Wireless roaming optimization method and system for mobile device

Abstract

The present invention discloses a wireless roaming optimization method and system for a mobile device, and relates to the field of wireless roaming technology. A wireless roaming optimization system for a mobile device includes: a wireless roaming access selection module and a wireless roaming access optimization module. The present invention can achieve more accurate access point recommendations by comprehensively analyzing the wireless roaming environment and access point information of the target mobile device, combined with a wireless roaming access selection model, and can combine the multi-dimensional characteristics of the environment and the device to conduct a comprehensive analysis to improve the stability and quality of the roaming connection; by real-time monitoring of the dynamic environmental fluctuations and device fluctuations in the wireless roaming environment where the target mobile device is located, it can respond quickly when the device or environment changes; by introducing a wireless roaming optimization switching model, it can make judgments based on environmental fluctuations, device status, and preset access thresholds, and can make switching decisions at the appropriate time.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless roaming, and in particular to a wireless roaming optimization method and system for a mobile device. Background Art

[0002] Wireless roaming technology has made some progress in multi-device, high-density environments, but it still faces many challenges. There are currently some standard protocols used to accelerate the switching between devices and access points and reduce roaming delays, such as 802.11r and 802.11k. However, frequent switching, roaming failures in low-signal areas, and network congestion in large-scale device environments are still unresolved problems.

[0003] In order to improve the wireless roaming experience, it is necessary to optimize the switching strategy between devices and access points in real time based on the actual environment. Through intelligent roaming sensitivity adjustment, the device can automatically predict and pre-switching based on signal strength, load conditions and network fluctuations to avoid unnecessary delays. At the same time, deploying access points in advance and adjusting AP configuration according to dynamic traffic and location will help improve stability and speed during roaming. Summary of the invention

[0004] The present invention aims to provide a method and system for optimizing wireless roaming of a mobile device, so as to enhance the roaming experience of the mobile device in different environments and improve the stability and speed during the roaming process.

[0005] A method for optimizing wireless roaming of a mobile device comprises the following steps:

[0006] Based on the wireless roaming environment where the target mobile device is located, there are N wireless access points D n ; Set up wireless access point D n The corresponding access point information is X n ; Allocating a preset roaming access threshold based on the target mobile device;

[0007] Based on wireless access point D n Access point information X n A comprehensive analysis is performed with the wireless roaming access selection model to obtain a recommended wireless access point; wireless roaming selection is performed on the target mobile device based on the recommended wireless access point; the wireless roaming access selection model is constructed by model fusion based on multi-feature learning, and the multi-dimensional features of the wireless access point are extracted for access selection;

[0008] Monitor the dynamic fluctuation of the environment and the fluctuation of the equipment in the wireless roaming environment where the target mobile device is located in real time; make roaming optimization judgments according to the dynamic fluctuation of the environment, the fluctuation of the equipment and the wireless roaming optimization switching model, and obtain the switching judgment result; if the switching judgment result is yes, output the pre-switching wireless access point, and give priority to wireless roaming according to the pre-switching wireless access point when performing the next wireless roaming; otherwise, no operation;

[0009] If all wireless access points D n When none of them meet the preset roaming access threshold, the wireless roaming access optimization model is used for optimization analysis to obtain the optimized roaming access threshold; the access point is reselected according to the optimized roaming access threshold and the wireless roaming access selection model to obtain the optimized recommended wireless access point; the wireless roaming access selection model performs optimization selection based on a fast-converging group algorithm for fast roaming switching under special circumstances.

[0010] As a preferred technical solution of the present invention, the wireless roaming access selection model includes an access point feature extraction layer, a feature composite analysis layer, a wireless roaming judgment layer and a result output layer;

[0011] The access point feature extraction layer is used to extract access point information X n Perform feature extraction to obtain access point information feature X n ';

[0012] The feature composite analysis layer is used to analyze access point information feature X n 'Perform comprehensive analysis to obtain the access point evaluation index Z n ;

[0013] The wireless roaming judgment layer is used to evaluate the index Z of all access points n The access point evaluation index Z of the access point that meets the preset roaming access threshold is selected. n Corresponding wireless access point D n As a recommended wireless access point;

[0014] The result output layer is used to output the recommended wireless access points.

[0015] As a preferred technical solution of the present invention, the characteristic composite analysis layer includes a characteristic screening layer, a fluctuation analysis layer and a comprehensive analysis layer;

[0016] The feature screening layer is used to filter the access point information feature X based on the PCA model. n 'Perform principal component analysis to obtain relevant access point information features G n ;

[0017] The fluctuation analysis layer is used to analyze the access point information characteristics X based on the GRACH model. n'Perform fluctuation analysis to obtain access point fluctuation characteristics B n ;

[0018] The comprehensive analysis layer is used to analyze the relevant access point information characteristics G n and access point fluctuation characteristic B n Perform predictive analysis to obtain the access point evaluation index Z n ;

[0019] The specific steps for training the comprehensive analysis layer include:

[0020] Collecting several groups of access point comprehensive analysis training samples; each group of access point comprehensive analysis training samples includes access point main features, fluctuation features and corresponding evaluation indicators; combining several groups of access point comprehensive analysis training samples to obtain access point comprehensive analysis training sets;

[0021] The access point comprehensive analysis training set is input into the stacked LSTM model for model training to obtain the initial comprehensive analysis layer. The initial comprehensive analysis layer is evaluated. If the initial comprehensive analysis layer passes the model evaluation, the initial comprehensive analysis layer is used as the comprehensive analysis layer in the feature composite analysis layer. Otherwise, the access point comprehensive analysis training set is used to continue model training.

[0022] As a preferred technical solution of the present invention, the wireless roaming optimization switching model includes an environment stability analysis layer, a device feature analysis layer, a switching judgment layer and an access point re-output layer;

[0023] The environmental stability analysis layer is used to perform environmental stability analysis based on the dynamic fluctuation of the environment and obtain the environmental stability feature vector;

[0024] The equipment feature analysis layer is used to perform equipment stability analysis based on the dynamic fluctuation of the equipment and obtain the equipment stability feature vector;

[0025] The switching judgment layer is used to perform switching judgment analysis based on the environment stability feature vector and the device stability feature vector to obtain the switching judgment result;

[0026] The access point re-output layer is used to output the switching judgment result; if the switching judgment result is yes, the latest access point information corresponding to the current wireless access point is re-acquired; according to the latest access point information corresponding to the current wireless access point and the wireless roaming access selection model, the pre-switching wireless access point is obtained; otherwise, no operation is performed.

[0027] As a preferred technical solution of the present invention, the specific steps of performing the handover judgment analysis in the handover judgment layer include:

[0028] The switching judgment layer is constructed by stacking the MLP model and the CNN model; wherein, the convolutional network A is used to extract local vector features based on the CNN model, and the convolutional network E is used to extract global vector features based on the MLP model; the convolutional network A and the convolutional network E are stacked to obtain the switching judgment layer;

[0029] Specific steps for training the switching judgment layer:

[0030] Collecting several groups of switching judgment training samples; each group of switching judgment training samples contains a feature vector set with a target label value; combining several groups of switching judgment training samples to obtain a switching judgment training set;

[0031] Model training is performed based on the switching judgment training set to obtain an initial switching judgment layer; a model evaluation is performed on the initial switching judgment layer. If the initial switching judgment layer passes the model evaluation, the initial switching judgment layer is used as the switching judgment layer in the wireless roaming optimization switching model; otherwise, the model training is continued using the switching judgment training set.

[0032] As a preferred technical solution of the present invention, the wireless roaming access optimization model includes a data acquisition layer, an optimization analysis layer and a threshold output layer;

[0033] The data acquisition layer is used to obtain the environmental characteristic information of the wireless roaming environment where the target mobile device is located;

[0034] The optimization analysis layer is used to optimize the environment based on the characteristics of the wireless access point D n Perform optimization threshold analysis to obtain optimized roaming access thresholds;

[0035] The threshold output layer is used to output the optimized roaming access threshold.

[0036] As a preferred technical solution of the present invention, the specific steps of performing the optimization threshold analysis in the optimization analysis layer include:

[0037] Based on wireless access point D n The characteristics of all wireless access points are enhanced by the environmental characteristics information to obtain the enhanced wireless access point characteristics T n ;

[0038] A simulated digital twin model is constructed based on environmental feature information to obtain a virtual wireless roaming environment;

[0039] Construct K wireless roaming sensitive threshold individuals H k , k=1, 2, ...K; each wireless roaming sensitivity threshold individual H k Contains a simulated roaming access threshold for optimizing wireless roaming; K wireless roaming sensitive threshold individuals H kCombining to obtain the wireless roaming sensitivity threshold iteration population; setting the maximum number of iterations;

[0040] In a virtual wireless roaming environment, using enhanced wireless access point features T n and wireless roaming sensitivity threshold individual H k Perform simulated access point screening to obtain a simulated optimized wireless access point set; perform wireless quality analysis based on the simulated optimized wireless access point set to obtain a wireless roaming quality mean; use the wireless roaming quality mean as the wireless roaming sensitivity threshold individual H k The fitness Y k ;

[0041] The wireless roaming sensitive threshold iteration population is iterated and updated. When the number of iterations reaches 1 / 3 of the maximum number of iterations, the optimal iteration is performed, and the greedy strategy is used to select the optimal solution for individual mutation. At the same time, when the number of iterations reaches 2 / 3 of the maximum number of iterations, the elite individual retention strategy is performed, and some wireless roaming sensitive threshold individuals with larger fitness are retained without individual mutation operations. The remaining wireless roaming sensitive threshold individuals use the greedy strategy to select the optimal solution for individual mutation.

[0042] When the maximum number of iterations is reached, the wireless roaming sensitivity threshold individual corresponding to the maximum fitness is output, which is the optimal wireless roaming sensitivity threshold individual; the optimal wireless roaming sensitivity threshold individual is reset to output the optimized roaming access threshold.

[0043] A wireless roaming optimization system for a mobile device, comprising:

[0044] The wireless roaming access selection module includes a roaming selection unit; the roaming selection unit is used to select a wireless access point D based on the wireless roaming environment where the target mobile device is located. n ; Set up wireless access point D n The corresponding access point information is X n ; Based on the target mobile device to allocate a preset roaming access threshold; based on the wireless access point D n Access point information X n A comprehensive analysis is performed with the wireless roaming access selection model to obtain a recommended wireless access point; wireless roaming selection is performed on the target mobile device based on the recommended wireless access point; the wireless roaming access selection model is constructed by model fusion based on multi-feature learning, and the multi-dimensional features of the wireless access point are extracted for access selection;

[0045] The wireless roaming access optimization module includes an optimized roaming unit and an optimized access unit; the optimized roaming unit is used to monitor the dynamic fluctuation of the environment and the fluctuation of the equipment in the wireless roaming environment where the target mobile device is located in real time; the roaming optimization judgment is made according to the dynamic fluctuation of the environment, the fluctuation of the equipment and the wireless roaming optimization switching model, and the switching judgment result is obtained; if the switching judgment result is yes, the pre-switching wireless access point is output, and the wireless roaming is performed according to the pre-switching wireless access point first when performing the next wireless roaming; otherwise, no operation is performed; the optimized access unit is used to make the wireless roaming optimization judgment result. n When none of them meet the preset roaming access threshold, the wireless roaming access optimization model is used for optimization analysis to obtain the optimized roaming access threshold; the access point is reselected according to the optimized roaming access threshold and the wireless roaming access selection model to obtain the optimized recommended wireless access point; the wireless roaming access selection model performs optimization selection based on a fast-converging group algorithm for fast roaming switching under special circumstances.

[0046] The present invention has the following advantages:

[0047] 1. The present invention can achieve more accurate access point recommendation by comprehensively analyzing the wireless roaming environment and access point information of the target mobile device, combined with the wireless roaming access selection model, and can combine the multi-dimensional characteristics of the environment and the device to conduct a comprehensive analysis, so as to select the optimal access point in a dynamic environment, reduce the error of roaming selection, and improve the stability and quality of roaming connection; by real-time monitoring the dynamic environmental fluctuations and device fluctuations in the wireless roaming environment where the target mobile device is located, it can make a quick response when the device or environment changes; by introducing the wireless roaming optimization switching model, it can make judgments based on environmental fluctuations, device status and preset access thresholds, and can make switching decisions at appropriate times.

[0048] 2. The present invention can allocate network resources more efficiently by optimizing the selection of access points and adjusting the access threshold, avoiding excessive load or inefficient utilization of access points. The optimized roaming access threshold can effectively balance the loads of different access points, improve the resource utilization of the entire wireless network, and reduce resource waste. By enhancing access point characteristics and simulating digital twin environments, it can adapt to wireless roaming optimization in more access points and complex environments. As the network scale increases or the environmental complexity increases, the model can maintain efficient roaming access performance through continuous optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 The present invention is a schematic diagram of the structure of a wireless roaming optimization system for a mobile device used in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] In order to enable persons skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0051] Embodiment 1, a method for optimizing wireless roaming of a mobile device, comprising the following steps:

[0052] Based on the wireless roaming environment where the target mobile device is located, there are N wireless access points D n ; Set up wireless access point D n The corresponding access point information is X n ; Allocating a preset roaming access threshold based on the target mobile device;

[0053] Access point information includes signal strength, network performance, bandwidth, latency, throughput, load, supported technologies, historical device connection quality, interference, device characteristics, etc., which can be selectively obtained based on actual conditions. The preset roaming access threshold is set by professional technicians based on historical data and actual conditions, which meets the access threshold of most devices.

[0054] Based on wireless access point D n Access point information X n A comprehensive analysis is performed with the wireless roaming access selection model to obtain a recommended wireless access point; wireless roaming selection is performed on the target mobile device based on the recommended wireless access point; the wireless roaming access selection model is constructed by model fusion based on multi-feature learning, and the multi-dimensional features of the wireless access point are extracted for access selection;

[0055] The wireless roaming access selection model includes an access point feature extraction layer, a feature composite analysis layer, a wireless roaming judgment layer and a result output layer;

[0056] The access point feature extraction layer is used to extract access point information X n Perform feature extraction to obtain access point information feature X n ';

[0057] The feature composite analysis layer is used to analyze access point information feature X n 'Perform comprehensive analysis to obtain the access point evaluation index Z n ;

[0058] The wireless roaming judgment layer is used to evaluate the index Z of all access points n The access point evaluation index Z of the access point that meets the preset roaming access threshold is selected. n Corresponding wireless access point D n As a recommended wireless access point;

[0059] The result output layer is used to output the recommended wireless access points;

[0060] The wireless roaming access selection model can accurately select the most suitable access point from multiple access points through multiple levels of analysis and processing. First, the access point feature extraction layer converts access point information such as signal strength, bandwidth, and delay into meaningful features, making subsequent analysis more accurate. Then, a multi-dimensional comprehensive evaluation is performed through the feature composite analysis layer, and finally the evaluation index of each access point is obtained, ensuring that the access point selection does not only consider a single factor, but a multi-angle evaluation. The wireless roaming judgment layer can screen out the access point that meets the threshold conditions and has the highest evaluation by comparing the access point evaluation index with the preset roaming access threshold. The judgment mechanism of this layer ensures that the access point selection meets both the device requirements and the network conditions, ensuring that the device can always maintain the best connection quality during the roaming process.

[0061] The feature composite analysis layer includes feature screening layer, fluctuation analysis layer and comprehensive analysis layer;

[0062] The feature screening layer is used to filter the access point information feature X based on the PCA model. n 'Perform principal component analysis to obtain relevant access point information features G n ; In the feature screening layer, the dimension reduction processing of the access point information features based on PCA can effectively reduce the dimension of the original features, thereby reducing the complexity of the data and retaining the most important features. This dimensionality reduction processing helps to remove redundant and noisy data, so that subsequent analysis can focus more on key features, thereby improving the efficiency of feature selection and the accuracy of access point evaluation;

[0063] The fluctuation analysis layer is used to analyze the access point information characteristics X based on the GRACH model. n 'Perform fluctuation analysis to obtain access point fluctuation characteristics B n The fluctuation analysis layer performs fluctuation analysis based on the GRACH model, which can effectively identify and quantify the fluctuation characteristics of access points in time and space. The GRACH model has strong time series analysis capabilities and can capture the changing patterns of access points in different time periods or network environments. By incorporating fluctuation characteristics into the evaluation, the model can better predict the stability and reliability of access points.

[0064] The comprehensive analysis layer is used to analyze the relevant access point information characteristics G n and access point fluctuation characteristic B n Perform predictive analysis to obtain the access point evaluation index Z nThe comprehensive analysis layer jointly analyzes the principal component characteristics and fluctuation characteristics of the access point to obtain the access point evaluation index. Through this composite analysis, the model can consider the stability and performance of the access point and generate a more comprehensive and accurate access point evaluation. This process combines the static characteristics and dynamic fluctuation characteristics of the access point, thereby better predicting the comprehensive performance of the access point.

[0065] The specific steps for training the comprehensive analysis layer include:

[0066] Collecting several groups of access point comprehensive analysis training samples; each group of access point comprehensive analysis training samples includes access point main features, fluctuation features and corresponding evaluation indicators; combining several groups of access point comprehensive analysis training samples to obtain access point comprehensive analysis training sets;

[0067] Input the access point comprehensive analysis training set into the stacked LSTM model for model training to obtain the initial comprehensive analysis layer; perform model evaluation on the initial comprehensive analysis layer. If the initial comprehensive analysis layer passes the model evaluation, the initial comprehensive analysis layer is used as the comprehensive analysis layer in the feature composite analysis layer; otherwise, continue model training using the access point comprehensive analysis training set;

[0068] The access point comprehensive analysis training set is trained by stacking LSTM models. The comprehensive analysis layer can automatically learn the rules of access point selection based on historical data and make accurate predictions for future access point evaluations. Stacking LSTM can model complex time series relationships and conduct in-depth learning of the volatility and historical characteristics of access points, thereby improving the model's adaptability and predictive ability to environmental changes. The training process of the stacking LSTM model can gradually optimize the access point selection strategy through multiple stages of learning. As the training progresses, the model can gradually converge and eventually form an efficient decision-making mechanism to quickly evaluate the comprehensive index of each access point, thereby making rapid and accurate access point selection.

[0069] Monitor the dynamic fluctuation of the environment and the fluctuation of the equipment in the wireless roaming environment where the target mobile device is located in real time; make roaming optimization judgments according to the dynamic fluctuation of the environment, the fluctuation of the equipment and the wireless roaming optimization switching model, and obtain the switching judgment result; if the switching judgment result is yes, output the pre-switching wireless access point, and give priority to wireless roaming according to the pre-switching wireless access point when performing the next wireless roaming; otherwise, no operation;

[0070] The wireless roaming optimization switching model includes an environment stability analysis layer, a device feature analysis layer, a switching judgment layer, and an access point re-output layer;

[0071] The environmental stability analysis layer is used to perform environmental stability analysis based on the dynamic fluctuation of the environment and obtain the environmental stability feature vector;

[0072] The equipment feature analysis layer is used to perform equipment stability analysis based on the dynamic fluctuation of the equipment and obtain the equipment stability feature vector;

[0073] The switching judgment layer is used to perform switching judgment analysis based on the environment stability feature vector and the device stability feature vector to obtain the switching judgment result;

[0074] The access point re-output layer is used to output the switching judgment result; if the switching judgment result is yes, then re-acquire the latest access point information corresponding to the current wireless access point; according to the latest access point information corresponding to the current wireless access point and the wireless roaming access selection model, obtain the pre-switching wireless access point; otherwise, no operation;

[0075] The model can monitor and analyze dynamic environmental fluctuations (such as signal strength changes, network interference, etc.) and device fluctuations (such as device location, speed, etc.) in real time, thereby providing more accurate roaming decisions. Through dynamic analysis, roaming interruptions or signal instability caused by sudden network fluctuations or device status changes can be effectively avoided. Through environmental stability analysis and device feature analysis, the model can identify factors that affect network quality and make adjustments in advance, thereby optimizing switching judgments during roaming and reducing switching failures and delays.

[0076] When it is determined that switching is necessary, the model can identify and recommend a more stable access point in advance. This pre-switching mechanism reduces the delay when switching access points, avoids network interruptions caused by re-searching for access points, and ensures smoother network connections for devices. By comprehensively considering environmental stability and device characteristics, the switching judgment layer can make more accurate switching decisions, reducing switching failures or unnecessary frequent switching caused by misjudgment.

[0077] The specific steps of performing the switching judgment analysis in the switching judgment layer include:

[0078] The switching judgment layer is constructed by stacking the MLP model and the CNN model; wherein, the convolutional network A is used to extract local vector features based on the CNN model, and the convolutional network E is used to extract global vector features based on the MLP model; the convolutional network A and the convolutional network E are stacked to obtain the switching judgment layer;

[0079] Specific steps for training the switching judgment layer:

[0080] Collecting several groups of switching judgment training samples; each group of switching judgment training samples contains a feature vector set with a target label value; combining several groups of switching judgment training samples to obtain a switching judgment training set;

[0081] Model training is performed based on the handover judgment training set to obtain an initial handover judgment layer; model evaluation is performed on the initial handover judgment layer, and if the initial handover judgment layer passes the model evaluation, the initial handover judgment layer is used as the handover judgment layer in the wireless roaming optimization handover model; otherwise, the model training is continued using the handover judgment training set;

[0082] By stacking convolutional neural networks and multi-layer perceptron models, the switching judgment layer can extract local and global features at the same time; convolutional network A extracts local vector features, which is suitable for capturing small fluctuations in signals or local environmental changes, while convolutional network E can analyze the overall environment and dynamic fluctuations of the device from a global level. This structure enables the model to more comprehensively and accurately determine whether the access point needs to be switched; the joint analysis of local and global features allows the switching judgment to consider not only short-term dynamic changes, but also long-term environmental stability, avoiding misjudgments caused by over-reliance on features at a certain level;

[0083] By extracting local features through the CNN model and global features through the MLP model, the switching judgment layer can comprehensively consider various factors, such as the dynamic state of the device, local fluctuations in the wireless network, and changes in the overall environment. This multi-level feature fusion can effectively reduce misjudgments caused by a single factor.

[0084] If all wireless access points D n When none of them meet the preset roaming access threshold, the wireless roaming access optimization model is used for optimization analysis to obtain the optimized roaming access threshold; the access point is reselected according to the optimized roaming access threshold and the wireless roaming access selection model to obtain the optimized recommended wireless access point; the wireless roaming access selection model is optimized based on the fast convergence group algorithm for fast roaming switching under special circumstances; due to the joint stacking structure of CNN and MLP, the switching judgment layer has strong feature learning and representation capabilities, and can make more robust judgments in complex wireless environments (such as interference, network fluctuations, etc.);

[0085] The wireless roaming access optimization model includes a data acquisition layer, an optimization analysis layer and a threshold output layer;

[0086] The data acquisition layer is used to obtain the environmental characteristic information of the wireless roaming environment where the target mobile device is located;

[0087] The optimization analysis layer is used to optimize the environment based on the characteristics of the wireless access point D n Perform optimization threshold analysis to obtain optimized roaming access thresholds;

[0088] The threshold output layer is used to output the optimized roaming access threshold;

[0089] The specific steps of performing optimization threshold analysis in the optimization analysis layer include:

[0090] Based on wireless access point D n The characteristics of all wireless access points are enhanced by the environmental characteristics information to obtain the enhanced wireless access point characteristics T n ;

[0091] A simulated digital twin model is constructed based on environmental feature information to obtain a virtual wireless roaming environment;

[0092] Construct K wireless roaming sensitive threshold individuals H k , k=1, 2, ...K; each wireless roaming sensitivity threshold individual H k Contains a simulated roaming access threshold for optimizing wireless roaming; K wireless roaming sensitive threshold individuals H k Combine to obtain the wireless roaming sensitivity threshold iteration population; set the maximum number of iterations, which is set by professional technicians according to actual conditions;

[0093] In a virtual wireless roaming environment, using enhanced wireless access point features T n and wireless roaming sensitivity threshold individual H k Perform simulated access point screening to obtain a simulated optimized wireless access point set; perform wireless quality analysis based on the simulated optimized wireless access point set to obtain a wireless roaming quality mean; use the wireless roaming quality mean as the wireless roaming sensitivity threshold individual H k The fitness Y k ;

[0094] The wireless roaming sensitive threshold iteration population is iterated and updated. When the number of iterations reaches 1 / 3 of the maximum number of iterations, the optimal iteration is performed, and the greedy strategy is used to select the optimal solution for individual mutation. At the same time, when the number of iterations reaches 2 / 3 of the maximum number of iterations, the elite individual retention strategy is performed, and some wireless roaming sensitive threshold individuals with larger fitness are retained without individual mutation operations. The remaining wireless roaming sensitive threshold individuals use the greedy strategy to select the optimal solution for individual mutation.

[0095] When the maximum number of iterations is reached, the wireless roaming sensitivity threshold individual corresponding to the maximum fitness is output, which is the optimal wireless roaming sensitivity threshold individual; the optimal wireless roaming sensitivity threshold individual is reset and the optimized roaming access threshold is output;

[0096] By enhancing the features of wireless access points and building a virtual wireless roaming environment based on environmental feature information, the optimization analysis layer can more comprehensively consider the impact of the environment on access point selection. Feature enhancement enables the attributes of access points to better reflect the various complex factors in the real environment. The simulated digital twin model provides virtual environment support for further optimization, so that the optimization process is not only based on real data, but also can simulate possible future environmental changes. The optimization analysis layer conducts in-depth analysis of access point selection by integrating environmental feature information and access point features, and then obtains the optimized access threshold. This multi-dimensional optimization analysis improves the decision-making accuracy of the model and reduces the errors that may occur in traditional single condition judgment.

[0097] Through the greedy strategy, elite individual retention and maximum number of iterations, the model can efficiently converge to the optimal solution, avoiding the problems of high computational complexity and slow convergence that may occur in traditional optimization methods. This not only saves time and resources, but also ensures the accuracy of the optimization results. The mean value of wireless roaming quality is a key fitness indicator in the optimization process, which can comprehensively reflect the impact of the selection of wireless access points on network performance. By performing wireless quality analysis on the simulated optimized wireless access point set in each iteration, the optimization analysis layer ensures that the wireless roaming quality can be maximized when selecting access points.

[0098] By optimizing the selection of access points and adjusting the access threshold, network resources can be allocated more efficiently and overloading or inefficient utilization of access points can be avoided. The optimized roaming access threshold can effectively balance the load of different access points, improve the resource utilization of the entire wireless network, and reduce resource waste. By enhancing access point characteristics and simulating digital twin environments, it can adapt to wireless roaming optimization in more access points and complex environments. As the network scale increases or the environmental complexity increases, the model can maintain efficient roaming access performance through continuous optimization.

[0099] Example 2, a wireless roaming optimization system for a mobile device, see Figure 1 As shown, including:

[0100] The wireless roaming access selection module includes a roaming selection unit; the roaming selection unit is used to select a wireless access point D based on the wireless roaming environment where the target mobile device is located. n ; Set up wireless access point D n The corresponding access point information is X n ; Based on the target mobile device to allocate a preset roaming access threshold; based on the wireless access point D n Access point information X nA comprehensive analysis is performed with the wireless roaming access selection model to obtain a recommended wireless access point; wireless roaming selection is performed on the target mobile device based on the recommended wireless access point; the wireless roaming access selection model is constructed by model fusion based on multi-feature learning, and the multi-dimensional features of the wireless access point are extracted for access selection;

[0101] The wireless roaming access optimization module includes an optimized roaming unit and an optimized access unit; the optimized roaming unit is used to monitor the dynamic fluctuation of the environment and the fluctuation of the equipment in the wireless roaming environment where the target mobile device is located in real time; the roaming optimization judgment is made according to the dynamic fluctuation of the environment, the fluctuation of the equipment and the wireless roaming optimization switching model, and the switching judgment result is obtained; if the switching judgment result is yes, the pre-switching wireless access point is output, and the wireless roaming is performed according to the pre-switching wireless access point first when performing the next wireless roaming; otherwise, no operation is performed; the optimized access unit is used to make the wireless roaming optimization judgment result. n When none of them meet the preset roaming access threshold, the wireless roaming access optimization model is used for optimization analysis to obtain the optimized roaming access threshold; the access point is reselected according to the optimized roaming access threshold and the wireless roaming access selection model to obtain the optimized recommended wireless access point; the wireless roaming access selection model performs optimization selection based on a fast-converging group algorithm for fast roaming switching under special circumstances.

[0102] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A wireless roaming optimization method for a mobile device, characterized in that: The following steps are involved: Based on the wireless roaming environment where the target mobile device is located, there are N wireless access points D n ; Set up wireless access point D n The corresponding access point information is X n ; Allocating a preset roaming access threshold based on the target mobile device; Based on wireless access point D n Access point information X n A comprehensive analysis is performed with the wireless roaming access selection model to obtain a recommended wireless access point; wireless roaming selection is performed on the target mobile device based on the recommended wireless access point; the wireless roaming access selection model is constructed by model fusion based on multi-feature learning, and the multi-dimensional features of the wireless access point are extracted for access selection; Monitor the dynamic fluctuation of the environment and the fluctuation of the equipment in the wireless roaming environment where the target mobile device is located in real time; make roaming optimization judgments according to the dynamic fluctuation of the environment, the fluctuation of the equipment and the wireless roaming optimization switching model, and obtain the switching judgment result; if the switching judgment result is yes, output the pre-switching wireless access point, and give priority to wireless roaming according to the pre-switching wireless access point when performing the next wireless roaming; otherwise, no operation; If all wireless access points D n When none of them meet the preset roaming access threshold, the wireless roaming access optimization model is used for optimization analysis to obtain the optimized roaming access threshold; the access point is reselected according to the optimized roaming access threshold and the wireless roaming access selection model to obtain the optimized recommended wireless access point; the wireless roaming access selection model is optimized based on the fast convergence group algorithm for fast roaming switching in special circumstances; The wireless roaming access selection model includes an access point feature extraction layer, a feature composite analysis layer, a wireless roaming judgment layer and a result output layer; The access point feature extraction layer is used to extract access point information X n Perform feature extraction to obtain access point information feature X n '; The feature composite analysis layer is used to analyze access point information feature X n 'Perform comprehensive analysis to obtain the access point evaluation index Z n ; The wireless roaming judgment layer is used to evaluate the index Z of all access points n The access point evaluation index Z of the access point that meets the preset roaming access threshold is selected. n Corresponding wireless access point D n As a recommended wireless access point; The result output layer is used to output the recommended wireless access points.

2. The wireless roaming optimization method for a mobile device according to claim 1, characterized in that: The feature composite analysis layer includes feature screening layer, fluctuation analysis layer and comprehensive analysis layer; The feature screening layer is used to filter the access point information feature X based on the PCA model. n 'Perform principal component analysis to obtain relevant access point information features G n ; The fluctuation analysis layer is used to analyze the access point information characteristics X based on the GRACH model. n 'Perform fluctuation analysis to obtain access point fluctuation characteristics B n ; The comprehensive analysis layer is used to analyze the relevant access point information characteristics G n and access point fluctuation characteristic B n Perform predictive analysis to obtain the access point evaluation index Z n ; The specific steps for training the comprehensive analysis layer include: Collecting several groups of access point comprehensive analysis training samples; each group of access point comprehensive analysis training samples includes access point main features, fluctuation features and corresponding evaluation indicators; combining several groups of access point comprehensive analysis training samples to obtain access point comprehensive analysis training sets; The access point comprehensive analysis training set is input into the stacked LSTM model for model training to obtain the initial comprehensive analysis layer. The initial comprehensive analysis layer is evaluated. If the initial comprehensive analysis layer passes the model evaluation, the initial comprehensive analysis layer is used as the comprehensive analysis layer in the feature composite analysis layer. Otherwise, the access point comprehensive analysis training set is used to continue model training.

3. The wireless roaming optimization method for a mobile device according to claim 2, characterized in that: The wireless roaming optimization switching model includes an environment stability analysis layer, a device feature analysis layer, a switching judgment layer, and an access point re-output layer; The environmental stability analysis layer is used to perform environmental stability analysis based on the dynamic fluctuation of the environment and obtain the environmental stability feature vector; The equipment feature analysis layer is used to perform equipment stability analysis based on the dynamic fluctuation of the equipment and obtain the equipment stability feature vector; The switching judgment layer is used to perform switching judgment analysis based on the environment stability feature vector and the device stability feature vector to obtain the switching judgment result; The access point re-output layer is used to output the switching judgment result; if the switching judgment result is yes, the latest access point information corresponding to the current wireless access point is re-acquired; According to the latest access point information corresponding to the current wireless access point and the wireless roaming access selection model, the pre-switching wireless access point is obtained; otherwise, no operation is performed.

4. The wireless roaming optimization method for a mobile device according to claim 3, characterized in that: The specific steps of performing the switching judgment analysis in the switching judgment layer include: The switching judgment layer is constructed by stacking the MLP model and the CNN model; wherein, the convolutional network A is used to extract local vector features based on the CNN model, and the convolutional network E is used to extract global vector features based on the MLP model; the convolutional network A and the convolutional network E are stacked to obtain the switching judgment layer; Specific steps for training the switching judgment layer: Collecting several groups of switching judgment training samples; each group of switching judgment training samples contains a feature vector set with a target label value; combining several groups of switching judgment training samples to obtain a switching judgment training set; Model training is performed based on the switching judgment training set to obtain an initial switching judgment layer; a model evaluation is performed on the initial switching judgment layer. If the initial switching judgment layer passes the model evaluation, the initial switching judgment layer is used as the switching judgment layer in the wireless roaming optimization switching model; otherwise, the model training is continued using the switching judgment training set.

5. The wireless roaming optimization method for a mobile device according to claim 4, characterized in that: The wireless roaming access optimization model includes a data acquisition layer, an optimization analysis layer and a threshold output layer; The data acquisition layer is used to obtain the environmental characteristic information of the wireless roaming environment where the target mobile device is located; The optimization analysis layer is used to optimize the environment based on the characteristics of the wireless access point D n Perform optimization threshold analysis to obtain optimized roaming access thresholds; The threshold output layer is used to output the optimized roaming access threshold.

6. The wireless roaming optimization method for a mobile device according to claim 5, characterized in that: The specific steps of performing optimization threshold analysis in the optimization analysis layer include: Based on wireless access point D n The characteristics of all wireless access points are enhanced by the environmental characteristics information to obtain the enhanced wireless access point characteristics T n ; A simulated digital twin model is constructed based on environmental feature information to obtain a virtual wireless roaming environment; Construct K wireless roaming sensitive threshold individuals H k , k=1, 2, ...K; each wireless roaming sensitivity threshold individual H k Contains the simulated roaming access threshold for optimizing wireless roaming; K wireless roaming sensitive threshold individuals H k Combining to obtain the wireless roaming sensitivity threshold iteration population; setting the maximum number of iterations; In a virtual wireless roaming environment, using enhanced wireless access point features T n and wireless roaming sensitivity threshold individual H k Perform simulated access point screening to obtain a simulated optimized wireless access point set; perform wireless quality analysis based on the simulated optimized wireless access point set to obtain a wireless roaming quality mean; use the wireless roaming quality mean as the wireless roaming sensitivity threshold individual H k The fitness Y k ; The wireless roaming sensitive threshold iteration population is iterated and updated. When the number of iterations reaches 1 / 3 of the maximum number of iterations, the optimal iteration is performed, and the greedy strategy is used to select the optimal solution for individual mutation. At the same time, when the number of iterations reaches 2 / 3 of the maximum number of iterations, the elite individual retention strategy is performed, and some wireless roaming sensitive threshold individuals with larger fitness are retained without individual mutation operations. The remaining wireless roaming sensitive threshold individuals use the greedy strategy to select the optimal solution for individual mutation. When the maximum number of iterations is reached, the wireless roaming sensitivity threshold individual corresponding to the maximum fitness is output, which is the optimal wireless roaming sensitivity threshold individual; the optimal wireless roaming sensitivity threshold individual is reset to output the optimized roaming access threshold.

7. A wireless roaming optimization system for a mobile device, characterized in that: The system applies a wireless roaming optimization method for a mobile device as described in any one of claims 1 to 6, including: The wireless roaming access selection module includes a roaming selection unit; the roaming selection unit is used to select a wireless access point D based on the wireless roaming environment where the target mobile device is located. n ; Set up wireless access point D n The corresponding access point information is X n ; Based on the target mobile device to allocate a preset roaming access threshold; based on the wireless access point D n Access point information X n A comprehensive analysis is performed with the wireless roaming access selection model to obtain a recommended wireless access point; wireless roaming selection is performed on the target mobile device based on the recommended wireless access point; the wireless roaming access selection model is constructed by model fusion based on multi-feature learning, and the multi-dimensional features of the wireless access point are extracted for access selection; The wireless roaming access optimization module includes an optimized roaming unit and an optimized access unit; the optimized roaming unit is used to monitor the dynamic fluctuation of the environment and the fluctuation of the equipment in the wireless roaming environment where the target mobile device is located in real time; the roaming optimization judgment is made according to the dynamic fluctuation of the environment, the fluctuation of the equipment and the wireless roaming optimization switching model, and the switching judgment result is obtained; if the switching judgment result is yes, the pre-switching wireless access point is output, and the wireless roaming is performed according to the pre-switching wireless access point first when performing the next wireless roaming; otherwise, no operation is performed; the optimized access unit is used to make the wireless roaming optimization judgment result. n When none of them meet the preset roaming access threshold, the wireless roaming access optimization model is used for optimization analysis to obtain the optimized roaming access threshold; the access point is reselected according to the optimized roaming access threshold and the wireless roaming access selection model to obtain the optimized recommended wireless access point; the wireless roaming access selection model performs optimization selection based on a fast-converging group algorithm for fast roaming switching under special circumstances.

Images