devices, adjust the bandwidth allocation strategy to avoid bandwidth bottlenecks.

[0090] Channel selection: According to the interference degree I i,j between devices and the signal quality, dynamically select channels to avoid channel conflicts.

[0091] Device scheduling: Schedule the device connection order and transmission strategy to reduce interference and improve the overall network efficiency.

[0092] The output of this step is the optimized configuration solution set C opt , which contains the configuration solutions such as bandwidth, channel, and scheduling for each device, and is directly passed to the control device for implementation.

[0093] S3. Design a globally optimized objective function according to the targeted router configuration solution to maximize the bandwidth utilization rate, channel quality, and effectiveness of device scheduling, and minimize interference. At the same time, construct a multi-dimensional resource competition matrix based on the interference degree, distance, and bandwidth overlap degree between devices to quantify the interference and competition between devices. At the same time, use the incremental optimization gradient descent algorithm based on the globally optimized objective function for global optimization to obtain the optimized router configuration solution.

[0094] Specifically, the input of this step comes from the output of step two, that is, the optimized configuration scheme set C opt , where the configuration scheme C of each device i i = (B i , C i , D i ), including:

[0095] B i : The bandwidth allocation of device i;

[0096] C i : The channel selection of device i;

[0097] D i : The scheduling strategy of device i.

[0098] These configuration schemes consider factors such as the environmental status and network load among devices, and are prepared to provide a basis for subsequent multi-device collaborative optimization and resource sharing.

[0099] Furthermore, by multi-device collaboration and resource sharing, the overall network performance is optimized. The goal is to maximize the bandwidth utilization, channel quality, and effectiveness of device scheduling, and minimize interference. We propose a globally optimized objective function that combines the bandwidth utilization, channel quality, and interference among devices. The objective function is:

[0100]

[0101] where Util(B i ): The bandwidth utilization of device i; Qual(C i ): The channel quality of device i; M ij : The resource competition degree between device i and device j (composed of interference, distance, and bandwidth overlap); α, β, γ: The weight coefficients in the objective function, which respectively control the influences of bandwidth, channel quality, and interference.

[0102] Among them, the resource competition among devices affects the network performance. Therefore, we introduce a multi-dimensional resource competition matrix M ij , which is used to quantify the interference and competition among devices. The calculation of the matrix is based on the following factors:

[0103] I i,j : The interference degree between device i and device j;

[0104] D i,j : The distance between device i and device j;

[0105] L i,j : The bandwidth overlap degree between device i and device j.

[0106] Resource competition matrix M ij is:

[0107] M ij = α 1 ·I i,j + α 2 ·D i,j + α 3 ·L i,j (9)

[0108] where α 1 , α 2 , α 3 are weight coefficients that respectively control the impacts of interference, distance, and bandwidth overlap degree.

[0109] Furthermore, after obtaining the resource competition matrix, an incremental optimization gradient descent algorithm is adopted for global optimization to adjust the bandwidth, channels, and scheduling strategies of the devices to ensure the collaborative optimization of the network. The update formula is:

[0110]

[0111] where C p (t): the configuration of device i at time t; η: the learning rate that controls the update step size; is the gradient of the configuration of device i, representing the configuration adjustment direction. This optimization algorithm is iteratively adjusted based on the objective function to ensure that the configuration of each device tends to the global optimal solution.

[0112] Furthermore, to avoid over-optimization of device configurations, resulting in excessive resource occupancy of some devices, a first regularization term is introduced to balance the configuration differences and ensure the balanced allocation of network resources. The regularization term is:

[0113]

[0114] where C avg is the average value of all device configurations; λ: the regularization coefficient that controls the regularization strength.

[0115] Furthermore, the output of this step is the optimized configuration scheme set C ′ opt , which contains the bandwidth allocation, channel selection, and scheduling strategies of each device, after global optimization and regularization processing. This configuration scheme set will be transmitted as a control signal to the network devices to ensure resource sharing and optimized collaboration among multiple devices, improving the network performance.

[0116] Through the multi-device collaboration optimization and resource sharing solution in this step, we have achieved efficient optimization of bandwidth, channel selection, and scheduling strategies among devices. This solution solves the resource allocation and interference problems in multi-device collaboration through innovative objective functions, multi-dimensional resource competition matrices, and incremental optimization algorithms, improving the overall performance and stability of the network. Finally, the optimized configuration solution set provides effective support for device scheduling and resource management in practical applications.

[0117] S4. Perform adaptive feedback learning on the optimized router configuration solution based on network-based real-time feedback, and dynamically adjust the configuration strategy.

[0118] Specifically, the input of this step comes from the output of step three, that is, the optimized configuration solution set C ′ opt , where the configuration solution C ′ ′ of each device i = (B i ′ , C i ′ , D i ′ ), including:

[0119] B i ′ : The bandwidth allocation of device i after collaborative optimization;

[0120] C i ′ : The channel selection of device i after collaborative optimization;

[0121] D i ′ : The scheduling strategy of device i after collaborative optimization.

[0122] These configuration solution sets C ′ opt represent the resource allocation state reached by the system after multi-device collaborative optimization, and will serve as the basis for adaptive learning and optimization feedback.

[0123] Furthermore, in order to achieve dynamic adaptive optimization of the system, we need to adjust the system configuration based on network-based real-time feedback. The feedback mechanism is based on the following factors:

[0124] Network performance metrics, such as throughput, latency, stability, etc.;

[0125] Environmental changes, such as device load, bandwidth changes, channel quality fluctuations, etc.

[0126] Based on these feedback data, we propose an optimization mechanism based on an adaptive feedback learning algorithm, aiming to dynamically adjust the configuration strategy. We use an update formula with adaptive weights to adjust the optimization degree of each device configuration.

[0127] Set the feedback adjustment factor for each device configuration as F i (t), which is determined by the difference between the current feedback R i (t) and the desired feedback R target :

[0128] F i (t) = F i (t - 1)+α i ·(R i (t)-R target ) (12)

[0129] Where, F i (t): The feedback adjustment factor of device i's configuration, the value at time t; R i (t): The real-time feedback of device i at time t, representing network performance (such as throughput, latency, etc.); R target : The target feedback value (such as the desired throughput or latency); α i : The adjustment step of device i, controlling the learning rate. This formula iteratively updates the feedback adjustment factor of the device configuration, making the device configuration gradually approach the optimal solution, thereby continuously improving the network performance.

[0130] Furthermore, combined with the feedback adjustment factor, we update the configuration strategy of each device through an adaptive learning algorithm. To achieve optimal resource sharing among multiple devices, we introduce an optimization strategy update function, which combines the configuration adjustment factor F i (t) of device i with the current configuration C i ′ (t) for dynamic optimization. The optimization function is as follows:

[0131] C i ′ (t + 1)=C i ′ (t)+γ i ·F i (t) (13)

[0132] Where, C i ′ (t): The configuration of device i at time t; γ i : The adjustment coefficient of device i, controlling the intensity of configuration update; F i (t): The feedback adjustment factor of device i, calculated from the difference between the real-time feedback and the target feedback.

[0133] Through this function, the configuration of the device is gradually adjusted over time and in response to real-time feedback, ensuring that the device configuration is continuously optimized to respond to network status and environmental changes.

[0134] Furthermore, to prevent excessive configuration adjustment, which may lead to network performance fluctuations or some devices over-occupying resources, we introduce a regularization term in the optimization process. This regularization term is used to balance the resource allocation differences between devices, making the resource allocation more balanced. The definition of the second regularization term is:

[0135]

[0136] where C i ′ (t): The configuration of device i at time t; The average value of all device configurations; λ i : The regularization coefficient of device i, which controls the regularization strength to avoid excessive configuration adjustment. The regularization term will be added to the optimization objective to prevent the configuration of some devices from deviating too much from the equilibrium value and ensure the reasonable allocation of system resources.

[0137] Furthermore, the output of this step is the adjusted optimized configuration scheme set C″ opt , which includes the bandwidth allocation, channel selection, and scheduling strategy of each device at the current time t. These configurations are dynamically adjusted according to real-time feedback and the regularization mechanism to ensure that the network maintains optimal performance in a changing environment.

[0138] This configuration set will be fed back to the control layer of the system and used for resource allocation and scheduling decisions in the actual network.

[0139] Through the adaptive learning and optimization feedback scheme of this step, we achieve the goal of dynamically adjusting the network configuration according to real-time feedback. This scheme ensures the efficient allocation of network resources and the collaborative cooperation between devices through innovative feedback adjustment factors, optimization strategy update functions, and regularization mechanisms. At the same time, through continuous optimization, this scheme improves the stability and overall performance of the network, responds to environmental changes, and solves the challenges of resource sharing in multi-device networks.

[0140] S5. Monitor the network status in real time, collect the key performance indicators of each device in the network, dynamically score the devices by combining historical data and target indicators, and dynamically adjust the device configuration based on real-time feedback and dynamic scoring to ensure that the network always operates efficiently under changing environmental conditions.

[0141] The input of this step comes from the output of step four, that is, the configuration scheme set C after optimized feedback adjustment ″o ″ pt , including the configuration scheme C for each device i i ′″ = (B i ″″ , C i ″″ , D i ″″ ), which includes the bandwidth allocation, channel selection, and scheduling strategy of device i. These optimized configuration schemes reflect the current state of the system after adaptive learning and optimization feedback, and provide basic data for real-time performance monitoring and continuous optimization.

[0142] Specifically, in this step, we need to monitor the running state of the system in real time to ensure that the configurations of each device can adapt to changes in the network environment. We use a real-time performance evaluation method based on a dynamic weighted scoring system. This method dynamically scores by collecting key performance indicators (such as throughput, latency, signal strength, etc.) of each device in the network in real time and combining historical data and target metrics.

[0143] For device i, we define the real-time performance scoring function P i (t), which combines the current feedback R i (t) of the device and the target feedback R target , and takes into account the contribution of the device configuration to the overall system performance. The performance scoring calculation formula is as follows:

[0144] P i (t) = ω i (t) · [R i (t) - R target + (1 - ω i (t)) · P i (t - 1) (15)

[0145] where, P i (t): the real-time performance score of device i at time t; ω i (t): the dynamic weighting coefficient of device i at time t, reflecting the importance of the device configuration; R i (t): the real-time feedback of device i; R target : the target feedback value (such as the desired throughput, latency, etc.); P i (t - 1): the performance score of device i at the previous time. This scoring function weights and fuses the real-time feedback and historical data, and improves the response speed to key devices and network state changes by dynamically adjusting the importance weight ω i (t).

[0146] Furthermore, the real-time performance score Pi (t) can serve as the basis for continuous optimization, which reflects the actual performance of each device in the network. Based on this, we adopt an adaptive adjustment mechanism to dynamically adjust the device configuration through feedback and scoring, achieving the effect of continuous optimization.

[0147] For each device i, we calculate the adjustment factor A i (t) according to its real-time performance score, and update the device configuration based on this adjustment factor. The adjustment factor formula is as follows:

[0148] A i (t) = β i ·)P i (t) - P target ) (16)

[0149] Where, A i (t): the configuration adjustment factor of device i at time t; β i : the adjustment step of device i, controlling the learning rate; P i (t): the real-time performance score of device i; P target : the target performance score of device i. The role of this formula is to adjust the device configuration according to the difference between the real-time performance score and the target performance, and control the adjustment amplitude according to the adjustment step β i of each device to avoid over-adjustment.

[0150] Furthermore, based on the adjustment factor A i (t), we will update the configuration scheme set C″ opt in each time period and feedback these adjustments to the system. To ensure the reasonable allocation of resources and avoid the situation where individual devices over-occupy resources, we introduce a resource reallocation mechanism. The resource reallocation is based on the optimization results of the devices, performs dynamic resource allocation, and optimizes the balance among bandwidth, channel, and scheduling strategies through the weighted average method.

[0151] Specifically, for the bandwidth allocation B″′ i of each device, we optimize it according to its adjustment factor and real-time performance score:

[0152] B″′ i = B″ i + γ i ·A i (t) (17)

[0153] Where, B″′ i : the adjusted bandwidth allocation of device i at time t; γ i : the bandwidth adjustment coefficient of device i, controlling the sensitivity of bandwidth update; A i(t): The configuration adjustment factor of device i. Through this mechanism, the bandwidth allocation of the device is dynamically updated according to the real-time feedback adjustment factor, so as to maintain the efficient use of network resources.

[0154] Furthermore, the output of this step is the new configuration scheme set C″′ after real-time monitoring and optimization adjustment opt , which includes the bandwidth, channel and scheduling strategy of each device, reflecting the optimal state achieved by the system during continuous optimization. This configuration set will be used for the next round of decision execution to ensure that the network always operates efficiently under changing environmental conditions.

[0155] Through the real-time performance monitoring and continuous optimization steps, we have achieved dynamic adjustment of device configurations and efficient reallocation of resources. This solution enables the network to automatically adapt to environmental changes and continuously optimize performance in each time period by introducing real-time performance scoring, dynamic weighting coefficients and adjustment factors. At the same time, the resource reallocation mechanism ensures the balanced utilization of network resources, avoids the problem of over-occupation of resources by some devices, and ensures the efficiency of multi-device collaborative work.

[0156] In one or more embodiments, as Figure 2 shown, an intelligent router automatic configuration system based on artificial intelligence is disclosed. The system includes:

[0157] A signal acquisition unit 010, which is used to directly collect the original data set from various sensors and device terminals and perform preprocessing to obtain an environmental feature data set; the original data set includes signal quality, network topology, environmental signal data and physical environment information; the environmental feature data set includes signal quality, interference between devices and bandwidth utilization rate;

[0158] A signal analysis unit 011, which is used to analyze the environmental feature data set and generate a targeted router configuration scheme based on the analysis results using the adaptive multi-armed bandit algorithm; wherein, the router configuration scheme includes bandwidth allocation, channel selection and device scheduling;

[0159] A configuration analysis unit 012, which is used to design a globally optimized objective function according to the targeted router configuration scheme, maximize the bandwidth utilization rate, channel quality, and effectiveness of device scheduling, and minimize interference. At the same time, a multi-dimensional resource competition matrix is constructed based on the interference degree, distance and bandwidth overlap degree between devices to quantify the interference and competition between devices. At the same time, based on the globally optimized objective function, an incremental optimization gradient descent algorithm is used for global optimization to obtain an optimized router configuration scheme;

[0160] A configuration adjustment unit 013, which is used to perform adaptive feedback learning on the optimized router configuration scheme based on the real-time feedback of the network and dynamically adjust the configuration strategy;

[0161] The dynamic configuration unit 014 is used to monitor the network status in real time, collect the key performance indicators of each device in the network, dynamically score the devices by combining historical data and target indicators, and dynamically adjust the device configuration based on real-time feedback and dynamic scoring to ensure that the network always operates efficiently under changing environmental conditions.

[0162] In summary, the present invention not only solves the defects of static configuration, lack of environmental perception and multi-device collaboration existing in the prior art, but also continuously optimizes the network configuration through an adaptive learning mechanism, providing a more intelligent, dynamic and self-optimizing network configuration solution. Finally, the present invention can achieve more efficient and stable network performance in actual use, greatly improving the user's network experience and meeting the needs of complex network environments in modern families and enterprises.

[0163] It should be noted that the specific working process of the intelligent router automatic configuration system based on artificial intelligence provided in the embodiments of the present invention is the same as the process of the intelligent router automatic configuration method based on artificial intelligence described in the above embodiments, and will not be elaborated here.

[0164] Compared with the prior art, the intelligent router automatic configuration system based on artificial intelligence provided in the embodiments of the present invention directly collects the original data set from various sensors and device terminals, and performs preprocessing to obtain the environmental feature data set; the original data set includes signal quality, network topology, environmental signal data and physical environment information; the environmental feature data set includes signal quality, interference between devices and bandwidth utilization rate; analyze the environmental feature data set, and use the adaptive multi-armed bandit algorithm to generate a targeted router configuration plan based on the analysis results; wherein, the router configuration plan includes bandwidth allocation, channel selection and device scheduling; design a globally optimized objective function according to the targeted router configuration plan to maximize the bandwidth utilization rate, channel quality and effectiveness of device scheduling, and minimize interference, and at the same time construct a multi-dimensional resource competition matrix based on the interference degree, distance and bandwidth overlap degree between devices to quantify the interference and competition between devices, and at the same time perform global optimization based on the globally optimized objective function using the incremental optimization gradient descent algorithm to obtain the optimized router configuration plan; perform adaptive feedback learning on the optimized router configuration plan based on the real-time feedback of the network, and dynamically adjust the configuration strategy; monitor the network status in real time, collect the key performance indicators of each device in the network, dynamically score the devices by combining historical data and target indicators, and dynamically adjust the device configuration based on real-time feedback and dynamic scoring to ensure that the network always operates efficiently under changing environmental conditions.

[0165] An embodiment of the present invention further provides an intelligent router automatic configuration device based on artificial intelligence, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the steps in the embodiment of the above-mentioned intelligent router automatic configuration method based on artificial intelligence are implemented, for example Figure 1 the steps S1 to S5 described therein; or, when the processor executes the computer program, the functions of each module in the above-mentioned system embodiments are implemented.

[0166] Exemplarily, the computer program may be divided into one or more modules. The one or more modules are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program in the intelligent router automatic configuration device based on artificial intelligence.

[0167] The intelligent router automatic configuration device based on artificial intelligence may be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The intelligent router automatic configuration device based on artificial intelligence may include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the intelligent router automatic configuration device based on artificial intelligence may further include input / output devices, network access devices, a bus, etc.

[0168] The processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The processor is the control center of the intelligent router automatic configuration device based on artificial intelligence, and connects various parts of the entire intelligent router automatic configuration device through various interfaces and lines.

[0169] The memory can be used to store the computer program and / or modules. By running or executing the computer program and / or modules stored in the memory, and by invoking the data stored in the memory, the processor can implement various functions of the intelligent router automatic configuration device based on artificial intelligence. The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function, etc.; the data storage area can store data created according to the operation of the air conditioner controller, etc. In addition, the memory can include high-speed random access memory, and can also include non-volatile memory, such as a hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage devices.

[0170] Among them, if the modules integrated in the intelligent router automatic configuration device based on artificial intelligence are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, to implement all or part of the processes in the above-mentioned embodiment methods of the present invention, it can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0171] Those of ordinary skill in the art can understand that to implement all or part of the processes in the above-mentioned embodiment methods, it can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the above-mentioned method embodiments. Among them, the storage medium can be a magnetic disk, optical disc, read-only memory (ROM), or random access memory (RAM), etc.

[0172] The above are the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and modifications can be made, and these improvements and modifications are also regarded as the protection scope of the present invention.

Claims

1. An automatic configuration method for intelligent routers based on artificial intelligence, characterized in that: The method comprises: S1. The original data set is directly collected from various sensors and device ends, and preprocessed to obtain an environmental feature data set; the original data set includes signal quality, network topology, environmental signal data and physical environment information; the environmental feature data set includes signal quality, interference between devices and bandwidth utilization; S2. Analyze the environmental feature data set, and generate a targeted router configuration scheme using an adaptive multi-armed bandit algorithm based on the analysis result; wherein the router configuration scheme includes bandwidth allocation, channel selection, and device scheduling; S3. Design a global optimization objective function based on the targeted router configuration scheme to maximize bandwidth utilization, channel quality, and device scheduling effectiveness, and minimize interference. At the same time, build a multi-dimensional resource competition matrix based on the interference, distance, and bandwidth overlap between devices to quantify the interference and competition between devices. At the same time, use an incremental optimization gradient descent algorithm based on the global optimization objective function to perform global optimization and obtain the optimized router configuration scheme. S4, based on the real-time feedback from the network, adaptive feedback learning is performed on the optimized router configuration scheme to dynamically adjust the configuration strategy; S5. Monitor the network status in real time, collect key performance indicators of each device in the network, dynamically score the devices based on historical data and target indicators, and dynamically adjust device configurations based on real-time feedback and dynamic scores to ensure that the network always maintains efficient operation under changing environmental conditions.

2. According to claim 1, the method for automatic configuration of intelligent routers based on artificial intelligence is characterized in that: The environmental signal data includes the signal strength, channel quality, interference and noise received by each device; The preprocessing is weighted moving average filtering, and the specific processing process is as follows: Smoothing the signal strength by weighted averaging, where the weights are set based on the device location and the signal strength attenuation function; Extract key features from environmental data: The signal quality represents the signal-to-noise ratio between device i and device j, and is used to measure the communication quality between devices; The inter-device interference is an assessment of the interference degree between devices in the network; The bandwidth utilization is the bandwidth occupancy rate of computing device i; The key features are extracted to form a data set and the principal component analysis is performed to reduce the dimensionality in order to reduce redundant features and retain the key influencing factors to obtain an environmental feature data set.

3. The method for automatic configuration of intelligent routers based on artificial intelligence according to claim 1, characterized in that: The specific process of analyzing the environmental feature data set is as follows: By evaluating the environmental feature data set based on an adaptive Gaussian mixture model, different environmental states of the network are identified; wherein the adaptive Gaussian mixture model uses maximum likelihood estimation to model the environmental feature data set and obtains the mean μ of each category i and variance σ i , thus obtaining the probability distribution of different environmental states.

4. The method for automatic configuration of intelligent routers based on artificial intelligence according to claim 3 is characterized in that: The adaptive multi-armed bandit algorithm is used to generate a targeted router configuration solution based on the analysis results. The specific processing process is as follows: For each environmental state, a configuration scheme is configured, and the configuration scheme includes bandwidth allocation, channel selection, and device scheduling; the adaptive multi-armed slot machine algorithm is dynamically adjusted according to the real-time feedback of each configuration scheme to maximize network performance. The adaptive multi-armed slot machine algorithm update rule is: Among them, Q i (t) represents the state S i The return estimate of the current configuration option, R i (t) is real-time feedback, n i (t) is the number of times this configuration is selected.

5. The method for automatic configuration of intelligent routers based on artificial intelligence according to claim 1 is characterized in that: The objective function of the global optimization It is expressed as: Among them, Util(B i ) represents the bandwidth utilization of device i; Qual(C i ) represents the channel quality of device i; M ij represents the resource contention degree between device i and device j; α, β, and γ represent the weight coefficients in the objective function, which control the influence of bandwidth, channel quality, and interference respectively; The resource contention degree is based on the interference degree I between device i and device j. i,j , the distance D between device i and device j i,j and the bandwidth overlap L between device i and device j i,j Sure.

6. The method for automatic configuration of intelligent routers based on artificial intelligence according to claim 5 is characterized in that: The optimized router configuration scheme also needs to undergo a first regularization process, and the first regularization process is used to balance configuration differences and ensure balanced allocation of network resources; the first regularization term of the first regularization is constrained based on an average value of all device configurations.

7. The method for automatic configuration of intelligent routers based on artificial intelligence according to claim 6 is characterized in that: The real-time feedback includes network performance indicators and environmental changes; Then the network-based real-time feedback performs adaptive feedback learning on the optimized router configuration scheme and dynamically adjusts the configuration strategy. The specific processing process is as follows: Based on the current feedback R i (t) and expected feedback R target The difference between determines the feedback adjustment factor for each device configuration; Combined with the feedback adjustment factor of each device configuration, the configuration strategy of each device is updated through an adaptive learning algorithm. In order to prevent excessive updates of configuration adjustments, a second regularization term is introduced in the update process to balance the resource allocation differences between devices, making resource allocation more balanced. Generate an adjusted set of optimized configuration solutions.

8. The method for automatic configuration of intelligent routers based on artificial intelligence according to claim 7 is characterized in that: The dynamic scoring is based on a real-time performance scoring function, expressed as: P i (t)=ω i (t)·[R i (t)-R target ]+(1-ω i (t))·P i (t-1) Among them, P i (t) represents the real-time performance score of device i at time t; ω i (t) represents the dynamic weight coefficient of device i at time t, reflecting the importance of device configuration; R i (t) represents the real-time feedback of device i; R target Indicates the target feedback value; P i (t-1) represents the performance score of device i at the previous moment.

9. The method for automatic configuration of intelligent routers based on artificial intelligence according to claim 8, characterized in that: The feedback adjustment factor of each device configuration is adjusted based on the real-time performance scoring function; and the configuration solution set is updated in each time period based on the adjusted feedback adjustment factor.

10. The intelligent router automatic configuration system based on artificial intelligence is characterized by: The system comprises: The signal acquisition unit is used to directly collect the original data set from various sensors and device ends, and pre-process it to obtain an environmental feature data set; the original data set includes signal quality, network topology, environmental signal data and physical environment information; the environmental feature data set includes signal quality, interference between devices and bandwidth utilization; A signal analysis unit, configured to analyze the environmental feature data set and generate a targeted router configuration scheme based on the analysis result using an adaptive multi-armed bandit algorithm; wherein the router configuration scheme includes bandwidth allocation, channel selection, and device scheduling; The configuration analysis unit is used to design a global optimization objective function according to a targeted router configuration scheme, maximize bandwidth utilization, channel quality, and the effectiveness of device scheduling, and minimize interference. At the same time, a multi-dimensional resource competition matrix is ​​constructed based on the interference degree, distance, and bandwidth overlap between devices to quantify the interference and competition between devices. At the same time, an incremental optimization gradient descent algorithm is used based on the global optimization objective function to perform global optimization and obtain an optimized router configuration scheme. A configuration adjustment unit, used to perform adaptive feedback learning on the optimized router configuration scheme based on real-time feedback from the network and dynamically adjust the configuration strategy; The dynamic configuration unit is used to monitor the network status in real time, collect the key performance indicators of each device in the network, dynamically score the devices based on historical data and target indicators, and dynamically adjust the device configuration based on real-time feedback and dynamic scores to ensure that the network always maintains efficient operation under changing environmental conditions.

Images