Method and device for switching credible wireless local area network (WLAN) of substation

Abstract

The invention discloses a switching method and device for a credible WLAN of a transformer substation, and belongs to the technical field of wireless communication of a power system, and the method comprises the steps: collecting historical network switching data of the transformer substation, and generating a time sequence feature matrix as an input data set; performing normalization processing on the receiving power, the time delay, the transmission rate and the network load, and calculating the weights of the four attributes according to the similarity matrix; calculating a multi-attribute effect value, supplementing the corresponding value of the training data into the input data set, and predicting the values of the current service network and the network to be switched at the next moment based on the historical feature data by the improved random forest model; when a switching condition is met, selecting a target network for switching; and filling the predicted value of the target network into the output data set to enable the output data set to become a new input data set for iterative updating, and continuously adjusting the parameters of the prediction model to optimize the prediction effect. According to the invention, the switching decision-making efficiency and stability in a complex environment are improved.

Description

Technical Field

[0001] The invention relates to a switching method and a device for a trusted WLAN in a transformer substation, belonging to the technical field of wireless communication in power systems. Background Art

[0002] As the power grid system becomes increasingly complex, this approach is prone to lead to incorrect switching decisions, thereby reducing the efficiency and stability of the power system. Therefore, how to comprehensively consider multiple attributes during the substation switching process, especially to achieve optimal switching under variable operating conditions, has become an urgent problem to be solved. In the research on inter-zone switching algorithms, the early focus was on the single attribute switching decision method based on signal strength, that is, deciding when to switch by setting a fixed signal strength threshold value. Although this method is simple and direct, it is easy to cause frequent switching, especially in scenarios with drastic signal fluctuations, which will produce a ping-pong effect, affecting the user experience and increasing the burden on network resources.

[0003] Existing studies have made important progress in optimizing the efficiency and accuracy of inter-zone switching. Researchers have explored a method based on a dual-attribute decision model, which predicts the switching threshold value at future moments by comprehensively considering multiple key parameters, significantly improving the accuracy and stability of switching. At the same time, optimization schemes based on deep reinforcement learning are widely used. This type of algorithm reduces the ping-pong effect caused by frequent switching in traditional switching methods by simulating multiple environmental states and actions. In addition, in response to the WLAN switching problem in industrial environments, edge computing and multi-input multi-output technologies have been introduced internationally to make switching more seamless and accurate, effectively ensuring network quality in complex scenarios.

[0004] In general, existing research has achieved higher switching performance and stability in complex network environments by combining multi-attribute decision making, machine learning, optimization algorithms and emerging technologies, providing a solid foundation for further improving network switching efficiency in the future. Therefore, the present invention provides a switching method and device for a trusted WLAN in a substation. Summary of the invention

[0005] In order to solve the above problems, the present invention proposes a switching method and device for a trusted WLAN in a substation, which can improve the efficiency and stability of switching decisions in complex environments by combining multi-attribute utility theory with an improved random forest model.

[0006] The technical solution adopted by the present invention to solve the technical problem is:

[0007] In a first aspect, an embodiment of the present invention provides a switching method for a substation trusted WLAN, comprising the following steps:

[0008] Step S1, collect the historical network switching data of the substation, generate a time series feature matrix as the input data set, and the input data set includes received power, delay, transmission rate, and network load;

[0009] Step S2, perform normalization processing on the received power, delay, transmission rate, and network load, and calculate the weights of the four attributes according to the similarity matrix (Analytic Network Process, ANP);

[0010] Step S3, calculate the multi-attribute effect value, fill the corresponding values of the training data into the input data set, select the improved random forest (Enhanced Random Forest, ES-RF) model as the prediction tool, and predict the values of the current service network and the network to be switched at the next moment based on the historical feature data;

[0011] Step S4, when the difference in utility values is greater than the switching threshold and reaches the delay set by the timer, and the load of the target switching network does not exceed 80%, select the target network for switching;

[0012] Step S5, fill the predicted values of the target network into the output data set to make it a new input data set for iterative update, and optimize the prediction effect by continuously adjusting the parameters of the prediction model.

[0013] As a possible implementation manner of this embodiment, the step S1 includes:

[0014] Select switching decision parameters, and the switching decision parameters include received power collected at the physical layer, delay, transmission rate, and network load concerned at the network layer;

[0015] Use the measurement configuration to periodically capture switching data packets and report them;

[0016] Use the sliding window method to process the reported switching data packet data and generate a time series feature matrix;

[0017] Use the time series feature matrix as the training input data set to capture the historical network performance.

[0018] As a possible implementation manner of this embodiment, the step S2 includes:

[0019] Construct a similarity matrix, and the similarity matrix is used to represent the importance of an attribute relative to an attribute;

[0020] Based on the analysis of historical data, further calculate the weights of each network attribute based on the similarity matrix;

[0021] Perform a consistency test and calculate the consistency ratio of the judgment matrix;

[0022] Further optimize the weights using a hypermatrix to better reflect the feedback and interaction between network attributes;

[0023] Calculate the utility value of historical data using a multi-attribute utility linear weighted model of Quality of Experience (QoE). The calculation formula for the utility value is:

[0024]

[0025] Where P recv (n) is the received power, D(n) is the delay, L(n) is the network load, R(n) is the transmission rate, and Q t is the comprehensive utility value of the i-th handover network, satisfying w 1 +w 2 +w 3 +w 4 = 1 and where k is the interaction threshold to prevent the excessive influence of the power and rate interaction terms;

[0026] Use the calculated utility value to improve the training dataset of machine learning.

[0027] As a possible implementation of this embodiment, the calculation formula for the weight of each network attribute is:

[0028]

[0029] Where, represents the weight of the i-th attribute, and C ij represents the element in the i-th row and j-th column of the C matrix.

[0030] As a possible implementation of this embodiment, step S3 includes:

[0031] Introduce an improved random forest model and dynamically adjust its weights according to the contribution degree of each feature to the utility value. The weights are updated according to the performance of historical data and the current network state;

[0032] Introduce a dynamic update of the weight vector to change the importance of four key attributes and form an adaptive feature input;

[0033] Adopt a stratified sampling strategy to divide the samples into different categories and uniformly sample from each category to ensure the representativeness of the training data.

[0034] As a possible implementation of this embodiment, step S4 includes:

[0035] Determine a credible handover decision model and set a given handover threshold;

[0036] Determine whether the following conditions are simultaneously satisfied:

[0037] (a) The difference between the utility value of the target network and the utility value of the currently connected network is greater than a given handover threshold and reaches the delay set by the timer;

[0038] (b) The load of the target handover network does not exceed 80%;

[0039] When conditions (a) and (b) are simultaneously satisfied, determine to perform the handover.

[0040] As a possible implementation manner of this embodiment, the step S5 includes:

[0041] Fill the predicted value of the target network into the output data set to make the output data set become a new input data set;

[0042] Use the new input data set to perform iterative updates of steps S1 to S4;

[0043] During the iterative update process, maintain periodic value measurement reporting, and use the periodic value measurement as the calibration value of the data set;

[0044] Utilize the goodness of fit of the measurement model to the data, continuously update the accuracy of the machine learning model, and obtain an optimized handover result to ensure the superiority of the handover algorithm;

[0045] Apply the optimized handover result to the substation trusted WLAN handover system to achieve efficient and stable wireless network handover decision-making.

[0046] In a second aspect, a handover device for a substation trusted WLAN provided by an embodiment of the present invention includes:

[0047] A data acquisition module, configured to collect historical network handover data of the substation, generate a time series feature matrix as an input data set, and the input data set includes received power, delay, transmission rate, and network load;

[0048] A data processing module, configured to perform normalization processing on the received power, delay, transmission rate, and network load, and calculate the weights of the four attributes according to the similarity matrix (ANP);

[0049] A data prediction module, configured to calculate the multi-attribute effect value, fill the corresponding value of the training data into the input data set, select an improved random forest (ES-RF) model as the prediction tool, and predict the values of the current serving network and the network to be handed over at the next moment based on the historical feature data;

[0050] A network switching module, configured to select a target network for switching when the difference in utility values is greater than a switching threshold and reaches the delay set by a timer, and the load of the target switching network does not exceed 80%;

[0051] A parameter adjustment module, configured to fill the predicted value of the target network into the output data set to make it a new input data set for iterative update, and optimize the prediction effect by continuously adjusting the parameters of the prediction model.

[0052] In a third aspect, an electronic device provided by an embodiment of the present invention includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device runs, the processor communicates with the memory through the bus, and the processor executes the machine-readable instructions to perform the steps of any of the above-mentioned switching methods for a trusted WLAN in a substation.

[0053] In a fourth aspect, a storage medium provided by an embodiment of the present invention stores a computer program, and when the computer program is run by a processor, it performs the steps of any of the above-mentioned switching methods for a trusted WLAN in a substation.

[0054] The beneficial effects of the technical solution of the embodiment of the present invention are as follows:

[0055] The present invention adopts a multi-attribute decision-making model based on key performance indicators such as received power, delay, transmission rate, and network load, avoiding the ping-pong effect caused by single-attribute switching and improving the accuracy of switching decisions. At the same time, by using the sliding window method and time series analysis, the switching trigger conditions are optimized, reducing the negative impact of feedback delay on switching efficiency. In addition, the present invention adopts an improved random forest model, enhancing the adaptability of the decision-making and ensuring that the switching algorithm still has a high success rate and network stability in a complex environment.

[0056] The present invention combines the ANP similarity matrix and the supermatrix to optimize the weight allocation, taking into account the user's subjective preferences and actual data to ensure that the switching decision is more personalized and efficient. To further improve the utilization rate of network resources, the present invention introduces a limit on network load in the switching conditions, avoiding overloading of the target network. By introducing a feedback mechanism and using the mean square error to evaluate the model performance, the accuracy of the machine learning model is continuously optimized to ensure that the switching algorithm self-improves during long-term use, thereby improving the overall operation effect of the system.

[0057] The strong electromagnetic environment in a substation poses higher requirements for handover technology. To adapt to this complex environment, it is necessary to optimize handover performance to ensure the stability and reliability of power grid operation. By introducing a multi-attribute utility model and machine learning algorithms, the present invention accurately predicts handover decision parameters, effectively reduces the impact of system feedback delay on handover success rate, and reduces the handover failure rate, providing a solution to the high requirements for handover efficiency in a trustworthy WLAN handover under strong electromagnetic interference in a substation.

[0058] The present invention selects the optimal handover scheme by comparing the utility values of each candidate handover scheme, ensuring the efficient and stable operation of the power system in a complex environment. By comprehensively considering multiple attributes such as received power, transmission rate, delay, and network load, and combining with a machine learning model, the present invention more efficiently and reliably realizes the optimal handover in the substation operation scenario. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 is a flowchart of a method for handover of a trustworthy WLAN in a substation shown according to an exemplary embodiment;

[0060] Figure 2 is a display diagram of a handover of a trustworthy WLAN in a substation shown according to an exemplary embodiment;

[0061] Figure 3 is a schematic structural diagram of a handover device of a trustworthy WLAN in a substation shown according to an exemplary embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0062] To more clearly illustrate the technical features of the solution of the present invention, the present invention will be described in detail below through specific embodiments and in conjunction with its accompanying drawings.

[0063] As Figure 1 shown, a method for handover of a trustworthy WLAN in a substation provided by an embodiment of the present invention includes the following steps:

[0064] Step S1, collect historical network handover data of the substation, generate a time series feature matrix as an input data set, and the input data set includes received power, delay, transmission rate, and network load;

[0065] Step S2, perform normalization processing on received power, delay, transmission rate, and network load, and calculate the weights of the four attributes according to the similarity matrix (ANP);

[0066] Step S3, calculate the multi-attribute effect value, supplement the corresponding values of the training data into the input data set, select an improved random forest (ES-RF) model as a prediction tool, and predict the values of the current serving network and the network to be handed over at the next moment based on the historical feature data;

[0067] Step S4, when the difference in utility values is greater than the handover threshold and reaches the delay set by the timer, and the load of the target handover network does not exceed 80%, select the target network for handover;

[0068] Step S5, fill the predicted value of the target network into the output data set to make it a new input data set for iterative update, and optimize the prediction effect by continuously adjusting the parameters of the prediction model.

[0069] As a possible implementation manner of this embodiment, the step S1 includes:

[0070] Select handover decision parameters, where the handover decision parameters include the received power collected at the physical layer, the delay, transmission rate, and network load concerned at the network layer;

[0071] Periodically capture handover data packets using the measurement configuration and report them;

[0072] Use the sliding window method to process the reported handover data packet data to generate a time series feature matrix;

[0073] Use the time series feature matrix as the training input data set to capture the historical network performance.

[0074] The present invention first selects four handover decision parameters: the parameter collected at the physical layer is the received power, and the key performance indicators concerned at the network layer are the delay, transmission rate, and network load. During the inspection process, the inspection robot will periodically capture handover data packets using software such as Wireshark according to the measurement configuration, process these data using the sliding window method to generate a time series feature matrix, and generate a training input data set, thereby capturing the historical network performance. By selecting multiple network parameters as handover decision parameters, the present invention can more comprehensively reflect the change trend of the network condition, improve the accuracy and timeliness of handover decisions; using the sliding window method to process data can capture the time dependence of the data and improve the accuracy and reliability of the prediction model; applied to the network handover decision during the inspection process of the inspection robot, it has a wide range of application prospects.

[0075] As a possible implementation manner of this embodiment, the step S2 includes:

[0076] Construct a similarity matrix, where the similarity matrix is used to represent the importance of an attribute relative to an attribute;

[0077] Based on the similarity matrix, further calculate the weight of each network attribute by analyzing historical data;

[0078] Perform a consistency check and calculate the consistency ratio of the judgment matrix;

[0079] Further optimize the weights using a hypermatrix to better reflect the feedback and interactions between network attributes;

[0080] Calculate the utility value of historical data using a Quality of Experience (QoE) multi-attribute utility linear weighted model. The calculation formula for the utility value is:

[0081]

[0082] where P recv (n) is the received power, D(n) is the delay, L(n) is the network load, R(n) is the transmission rate, and Q t is the comprehensive utility value of the i-th handover network, satisfying w 1 + w 2 + w 3 + w 4 = 1 and where k is the interaction threshold to prevent the excessive influence of the power and rate interaction terms;

[0083] Use the calculated utility value to improve the training dataset of machine learning.

[0084] The present invention comprehensively considers the similarity, weight assignment, consistency, and comprehensive utility value between network attributes, improving the accuracy of model prediction; by introducing a hypermatrix to optimize the weights, it better reflects the feedback and interactions between network attributes; multiple attributes are considered in the calculation formula and an interaction threshold is set to avoid the problem of excessive influence; the training dataset of machine learning is improved, enhancing the training effect of the model.

[0085] As a possible implementation of this embodiment, the calculation formula for the weight of each network attribute is:

[0086]

[0087] where represents the weight of the i-th attribute, and C ij represents the element in the i-th row and j-th column of the C matrix.

[0088] As a possible implementation of this embodiment, step S3 includes:

[0089] Introduce an improved random forest model and dynamically adjust its weights according to the contribution degree of each feature to the utility value. The weights are updated based on the performance of historical data and the current network state;

[0090] Introduce a weight vector to dynamically update the importance of four key attributes to form an adaptive feature input;

[0091] Adopt a hierarchical sampling strategy to divide the samples into different categories and uniformly sample from each category to ensure that the training data is representative.

[0092] The advantages of the present invention are as follows: By dynamically adjusting the feature weights, the prediction accuracy is improved; By the hierarchical sampling strategy, the generalization ability of the model is enhanced; Using the time-dependent characteristics of the measurement parameters for handover prediction reduces the impact of system feedback delay on the handover success rate, and is particularly suitable for reliable WLAN handover in the strong electromagnetic interference environment of substations.

[0093] As a possible implementation manner of this embodiment, the step S4 includes:

[0094] Determine a reliable handover decision model and set a given handover threshold.

[0095] Judge whether the following conditions are simultaneously satisfied:

[0096] (a) The difference between the utility value of the target network and the utility value of the currently connected network is greater than the given handover threshold and reaches the delay set by the timer.

[0097] (b) The load of the target handover network does not exceed 80%.

[0098] When conditions (a) and (b) are simultaneously satisfied, determine to perform a handover.

[0099] The present invention improves the reliability and stability of handover decisions by comprehensively considering the handover delay and the load of the target handover network; Avoids unnecessary handovers in high-load or unstable network environments, reducing the risk of service interruption and data loss; Improves the user experience and ensures that users obtain continuous and stable services during the network handover process.

[0100] As a possible implementation manner of this embodiment, the step S5 includes:

[0101] Fill the predicted value of the target network into the output data set to make the output data set become a new input data set.

[0102] Use the new input data set to perform iterative updates of steps S1 to S4.

[0103] During the iterative update process, maintain periodic value measurement reporting and use the periodic value measurement as the calibration value of the data set.

[0104] Utilize the goodness of fit of the model to the data to continuously update the accuracy of the machine learning model and obtain optimized handover results to ensure the superiority of the handover algorithm.

[0105] Apply the optimized handover result to the substation trusted WLAN handover system to achieve efficient and stable wireless network handover decision-making.

[0106] The present invention improves the efficiency and stability of wireless network handover. Through iterative optimization and machine learning techniques, the handover strategy is dynamically adjusted to adapt to the complex and changing network environment; enhances the adaptability and superiority of the handover algorithm, and ensures the accuracy and reliability of handover decisions by continuously updating the accuracy of the machine learning model; provides a periodic value measurement reporting mechanism as the calibration value of the data set, further improving the accuracy and stability of handover decisions.

[0107] As Figure 2 shown, a handover device for a substation trusted WLAN provided by an embodiment of the present invention includes:

[0108] A data acquisition module, configured to collect historical network handover data of a substation, generate a time series feature matrix as an input data set, and the input data set includes received power, delay, transmission rate, and network load;

[0109] A data processing module, configured to perform normalization processing on received power, delay, transmission rate, and network load, and calculate the weights of the four attributes according to the similarity matrix (ANP);

[0110] A data prediction module, configured to calculate the multi-attribute effect value, fill the corresponding value of the training data into the input data set, select an improved random forest (ES-RF) model as a prediction tool, and predict the values of the current serving network and the network to be handed over at the next moment based on historical feature data;

[0111] A network handover module, configured to select a target network for handover when the difference in utility values is greater than the handover threshold and reaches the delay set by the timer, and the load of the target handover network does not exceed 80%;

[0112] A parameter adjustment module, configured to fill the predicted value of the target network into the output data set to make it a new input data set for iterative update, and optimize the prediction effect by continuously adjusting the parameters of the prediction model.

[0113] The present invention provides a trusted WLAN handover algorithm based on multi-attribute decision-making to solve the following two technical problems existing in the prior art.

[0114] (1) The ping-pong effect is serious. When the utility function value or the received signal strength fluctuates around the handover threshold, if the traditional handover algorithm is used, the terminal will switch back and forth between the two access networks, resulting in a decline in system performance and a waste of network resources. The handover algorithm proposed by the present invention can effectively alleviate the impact of the ping-pong effect and improve the user experience quality.

[0115] (2) The network parameter weights are unreasonable. The existing credible WLAN handover algorithms for multi-attribute decision-making are mainly implemented based on the analytic hierarchy process or the entropy weight method. The former adopts the idea of expert scoring, which will lead to the weights of each parameter being too subjective, while the latter will lead to the weights being too objective. The credible WLAN handover algorithm proposed by the present invention combines random forest prediction and multi-attribute decision-making, improving the rationality of the network parameter weights.

[0116] As Figure 3 shown, the specific implementation process of the present invention for the handover of the credible WLAN in the substation is as follows.

[0117] First, collect the historical data of four key network attributes from the network devices in the substation: received power, delay, transmission rate, and network load, and store them through a database or other storage devices.

[0118] To capture handover data packets, an extended antenna can be used. The adapter Riverbed AirPcap Nx802.11 can be used to capture management, control, and data packets from multiple channels of the 802.11 standard, and these packets can be viewed together with Wireshark or Eye P.A.

[0119] After collecting the historical data, through the sliding window method, the historical time series data of each attribute will be converted into a feature matrix X(t) suitable for the input of the random forest:

[0120]

[0121] where P is the received power, D is the delay, L is the network load, and R is the transmission rate.

[0122] The sliding window method takes multiple data points within a period of time as the input of one time step to capture the time dependence of the data. Here, t is the current time step, and t - 4 to t represent the historical data of the past 5 time steps. Through this matrix, the system can capture the changing trends of each attribute over time and provide input for the prediction model.

[0123] Set the time step, the initial position of the user equipment, the received power, and the number of loops, and the path loss model where PL 0 is the path loss at the reference distance d 0 , usually taking the loss at 1 meter, d is the distance between the user and the base station, n is the path loss exponent, and X δ is the additional attenuation amount considering the environmental impact, used to simulate the additional loss caused by the strong electromagnetic environment in the substation.

[0124] Normalize the collected network attributes to ensure that all attributes are in the same dimension. The normalization formula is as follows:

[0125]

[0126] where f i (x) is the normalized value, x is the original value, and X is the total set of attributes.

[0127] While processing the data, it is also necessary to construct an ANP similarity matrix W for subsequent weight calculation. First, expert scoring is used to reflect the relative importance between each network attribute. According to the specific requirements and environment of the substation, compare the importance of each attribute in combination with Table 1.

[0128] Table 1 Comparison Table of Element Importance Degrees

[0129] 1 Indicates that the importance of two elements is the same 3 Indicates that one element is slightly more important than the other 5 Indicates that one element is significantly more important than the other 7 Indicates that one element is strongly more important than the other 9 Indicates that one element is extremely more important than the other 2、4、6、8 Indicates the median of the above importance levels

[0130] When constructing the judgment matrix, through the 1-9 scale method, convert the qualitative problem into a quantitative problem, and then use these scales to construct the judgment matrix C:

[0131]

[0132] where, c 11 , c 22 , c 33 , c 44 are usually set to 1 because the comparison of any attribute with itself is always equal. c ij If it is greater than 1, it means that attribute i is more important than attribute j. c ij If it is less than 1, it means that attribute i is less important than attribute j, ensuring the consistency of the matrix.

[0133] Using the C matrix, the weight of each network attribute can be further calculated:

[0134]

[0135] where, represents the weight of the i-th attribute, and c ij represents the element in the i-th row and j-th column of the C matrix. By analyzing historical data, the weight of each network attribute is further calculated. Although this method has a certain degree of subjective randomness, it can fully ensure user preferences. Since the judgment matrix is obtained by subjectively comparing the importance degrees of parameters, it is necessary to conduct a consistency test and calculate the consistency ratio of the judgment matrix W:

[0136] CR = λ max -n / RI(n - 1) (5)

[0137] Among them, λ max is the largest eigenvalue of W, RI is the average consistency index, which can be obtained by looking up the table, and n is the number of network parameters (n = 4). When CR < 0.1, it indicates that the consistency of the judgment matrix is acceptable; otherwise, the judgment matrix needs to be reconstructed. Conduct a consistency test and calculate the consistency ratio of the judgment matrix. To better reflect the feedback and interaction between network attributes, a supermatrix is introduced to further optimize the weights.

[0138] The similarity matrix provides the direct relative importance between each network attribute, but the complex interdependent relationships between network attributes are captured by the supermatrix. The supermatrix can further optimize the weights to better reflect the feedback and interaction between network attributes.

[0139]

[0140] For example, network load affects latency, and conversely, an increase in latency may also lead to an increase in network load. To ensure that the weights of each attribute are reasonable, the supermatrix needs to be normalized. Normalize each column so that the sum of each column is 1 to obtain the new weight w i .

[0141] Introduce a multi-attribute decision-making model, normalize each attribute, and then calculate the network contribution degree. The utility value is a quantitative indicator for evaluating the overall performance of the network, and its level will directly affect whether the system chooses to switch networks. The formula is as follows:

[0142]

[0143] Among them, the received power is P recv (n), the latency is D(n), the network load is L(n), the transmission rate is R(n), and Q t is the comprehensive utility value of the i-th switched network. Satisfy w 1 +w 2 +w 3 +w 4 = 1 and where k is the interaction threshold to prevent the excessive influence of the power and rate interaction terms.

[0144] In the time series prediction task, the random forest uses the data of historical time steps as input to predict the network state of future time steps.

[0145] Input data: The feature matrix X(t) consists of data of n time steps, and each row represents the historical value of a network attribute.

[0146] Output data: The network utility value at the next time step \(t + 1\). The ES-RF model dynamically adjusts its weights according to the contribution of each feature to the utility value. The mathematical representation of the training process is as follows:

[0147] For the four key attributes, by introducing the weight vector \(W=[w P ,w D ,w L ,w R \), the importance of each feature is updated dynamically.

[0148] The adaptive feature input can be expressed as:

[0149] \(X = [w P \cdot P recv ,w D \cdot D,w L \cdot L,w R \cdot R]\ (8)

[0150] Among them, the weights are updated according to the performance of historical data and the current network state, so as to improve the prediction accuracy in different scenarios. In ES-RF, the voting weights are weighted based on the prediction errors of each tree, so that the trees with better prediction effects have a greater impact on the final decision. The weight calculation formula is as follows:

[0151]

[0152] Among them, \(MSE i \) represents the mean squared error of the \(i\)-th tree. The smaller the mean squared error of a tree, the greater its weight. The final comprehensive utility value \(Q switch \) is obtained through weighted voting:

[0153]

[0154] Here, \(h i (X)\) represents the prediction result of the \(i\)-th tree.

[0155] At the same time, the ES-RF model adopts a stratified sampling strategy. This strategy divides the samples into different categories and uniformly samples from each category to ensure that the training data is representative, thereby enhancing the generalization ability of the model.

[0156] Because the robot's driving route is fixed, and measurement parameters such as the Reference Signal Received Power (RSRP) have strong spatio-temporal correlation, the time-dependent characteristics of the measurement parameter sequence can be used to predict the decision parameters at future time steps, thereby promoting handover triggering and reducing the impact of system feedback delay on the handover success rate, providing a solution for the high requirements of reliable WLAN handover under strong electromagnetic interference in substations for handover efficiency.

[0157] At each time step, based on the predicted network utility value, the optimal network is selected for connection. If the difference between the utility value of the target network and the utility value of the currently connected network is higher than the set threshold and the network load is less than 80%, a handover is triggered:

[0158]

[0159] where Q 1 and Q 2 are the network utility values of WiFi 1 and WiFi 2 respectively. During the iteration process, according to the preset number of iterations, data loading and preprocessing, calculation of attribute weights, application of an improved random forest model, and comprehensive evaluation of the multi-attribute utility function are carried out to complete the handover. At the same time, periodic Q-value measurement reports are still maintained as calibration values for the dataset. Using to measure the goodness of fit of the model to the data, continuously update the machine learning accuracy to ensure the superiority of the handover algorithm; through this series of iterative optimizations, apply the handover result H q =l to the substation trusted WLAN handover system to achieve efficient and stable wireless network handover decisions.

[0160] The present invention comprehensively considers multiple key performance indicators, avoids the ping-pong effect caused by single-attribute handovers, and improves the accuracy of handover decisions; optimizes the handover trigger conditions through the sliding window method and time series analysis, reducing the negative impact of feedback delay on handover efficiency; adopts an improved random forest model to enhance the adaptability of decisions, ensuring high success rates and network stability of the handover algorithm in complex environments; combines the ANP similarity matrix and supermatrix to optimize weight allocation, taking into account user subjective preferences and actual data to ensure more personalized and efficient handover decisions; introduces network load limitations to avoid overloading of the target network and improve the utilization rate of network resources.

[0161] An electronic device provided by an embodiment of the present invention includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device runs, the processor communicates with the memory through the bus, and the processor executes the machine-readable instructions to perform the steps of the handover method of any of the above substation trusted WLANs.

[0162] Specifically, the above memory and processor can be general-purpose memory and processor, which are not specifically limited here. When the processor runs the computer program stored in the memory, it can execute the handover method of the above substation trusted WLAN.

[0163] Those skilled in the art can understand that the structure of the computer device does not constitute a limitation on the computer device, and it may include more or fewer components than shown in the figure, or combine some components, or split some components, or have different component arrangements.

[0164] In some embodiments, the computer device may further include a touch screen, which can be used to display a graphical user interface (e.g., the startup interface of an application) and receive user operations on the graphical user interface (e.g., the startup operation for an application). Specifically, the touch screen may include a display panel and a touch panel. The display panel can be configured in forms such as LCD (Liquid Crystal Display) and OLED (Organic Light-Emitting Diode). The touch panel can collect contact or non-contact operations of the user on or near it and generate preset operation instructions. For example, the user uses any suitable object such as a finger, a stylus, or an accessory to operate on or near the touch panel. Additionally, the touch panel may include two parts: a touch detection device and a touch controller. Among them, the touch detection device detects the touch orientation and posture of the user, and detects the signal brought by the touch operation, and transmits the signal to the touch controller; the touch controller receives the touch information from the touch detection device, converts it into information that the processor can process, then sends it to the processor, and can receive and execute the commands sent by the processor. Furthermore, various types such as resistive, capacitive, infrared, and surface acoustic wave can be used to implement the touch panel, and any technology developed in the future can also be used to implement the touch panel. Further, the touch panel can cover the display panel, and the user can operate on or near the touch panel covering the display panel according to the graphical user interface displayed on the display panel. After the touch panel detects the operation on or near it, it transmits it to the processor to determine the user input, and then the processor provides a corresponding visual output on the display panel in response to the user input. Additionally, the touch panel and the display panel can be implemented as two independent components or integrated.

[0165] Corresponding to the above application startup method, an embodiment of the present invention further provides a storage medium, on which a computer program is stored. When the computer program is run by a processor, it executes the steps of any of the above switching methods for a substation trusted WLAN.

[0166] The startup device of the application provided by the embodiments of the present application can be specific hardware on the device, or software or firmware installed on the device, etc. For the device provided by the embodiments of the present application, its implementation principle and the resulting technical effects are the same as those of the foregoing method embodiments. For the sake of brief description, for the parts not mentioned in the device embodiments, reference can be made to the corresponding content in the foregoing method embodiments. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the foregoing described systems, devices, and units can all refer to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0167] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0168] In the embodiments provided by the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only illustrative. For example, the division of modules is only a logical function division, and there can be other division methods in actual implementation. For another example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed couplings, direct couplings, or communication connections to each other can be through some communication interfaces. The indirect couplings or communication connections of devices or modules can be in electrical, mechanical, or other forms.

[0169] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they can be located in one place, or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0170] In addition, the various functional modules in the embodiments provided by the present application can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0171] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices generate means for implementing the functions specified in one or more of the flows Figure 1 one or more of the flows and / or blocks Figure 1 or means for implementing the functions specified in one or more of the blocks.

[0172] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means for implementing the functions specified in one or more of the flows Figure 1 one or more of the flows and / or blocks Figure 1 or means for implementing the functions specified in one or more of the blocks.

[0173] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more of the flows Figure 1 one or more of the flows and / or blocks Figure 1 or means for implementing the functions specified in one or more of the blocks.

[0174] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that the specific implementation manners of the present invention can still be modified or equivalently replaced, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the protection scope of the claims of the present invention.

Claims

1. A switching method for a trusted WLAN in a substation, characterized in that: The steps include: Step S1, collecting historical network switching data of the substation and generating a time series feature matrix as an input data set, wherein the input data set includes received power, delay, transmission rate and network load; Step S2, normalizing the received power, delay, transmission rate and network load, and calculating the weights of the four attributes according to the similarity matrix; Step S3, calculating the multi-attribute effect value, adding the corresponding value of the training data to the input data set, selecting the improved random forest model as a prediction tool, and predicting the values ​​of the current serving network and the network to be switched at the next moment based on the historical feature data; Step S4, when the difference in utility values ​​is greater than the switching threshold and reaches the delay set by the timer, and the load of the target switching network does not exceed 80%, select the target network for switching; Step S5, fill the predicted value of the target network into the output data set to make it a new input data set for iterative update, and optimize the prediction effect by continuously adjusting the parameters of the prediction model.

2. The switching method of the substation trusted WLAN according to claim 1, characterized in that: The step S1 comprises: Selecting a switching decision parameter, wherein the switching decision parameter includes a received power collected at the physical layer, a delay concerned at the network layer, a transmission rate, and a network load; Use measurement configuration to periodically capture switching data packets and report them; Use the sliding window method to process the reported handover packet data and generate a time series feature matrix; The time series feature matrix is ​​used as a training input dataset to capture historical network performance.

3. The switching method of the substation trusted WLAN according to claim 1, characterized in that: The step S2 comprises: constructing a similarity matrix, wherein the similarity matrix is ​​used to represent the importance of attributes relative to each other; By analyzing historical data, the weight of each network attribute is further calculated based on the similarity matrix; Conduct consistency test and calculate the consistency ratio of the judgment matrix; further optimizing the weights using a supermatrix to better reflect feedback and interactions between network attributes; The utility value of historical data is calculated using the quality experience multi-attribute utility linear weighted model. The calculation formula of the utility value is: Among them, P recv (n) is the received power, D(n) is the delay, L(n) is the network load, R(n) is the transmission rate, Q t is the comprehensive utility value of the i-th switching network, satisfying w1+w2+w3+w4=1 and Where k is the interaction threshold, which prevents excessive influence of the interaction term between power and rate; The calculated utility values ​​are used to improve the training data set for machine learning.

4. The switching method of the substation trusted WLAN according to claim 3 is characterized in that: The weight of each network attribute is calculated as: in, represents the weight of the i-th attribute, C ij Represents the element in the i-th row and j-th column of the C matrix.

5. The switching method of the substation trusted WLAN according to claim 1, characterized in that: The step S3 comprises: An improved random forest model is introduced, and the weight of each feature is dynamically adjusted according to its contribution to the utility value. The weight is updated according to the performance of historical data and the current network status. Introducing a weight vector to dynamically update the importance of four key attributes to form an adaptive feature input; A stratified sampling strategy is adopted to divide the samples into different categories and sample evenly from each category to ensure that the training data is representative.

6. The switching method of the substation trusted WLAN according to claim 1, characterized in that: The step S4 comprises: Determine a trusted switching decision model and set a given switching threshold; Determine whether the following conditions are met at the same time: (a) The difference between the utility value of the target network and the utility value of the currently connected network is greater than the given switching threshold and reaches the delay set by the timer; (b) The load of the target switching network does not exceed 80%; When both conditions (a) and (b) are satisfied, switching is determined.

7. The switching method of a substation trusted WLAN according to any one of claims 1 to 6, characterized in that: The step S5 comprises: Filling the predicted value of the target network into the output data set, so that the output data set becomes a new input data set; Perform iterative updates of steps S1 to S4 using the new input data set; During the iterative update process, periodic value measurement reporting is maintained, and the periodic value measurement is used as a calibration value of the data set; By measuring the goodness of fit of the model to the data, the accuracy of the machine learning model is continuously updated to obtain optimized switching results to ensure the superiority of the switching algorithm; The optimized switching results are applied to the trusted WLAN switching system of substations to achieve efficient and stable wireless network switching decisions.

8. A switching device for a trusted WLAN in a substation, characterized in that: include: A data acquisition module, used to collect historical network switching data of the substation and generate a time series feature matrix as an input data set, wherein the input data set includes received power, delay, transmission rate and network load; The data processing module is used to normalize the received power, delay, transmission rate and network load, and calculate the weights of the four attributes according to the similarity matrix; The data prediction module is used to calculate the multi-attribute effect value, add the corresponding value of the training data to the input data set, select the improved random forest model as the prediction tool, and predict the value of the current service network and the network to be switched at the next moment based on the historical feature data; A network switching module, used for selecting a target network for switching when the difference in utility values ​​is greater than a switching threshold and reaches a delay set by a timer, and the load of the target switching network does not exceed 80%; The parameter adjustment module is used to fill the predicted value of the target network into the output data set, turning it into a new input data set for iterative update, and continuously adjusting the parameters of the prediction model to optimize the prediction effect.

9. An electronic device, characterized in that: The electronic device comprises a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory through the bus, and the processor executes the machine-readable instructions to perform the steps of the switching method of the substation trusted WLAN as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the switching method for a substation trusted WLAN as claimed in any one of claims 1 to 7 are executed.

Images