Network adaptive handover method and device, storage medium and network equipment

Abstract

The application discloses a network adaptive switching method and device, a storage medium and a network device. The method comprises the following steps: acquiring network exception features of a target network when network exception occurs in a preset time period and a network parameter set of the target network for processing target user service of a target user terminal in the preset time period; screening at least one target network attribute affecting the target user service from a plurality of network attributes in the network parameter set according to the network exception features; integrating network parameters of each target network attribute at each moment in the preset time period to obtain performance estimation values of the target network at each moment; and performing network switching management on the target user terminal according to the performance estimation values of the target network at each moment. The application solves the technical problem that the related art does not consider that network attributes are not uniform in different scenarios when determining the weight of the network attributes, so that the decided network switching node is inaccurate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless communication, in particular to a network adaptive switching method and device, a storage medium and a network equipment. BACKGROUND

[0002] At present, in order to meet the complex service demand of people for mobile communication, the 3rd Generation Partnership Project (3GPP) proposes a heterogeneous network, that is, by introducing various micro base stations to change the traditional single-layer network structure to build a multi-technology and multi-protocol network system. However, due to the limited coverage of a single network, the communication network quality will inevitably change with the change of user location information during the heterogeneous network switching. Therefore, in the existing heterogeneous network environment, how to realize the seamless switching of user network by using the difference of network quality is a key technical problem that needs to be solved in the development of heterogeneous network. In the heterogeneous network environment, the switching process mainly includes session logout and re-registration of the original network, during which the original network is interrupted and accompanied by ping-pong effect. In order to solve this problem, a large number of researches have been carried out at home and abroad in recent years, which can be divided into the following categories according to the decision-making method:

[0003] (1) Threshold-based switching: it mainly uses RSS (Received Signal Strength) as the evaluation characteristic parameter, and the system performs network switching when the RSS exceeds the threshold. However, this method does not consider the dynamic change of network, so it cannot meet the more complex heterogeneous network environment, and it is easy to produce ping-pong effect in the actual application process;

[0004] (2) Multi-attribute decision-based switching: it is based on deep learning, reinforcement learning, neural network and other algorithms to realize the switching of heterogeneous network. After training, evaluation and optimization of historical data model, the network switching decision is made according to the final optimized model. However, since this method involves many algorithms and has a complex decision structure, the training set directly affects the accuracy of subsequent switching decision;

[0005] (3) Artificial intelligence-based switching: it is a comprehensive switching decision by measuring multiple network attributes, such as delay, RSS, packet loss rate, bandwidth, and bit error rate. This method has high network switching accuracy and high feasibility. However, this kind of multi-attribute-based network switching method does not inform the network switching node, resulting in a long calculation time in the early stage of the switching process and greatly increasing the occurrence of ping-pong effect.

[0006] In summary, to some extent, the existing strategy method can realize the switching of heterogeneous network, but there are still many deficiencies. In view of the above problems, there is no effective solution at present. Summary of the Invention

[0007] This application provides a network adaptive switching method and apparatus, storage medium, and network device to at least solve the technical problem that related technologies do not consider the non-uniformity of network attributes in different scenarios when determining the weight of network attributes, resulting in inaccurate network switching node decisions.

[0008] According to one aspect of the embodiments of this application, a network adaptive handover method is provided, comprising: acquiring network anomaly characteristics of a target network when a network anomaly occurs within a preset time period and a set of network parameters of the target network processing target user services of a target user terminal within the preset time period, wherein the set of network parameters includes: a subset of network parameters at multiple times within the preset time period, and the subset of network parameters includes network parameters under at least one network attribute; filtering at least one target network attribute affecting the target user service from multiple network attributes in the set of network parameters based on the network anomaly characteristics; integrating the network parameters of each target network attribute at each time within the preset time period to obtain a performance estimate of the target network at each time; and performing network handover management on the target user terminal based on the performance estimate of the target network at each time.

[0009] Optionally, the service type of the target user's service includes at least: low-latency communication, wide-connectivity communication, or high-bandwidth communication; the network attribute type includes at least one of the following: loopback latency, network signal strength, network bandwidth, and packet loss rate.

[0010] Optionally, based on network anomaly characteristics, at least one target network attribute affecting the target user's service is selected from multiple network attributes within the network parameter set. This includes: calculating the first mutual information between each network attribute and the network anomaly characteristics, and the second mutual information between each network attribute; using the maximum correlation minimum redundancy (mRMR) algorithm to analyze the first mutual information between each network attribute and the network anomaly characteristics, and the second mutual information between each network attribute, to select multiple key network attributes affecting the target user's service from multiple network attributes within the network parameter set; constructing an objective function with the optimization objective of maximizing the first mutual information between the key network attributes and the network anomaly characteristics and minimizing the second mutual information between each key network attribute, and solving the objective function using a preset optimization algorithm, to select at least one target network attribute from multiple key network attributes.

[0011] Optionally, the objective function is: In the formula y=[y1,y2,...y p ] represents the indicator value of multiple key network attributes, and the indicator value y ifor reflecting whether the ith key network attribute is selected as a target network feature; d represents the number of target network attributes; D ′ represents a correlation matrix between the plurality of key network attributes and the network anomaly feature; R ′ represents a redundancy matrix between the plurality of key network attributes.

[0012] Optionally, the network parameters of each target network attribute at each time point in a preset time period are integrated to obtain a performance estimation value of the target network at each time point, including: for each time point in the preset time period, determining a zero-mean score noise of each target network attribute at the time point, and determining a weight of each target network attribute according to the zero-mean score noise of each target network attribute at the time point; calculating a sum value of the zero-mean score noise and the network parameter of each target network attribute at the time point, and taking a sum of products of the sum value of each target network attribute at the time point and the weight of the target network attribute as the performance estimation value of the target network at the time point.

[0013] Optionally, the weight of each target network attribute is determined according to the zero-mean score noise of each target network attribute at the time point, including: determining an unbiased estimator of the variance of the zero-mean score noise of each target network attribute at the time point; and calculating the weight of each target network attribute according to the unbiased estimator of the variance of the zero-mean score noise of each target network attribute at the time point through the following formula: wherein represents the weight of the kth target network attribute at the tth time point, represents the unbiased estimator of the variance of the zero-mean score noise of the kth target network attribute at the tth time point.

[0014] Optionally, the target user terminal is managed for network switching according to the performance estimation value of the target network at each time point, including: counting the number of time points at which the performance estimation value of the target network is lower than a preset threshold value in a preset time period; and in the case that the number of time points exceeds a preset number threshold, vertically switching the target user terminal from the target network to other heterogeneous networks.

[0015] According to another aspect of the embodiments of the present application, a network adaptive switching device is further provided, which comprises: an acquisition module, configured to acquire network abnormality features of a target network when network abnormality occurs in a preset time period and a network parameter set of the target network for processing target user service of a target user terminal in the preset time period, wherein the network parameter set comprises network parameter subsets at multiple time points in the preset time period, and each network parameter subset comprises network parameters at at least one network attribute; a determination module, configured to filter at least one target network attribute affecting the target user service from the multiple network attributes in the network parameter set according to the network abnormality features; an evaluation module, configured to integrate the network parameters of each target network attribute at each time point in the preset time period to obtain performance estimation values of the target network at each time point; and a management module, configured to perform network switching management on the target user terminal according to the performance estimation values of the target network at each time point.

[0016] According to another aspect of the embodiments of the present application, a nonvolatile storage medium is further provided, which comprises a stored computer program, wherein a device where the nonvolatile storage medium is located performs the network adaptive switching method described above by running the computer program.

[0017] According to another aspect of the embodiments of the present application, a network device is further provided, which comprises a memory and a processor, and the processor is configured to run a program stored in the memory, wherein the program performs the network adaptive switching method described above when running.

[0018] In the embodiments of the present application, the control system analyzes the network abnormality features of the target network when network abnormality occurs in the preset time period and intelligently filters the network parameter set, so as to identify and focus on the key network attributes affecting the target user service; the network parameters of the key network attributes at each time point are integrated to obtain accurate comprehensive performance estimation values generated dynamically, and the network switching decision is determined based on the comprehensive performance estimation values of the target network at different time points, so as to achieve the purposes of improving user experience and ensuring service continuity, thereby solving the technical problem that the related art does not consider the non-uniformity of network attributes in different scenarios when determining the weight of the network attributes, resulting in inaccurate network switching nodes. BRIEF DESCRIPTION OF DRAWINGS

[0019] The accompanying drawings, which are included to provide a further understanding of the present application, constitute a part of the present application and illustrate the illustrative embodiments of the present application and their description serve to explain the present application, but do not constitute improper limitations on the present application. In the drawings:

[0020] Figure 1 is a flow schematic diagram of an optional network adaptive switching method according to the embodiments of the present application;

[0021] Figure 2 is a schematic diagram of an optional network adaptive switching system according to an embodiment of the application;

[0022] Figure 3 is a schematic diagram of an optional network adaptive switching device according to an embodiment of the application;

[0023] Figure 4 is a schematic diagram of an optional network device according to an embodiment of the application. DETAILED DESCRIPTION

[0024] In order to make the personnel in the art better understand the scheme of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor should be within the scope of protection of the present application.

[0025] It should be noted that the terms "first", "second", and the like in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0026] In addition, the relevant information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, analyzed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties. For example, an interface is provided between the system and the relevant user or institution. Before obtaining the relevant information, the interface needs to send a request to the aforementioned user or institution, and after receiving the consent information fed back by the aforementioned user or institution, the relevant information is obtained.

[0027] In order to better understand the embodiments of the present application, the technical terms involved in the embodiments of the present application are explained as follows:

[0028] Heterogeneous Network: A type of network that is composed of computers, network devices and systems produced by different manufacturers. That is, different types of networks that overlap with each other can form a heterogeneous network to meet the diverse needs of future terminals. In order to access multiple networks at the same time, mobile terminals should have interfaces that can access multiple networks. Such mobile terminals are called multi-mode terminals. Since multi-mode terminals can access multiple networks, switching between different networks is certainly designed. Unlike horizontal handoff (HHO) in homogeneous wireless networks, the switching between different communication systems is called vertical handoff (VHO) here.

[0029] Ping-pong Effect: refers to the change between two different states. In a mobile communication system, if the signal strength of two base stations in a certain area changes dramatically, the mobile terminal will switch between the two base stations, resulting in a so-called "ping-pong effect".

[0030] Mutual Information: an important concept in probability wheel and information theory, used to measure the degree of mutual dependence between two random variables. Unlike the correlation coefficient, mutual information is not limited to real-valued random variables, it is more general and determines the similarity of the product of the joint distribution and the marginal distribution of the decomposition. Therefore, the mutual information of two discrete random variables X and Y can be defined as:

[0031]

[0032] where p(x,y) is the joint probability distribution function of X and Y, and p(x) and p(y) represent the marginal probability distribution functions of X and Y, respectively.

[0033] Unbiased Estimate: a kind of unbiased inference when using sample statistics to estimate population parameters. The mathematical expectation of the estimator is equal to the true value of the estimated parameter, and the estimator is called the unbiased estimate of the estimated parameter, that is, it has unbiasedness. The significance of unbiased estimate is that its average value is close to the true value of the estimated parameter when repeated multiple times.

[0034] Example 1

[0035] With the rapid rise of the network, mobile cloud computing, large-scale MIMO (Multiple-Input Multiple-Output), millimeter wave communication and other new generation information technology are developing rapidly, at the same time, people's demand for mobile communication is no longer limited to voice conversation services, therefore, a single communication network can no longer support future business needs. For this reason, the 3rd Generation Partnership Project (3GPP) proposes a heterogeneous network, that is, by introducing various micro base stations to change the traditional single-layer network structure to build a multi-technology, multi-protocol network system. However, due to the limited coverage of a single network, as the user's location information changes, the quality of the communication network will inevitably change when switching between heterogeneous networks, that is, when the user is in the edge coverage position of the original network, the original network will not be able to guarantee the user's communication quality, resulting in real-time sharing of business data. Therefore, in the existing heterogeneous network environment, how to use the difference in network quality to achieve seamless switching of user networks is a key technical problem that needs to be solved in the development of heterogeneous networks.

[0036] The switching of terminals in a heterogeneous network is divided into horizontal switching and vertical switching according to the type of the network after switching, wherein: horizontal switching refers to the network before and after switching belonging to the same type of network, therefore, the reference quantity of horizontal switching is generally received signal strength (RSS); while vertical switching is a multi-access technology-oriented heterogeneous network environment, such as switching between wireless local area networks and cellular data (i.e. switching between users in different types of networks), which is also the focus of this patent.

[0037] In the vertical switching process, the network will log out of the session and re-register. During this period, the client-side network will be interrupted and accompanied by the ping-pong effect problem. Therefore, how to achieve accurate and seamless switching of multiple networks in a heterogeneous network environment has important practical significance for improving customer perception. In recent years, a large number of studies on multi-network switching have been conducted at home and abroad, which can be divided into threshold-based switching, multi-attribute decision-based switching, artificial intelligence-based switching, etc., wherein:

[0038] (1) The threshold-based network vertical switching method mainly uses RSS as the evaluation characteristic parameter, and only when RSS exceeds the threshold value will the network switch.

[0039] For example: the related patent "CN 117377004 A" provides a cell switching method based on signal strength. When the terminal signal strength is less than the first signal strength threshold and greater than the second signal strength threshold, the terminal service cell is switched by comparing the service transmission rate. The method provided by the patent has poor stability. If the terminal RSS measurement fluctuates around the threshold, the switching will not be timely, and the ping-pong effect will be easily triggered, causing the interruption of the user end network.

[0040] The related patent "CN 116744337 A" designs a method for adjusting network transmission parameters based on terminal signal strength, location information and package information. The terminal probe transmits terminal network information to the base station side, and the base station adjusts the network parameters based on the collected information. However, the method provided by the patent has great limitations. When the terminal moves quickly, the probe collected information will have errors, causing the base station side to be unable to accurately adjust the network resources, resulting in network resource loss.

[0041] The related patent "CN 111093241 A" proposes a network switching method based on double thresholds. The network switching is completed by comparing the preset signal strength threshold of wired or wireless network. The method provided by the patent only applies to wired and wireless network switching, and cannot be compatible with complex network scenarios.

[0042] Therefore, the network vertical switching method based on threshold has high feasibility and can be directly applied to network decision-making, and the decision-making structure is simple. However, this method does not consider network dynamic changes, and the application scene of this method has great limitations, which cannot meet the more complex heterogeneous network environment, and is prone to produce ping-pong effect in actual application process.

[0043] (2) Network vertical switching method based on artificial intelligence. This method is based on deep learning, reinforcement learning, neural network algorithm to realize the switching of heterogeneous network. After training, evaluation and optimization of historical data model, the network switching decision is made according to the final optimized model.

[0044] For example, the related patent "CN 116017601 A" designs a network automatic switching device based on deep learning technology. The deep convolutional neural network model collects high-dimensional features of different receiving powers on the terminal side, and infers the network status of the transmitting base station according to the signal strength of each terminal, so as to realize intelligent switching of terminal network. The method provided by the patent can accurately judge the network status change and complete the user side network switching in time, but the amount of training set data is large in the early stage, and the collection of receiving power of all terminals under the same base station is required, which has low feasibility.

[0045] The related patent "CN 113784410 B" discloses a heterogeneous wireless network switching method based on reinforcement learning TD3 algorithm. It first obtains the weight value of different network state parameters by using entropy weight method, synchronously constructs reward function, and stores the reward value obtained by switching decision under different network states in experience pool, and finally obtains the optimal switching strategy by using loop iteration. To some extent, the method provided by the patent can complete the switching of heterogeneous network, but the large demand for data model in the early experience pool will cause errors in the switching strategy.

[0046] Therefore, compared with the threshold switching method, the network vertical switching method based on artificial intelligence has higher switching decision accuracy, but the strategy involves more algorithms and the decision structure is more complex. The training set directly affects the accuracy of subsequent switching decisions, so the demand for training set in the early stage is large and needs to be further optimized and improved.

[0047] (3) The network vertical switching method based on multi-attribute decision making is a comprehensive switching decision made by measuring network multi-attribute indicators such as delay, RSS, packet loss rate, bandwidth, and bit error rate. This method has high network switching accuracy and high feasibility.

[0048] For example, the related patent "CN 117998512 A" provides a heterogeneous network vertical switching method based on multi-attribute decision making, which uses AHP algorithm to determine subjective weight and EWM algorithm to determine objective weight. By deeply integrating subjective weight, objective weight, positive indicators and negative indicators, the current network total weight is obtained, thereby completing the vertical switching of heterogeneous network. The method provided by the patent can effectively reduce the occurrence of ping-pong effect, but it ignores the subjective demand of customers for network, resulting in poor customer communication perception and uneven network load.

[0049] The related patent "CN 112839366 B" designs a cognitive heterogeneous wireless network vertical switching method, which sorts and processes the customer side wireless network by gray correlation analysis method, and completes the optimal selection of wireless network by the customer by means of two-part graph maximum matching network selection algorithm. The method provided by the patent can greatly improve the customer's wireless network perception, but in the process of rapid customer movement, it cannot guarantee the premise condition of RSS> RSS0, thereby causing the weight values of subsequent parameters to be unable to be normally sorted.

[0050] The related patent "CN 113115411 A" proposes a heterogeneous network switching based on coordination theory. First, the method uses the utility function to obtain the weight value of each attribute, and combines user preference and business scenario to realize the sorting of the performance of each wireless network through the TOPSIS and MEW methods, so as to realize the access of the best network. The method provided by the patent can effectively combine user preference to switch the heterogeneous network, and can satisfy the load balancing of the network, but the patent does not explicitly indicate the node of network switching, and does not mention how to avoid the occurrence of ping-pong effect.

[0051] Therefore, the method of network vertical switching based on multi-attribute decision can realize the switching of the optimal network, and can meet the customer side network preference and specific business scenario requirements. However, the network switching method based on multi-attribute does not indicate the network switching node, resulting in a long calculation time in the early stage of the switching process, and greatly increasing the occurrence of ping-pong effect.

[0052] In summary, although the existing strategy method can realize the switching of the heterogeneous network to a certain extent, there are still many deficiencies:

[0053] (1) The switching process includes a switching trigger stage and a switching decision stage. The trigger stage refers to the triggering of the switching operation when the network stalls during customer use; the switching decision stage refers to how the customer selects the optimal network for switching to ensure business stability during the network stall process. The former and the latter are related to each other. When the network stalls, if the switching trigger is not timely, it will affect the normal operation of the subsequent decision and form a ping-pong effect. Conversely, if the network does not meet the user's business requirements after the switching decision, the network switching will be continuously triggered, causing network congestion and affecting the customer's network perception. Therefore, enhancing the accuracy of the network switching node and avoiding the ping-pong effect are the first key problems to be solved by the embodiments of the present application.

[0054] (2) At present, the method of heterogeneous network switching mainly calculates the weight value of each attribute through various algorithms, such as simple weighting method, analytic hierarchy process, grey correlation method, and utility function, and realizes the switching of the optimal network according to the high and low of the weight value or utility value, but it is not universally applicable. The attributes that affect the network status of each scene do not have uniformity. If all network application scenarios use full attribute weight calculation, it will cause the calculation complexity of the entire switching process to be large, and the calculation period will be prolonged. Therefore, improving the matching of the switching decision and optimizing the calculation process are the core technical problems to be solved by the embodiments of the present application.

[0055] To solve the problem, the embodiment of the present application provides a method for network adaptive switching, which will be described in detail below. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order from here.

[0056] Figure 1 is a flowchart of an optional network adaptive switching method according to an embodiment of the present application, as shown in Figure 1 The method comprises steps S102-S108, wherein:

[0057] In step S102, the network abnormality features of the target network when the network abnormality occurs in a preset time period and the network parameter set of the target network for processing the target user service in the preset time period are obtained.

[0058] In the technical solution provided in step S102, the above-mentioned target network can be understood as the network in which the user terminal device is located in a preset time period, such as a wireless local area network (WLAN), a cellular network, a satellite communication network, etc. The network abnormality features of the target network when the network abnormality occurs in a preset time period can be understood as the performance characteristics of the target network when the performance of the target network is degraded or fails, for example, a higher packet loss rate, an increased delay, a reduced signal strength (RSS), etc. In addition, the network parameter set of the target network for processing the target user service in the preset time period refers to the set of all network parameters related to the target network processing the target user service of the target user terminal, and this set includes network parameter subsets at multiple time points in the preset time period, and each network parameter subset contains at least one network parameter of a network attribute, such as delay, RSS (Received Signal Strength), bandwidth, jitter, packet loss rate, etc., which are used to describe the specific status of the network.

[0059] In step S104, at least one target network attribute affecting the target user service is selected from the multiple network attributes in the network parameter set according to the network abnormality features.

[0060] In the technical solution provided in step S104, the control system can analyze the network parameter set using the network abnormality features to identify the key network attributes affecting the service performance of the target user service. That is, the control system identifies the key network attributes most sensitive to the network abnormality response from the multiple network attributes, and monitors these key network attributes in order to efficiently evaluate the network status and make corresponding decisions subsequently.

[0061] Step S106, for each time within the preset time period, the network parameters corresponding to each target network attribute are integrated to obtain the performance estimation value of the target network at the time.

[0062] In the technical solution provided in step S106, the control system obtains the comprehensive performance estimation value of the target network at the time by integrating the network parameters of the target network attribute at each time within the preset time period. Specifically, once the key network attributes affecting the target user service are determined, the control system can perform data fusion on the network parameters of these attributes at each sampling time to calculate an estimation value representing the overall network performance.

[0063] Step S108, network switching management is performed on the target user terminal according to the performance estimation value of the target network at each time.

[0064] In the technical solution provided in step S108, the control system dynamically evaluates the network condition of the target network by using the performance estimation value of the target network at each time within the preset time period, and accordingly executes the network switching strategy to optimize the network connection experience of the user terminal device. Specifically, once the performance estimation value of the network at each time within the preset time period is obtained, these performance estimation values will be used as the basis for system decision to determine whether heterogeneous network switching is needed.

[0065] Based on the solutions defined in steps S102 to S108, in the embodiment, the control system analyzes the network abnormality characteristics and intelligently selects the network parameter set when the target network abnormally occurs within the preset time period to identify and focus on the key network attributes affecting the target user service; the network parameters at each time of the key network attributes are integrated to obtain an accurate comprehensive performance estimation value generated dynamically, and the network switching decision is determined based on the comprehensive performance estimation value of the target network at different times, which achieves the purpose of improving user experience and ensuring service continuity, thereby solving the technical problem that the related art does not consider the non-uniformity of network attributes in different scenarios when determining the weight of network attributes, resulting in inaccurate network switching nodes.

[0066] The steps of the method will be described below in conjunction with the architecture diagram and specific implementation process of the network adaptive switching system shown in Figure 2

[0067] ​As an optional implementation, when the user terminal device enters the initial moment of the wireless heterogeneous network environment, the user terminal device will select the initial network access point according to the received signal strength. When the user terminal device is in the working process (i.e., performing user services), the location point information of the user terminal device will change in real time, and because the coverage range, signal interference, bandwidth, etc. of different networks are not the same, the network stability of the user terminal device at the edge of the network is low, and the ping-pong effect is easy to occur. Therefore, in the embodiment of the present application, the loopback delay, RSS, bandwidth, jitter, packet loss rate, etc. are selected as the feature collection data.

[0068] Among them:

[0069] The loopback delay refers to the time required for the sending end to send a message to the receiving end and receive an acknowledgment instruction from the sending end. Generally, high loopback delay is mainly caused by data retransmission, data packet loss, etc. due to poor network communication quality. Therefore, the loopback delay is an important attribute for measuring network quality.

[0070] RSS reflects the signal strength of the network. In the process of space transmission, wireless networks are easily affected by other external factors, such as obstructions, weather conditions, and other wireless device interference sources. When the signal-to-noise ratio reaches a threshold, it will cause the interruption of service communication. Therefore, RSS is a basic attribute that reflects network quality information.

[0071] Bandwidth generally refers to the amount of data transmitted in a certain unit of time, and its unit is bit rate. Generally, when the bandwidth is low, the customer network delay increases, the transmission rate decreases, and the customer's network experience is greatly affected. Therefore, bandwidth is an important indicator for evaluating network performance and network speed.

[0072] Jitter reflects the stability of real-time transmission of the network. When the network is congested, the transmission delay of the same group of messages changes, affecting the real-time network communication of the customer. The smaller the network jitter, the more stable the network state.

[0073] Packet loss rate refers to the difference between the received message at the receiving end and the sent message at the sending end. When the user terminal device moves, there is a possibility of packet loss in the channel. Therefore, the packet loss rate can also be one of the evaluation criteria for network switching.

[0074] In addition, different types of services have different requirements for network attributes, so the user service type in the embodiment of the present application can be divided into the following categories,

[0075] (1) Low latency communication: as a high-speed communication method, it reduces the data transmission and processing delay to the minimum, ensuring the real-time response needs of people, and the application scenarios include remote surgery, autonomous driving, real-time translation, online games, etc.

[0076] (2) Wide connection communication: In practical application process, same local area network under part terminal rate demand is lower, time delay is not sensitive, etc., but access terminal quantity is more, signaling interaction proportion is big, etc. Its typical application scene includes wisdom city, wisdom building, smart home, environment monitoring, etc.

[0077] (3) Large bandwidth communication: Large bandwidth can realize large-scale data transmission, high-speed network connection and high data throughput, on the basis of improving the speed of data transmission, guaranteeing the stable connection performance of network peak period, its application scene mainly includes video monitoring, financial transaction, medical image, etc.

[0078] Further, since all network parameters of the target network in processing the target user business in the preset time period are real-time collected. Therefore, once the target network fluctuates, then the dynamic data collection will have problems such as interruption, abnormal collection, etc., thereby greatly reducing the authenticity of the data. In addition, due to the change of location information or environmental factors, it will also cause a large number of data to breed, seriously affecting the accuracy of the later calculation. Therefore, the embodiment of the present application needs to fill in the missing values of the network parameter subset collected in the preset time period. The control system can use linear interpolation method to complete the data, that is, the missing data is completed by the average value of the data before and after the missing point, that is:

[0079]

[0080] Wherein, x n represents the network parameter of a certain network attribute at the current time, x n-1 represents the network parameter of the network attribute at the previous time, x n+1 represents the network parameter of the network attribute at the next time.

[0081] In addition, considering that the units of network parameters collected at different times may not be consistent, the control system can also normalize the network parameters, that is:

[0082]

[0083] Wherein, x i ′ represents the normalized parameter of a certain network attribute, x i ′ ∈[-1,1], x i represents the original network parameter of the network attribute at any time, x max represents the maximum original network parameter of the network attribute, x min represents the minimum original network parameter of the network attribute.

[0084] As an optional implementation, in the technical solution provided in the step S104, the control system can complete the screening of the key network attributes in the following manner, comprising:

[0085] Step S1041, calculating the first mutual information between each network attribute in the network parameter set and the network anomaly feature and the second mutual information between each network attribute corresponding to the network parameter set.

[0086] Step S1042, using the mRMR algorithm to analyze the first mutual information between each network attribute and the network anomaly feature and the second mutual information between each network attribute, so as to screen the multiple key network attributes affecting the target user service from the multiple network attributes in the network parameter set.

[0087] Step S1043, constructing a target function with the optimization target of maximizing the first mutual information between the key network attributes and the network anomaly feature and minimizing the second mutual information between each key network attribute, and using a preset optimization algorithm to solve the target function, so as to screen at least one target network attribute from the multiple key network attributes.

[0088] Specifically, assuming that the network attribute parameter is N, the dimension of the corresponding network attribute is m, that is, each network attribute in the network parameter set can be denoted as X={x i ,x j ,i,j=1,2,3…,m};In addition, the network service fault type is denoted as c, and a p-dimensional network attribute is selected to construct an attribute set S, wherein p≤m. Therefore, the first mutual information between each network attribute and the network anomaly feature can be denoted as I(x i ;c), and the second mutual information between each network attribute can be denoted as I(x i ;x j ).

[0089] For the mRMR algorithm, when the correlation between the p-dimensional network attributes is the lowest, the correlation between the network anomaly in different service scenarios and the network attribute can be optimal, that is:

[0090]

[0091] Then, the above formula is integrated, and the final target is defined as finding the attribute set S with the maximum correlation and the minimum redundancy. Specifically, the operators D and R can be combined as follows:

[0092]

[0093] ​​On the basis of the above standard function, the key network attribute (i.e. the approximate optimal solution) causing the network anomaly can be obtained by using the incremental search optimization algorithm. It is assumed that a subset S of m-1 network attributes has been obtained m-1 , and m network attributes need to be selected from the remaining X-S m-1 subset of network attributes, so the maximum can be used for feature selection, i.e. maximizing:

[0094]

[0095] By calculating the above formula, the p key network attributes affecting the target user service can be preliminarily screened from the multiple network attributes in the network parameter set.

[0096] It should be noted that, in order to improve the correlation between different network fault features and network attributes, the fusion of multiple weight algorithms can also be used to complete the importance sorting of feature quantities in different network application scenarios, as follows:

[0097] Table 1

[0098] Evaluation feature algorithm Attribute feature T test Bandwidth Chi-square test Latency, jitter Relief algorithm RSS, packet loss rate

[0099] Further, in order to measure the target network attribute with the greatest impact, the control system can also set the indication vector of the p key network attributes as y = [y1, y2,..., y p ], and combine the first mutual information between the key network attributes and the network anomaly features and the second mutual information between the key network attributes to build a target function, i.e.:

[0100]

[0101] In the formula, y = [y1, y2,..., y p ] represents the indication values of the multiple key network attributes, and the indication value y i is a binary value, used to reflect whether the i-th key network attribute is selected as the target network feature, wherein if y i = 1, it means that the i-th key network attribute is selected as the target network feature, and if y i = 0, it means that the i-th key network attribute is not selected as the target network feature; d represents the number of target network attributes; D ′ represents the correlation matrix between the multiple key network attributes and the network anomaly features; and R ′ represents the redundancy matrix between the multiple key network attributes. Then, the above represents the average correlation between the selected network attributes and the network anomaly features, and y T R ′y represents the sum of the redundancy between all selected network attributes, so y represents the average redundancy between the selected network attributes.

[0102] Therefore, by solving the objective function y, the multiple target network attributes with the highest correlation with the network anomaly characteristics and the lowest redundancy can be obtained. For example, Table 2 below is a table of the relationship between the network attributes and the user service type using the above method.

[0103] Table 2

[0104] Network attribute Service type Loopback latency Low latency, large bandwidth RSS Wide connection, large bandwidth, low latency Bandwidth Large bandwidth, wide connection, low latency Jitter Low latency, large bandwidth, wide connection Packet loss rate Low latency, large bandwidth, wide connection

[0105] After obtaining the target network attributes that affect the target user service through the above steps, the control system can perform data fusion on the network parameters of these network attributes at each sampling time.

[0106] As an optional implementation, in the technical solution provided in step S106, the control system can implement data fusion according to the following steps, including:

[0107] Step S1061, for each time in the preset time period, determine the zero-mean score noise of each target network attribute at the time, and determine the weight of each target network attribute according to the unbiased estimate of the variance of the zero-mean score noise of each target network attribute at the time.

[0108] Specifically, the technical solution provided in step S1061 can be implemented by the following steps:

[0109] First, determine the unbiased estimate of the variance of the zero-mean score noise of each target network attribute at the time, that is, V k (t), k = 1, 2, …, d, then the variance of the zero-mean score noise of each target network attribute at each t time can be denoted as And the unbiased estimate of the variance of the zero-mean score noise of each target network attribute at each t time can be denoted as

[0110] Then, according to the unbiased estimate of the variance of the zero-mean score noise of each target network attribute at the time, the weight of each target network attribute is calculated by the following formula:

[0111]

[0112] In the formula, y represents the weight of the kth target network attribute at t time. represents an unbiased estimator of the variance of the zero-mean score noise of the kth target network attribute at time t.

[0113] Step S1062, the sum of the zero-mean score noise of each target network attribute at time t and the network parameter is calculated, and the sum of the product of each target network attribute at time t and the weight of the target network attribute is taken as the performance estimation value of the target network at time t. Therefore, the expression of the performance estimation value of the target network at time t can be written as:

[0114]

[0115] wherein the above X' k (t) = X k (t) + V k (t), and X k (t) represents the network parameter of the kth target network attribute at time t, and Z(t) represents the performance estimation value of the target network at time t. These estimation values reflect the processing capacity and health status of the target network in processing the target user service at each time.

[0116] Further, after obtaining the performance estimation value of the target network at each time through the above steps, the control system can manage the network switching of the user mobile terminal according to the following method, which comprises: counting the number of time points at which the performance estimation value of the target network in a preset time period is lower than a preset threshold value; and in the case that the number of time points exceeds a preset number threshold, vertically switching the target user terminal from the target network to other heterogeneous networks.

[0117] That is, the control system compares the network performance estimation value of the target network with a preset threshold value through dynamic monitoring, and then counts the number of performance decline moments as the basis for triggering network switching. Therefore, the system determines whether the network performance decline reaches the critical point of switching by setting a number threshold in advance. If the number of performance decline moments exceeds the number threshold, it indicates that the unstable state of network performance has lasted to a certain extent, which significantly affects the service continuity and user experience of the target user service, thereby triggering the network switching process. Once the switching condition is triggered, the system will perform vertical switching operation to switch the user mobile terminal from the current network to another heterogeneous network. Through the above network switching management strategy, the control system can respond and effectively manage the network performance decline in a timely manner, ensuring that the user terminal can automatically switch to a more stable network when the performance is lower than expected, thereby maintaining the continuity of the service and optimizing the user experience. This strategy avoids the continuous negative impact of network performance fluctuations on user services, and by setting reasonable performance thresholds and number thresholds, it balances the necessity of switching and the strategy of avoiding frequent switching, ensuring efficient use of network resources and the switching process.

[0118] In addition, the above-mentioned network adaptive switching method provided by the embodiments of the present application can effectively solve the problems of low switching accuracy, poor customer network perception, and complex calculation in the prior art, which are embodied in the following two points:

[0119] (1) Enhance the accuracy of network switching nodes and suppress the ping-pong effect.

[0120] Compared with the traditional attribute weight value comparison, the embodiments of the present application use the mRMR algorithm to extract the target network attributes, which can actually match the user service type on the one hand, and determine the target network attributes that affect the network under the service type on the other hand, thereby reducing the complexity of calculation and improving the accuracy of network switching. At the same time, the embodiments of the present application use the interval measurement method to determine whether to perform network switching by comparing the probability that the optimal estimation value is greater than the preset threshold within the preset collection period, thereby significantly reducing the occurrence of the ping-pong effect.

[0121] (2) High switching decision matching and high user experience quality.

[0122] Firstly, the embodiments of the present application integrate the parameters of the selected target network attributes to construct a data model associated with the target user service type, fully considering the service type requirements of users to improve the feasibility of practical application. Secondly, on the basis of the mRMR algorithm, the selected network attributes are sorted in importance by using T-test, chi-square test, and Relief algorithm, etc., which greatly improves the robustness of the algorithm to quickly and accurately extract the target network attributes that affect the target user service.

[0123] Embodiment 2

[0124] Based on the embodiment 1 of the present application, an embodiment of a network adaptive switching device is further provided, which executes the network adaptive switching method of the above embodiment when running. Wherein, Figure 3 is a structural schematic diagram of an optional network adaptive switching device according to the embodiment of the present application, as Figure 3 shown, the network adaptive switching device at least includes an acquisition module 32, a determination module 34, an evaluation module 36 and a management module 38, wherein:

[0125] The acquisition module 32 is configured to acquire network abnormality features of the target network when network abnormality occurs within a preset time period and a network parameter set of the target network for processing target user services of the target user terminal within the preset time period, wherein the network parameter set includes network parameter subsets at multiple time points within the preset time period, and the network parameter subset includes network parameters under at least one network attribute.

[0126] The determination module 34 is configured to filter at least one target network attribute affecting the target user service from the multiple network attributes in the network parameter set according to the network abnormality features.

[0127] The evaluation module 36 is configured to integrate the network parameters of each target network attribute at each time point within the preset time period to obtain a performance estimation value of the target network at each time point.

[0128] The management module 38 is configured to perform network switching management on the target user terminal according to the performance estimation value of the target network at each time point.

[0129] It should be noted that each module in the above network adaptive switching device can be a program module (for example, a set of program instructions for implementing a certain specific function) or a hardware module. For the latter, it can be in the following form, but not limited to: the form of each module is a processor, or the functions of each module are implemented by a processor.

[0130] Embodiment 3

[0131] According to the embodiment of the present application, a non-volatile storage medium is further provided, which stores a program, wherein when the program is running, the non-volatile storage medium controls the device where the non-volatile storage medium is located to execute the network adaptive switching method in the embodiment 1.

[0132] According to the embodiment of the present application, a computer program product is further provided, which includes a stored computer program, wherein the computer program is executed by a processor to implement the network adaptive switching method in the embodiment 1.

[0133] According to an embodiment of the present application, a processor is also provided, which is configured to execute a program. When the program is executed, the network adaptive handover method in the above embodiment 1 is implemented.

[0134] According to an embodiment of the present application, an electronic device is also provided, which comprises one or more processors, and a memory configured to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors are configured to implement a program configured to execute the network adaptive handover method in the above embodiment 1.

[0135] Optionally, when the computer program is executed, the following steps are implemented: obtaining a network abnormality feature of a target network when a network abnormality occurs in a preset time period and a network parameter set of the target network for processing a target user service of a target user terminal in the preset time period, wherein the network parameter set comprises network parameter subsets at multiple time points in the preset time period, and each network parameter subset comprises network parameters at at least one network attribute; filtering at least one target network attribute affecting the target user service from the multiple network attributes in the network parameter set according to the network abnormality feature; integrating the network parameters of each target network attribute at each time point in the preset time period to obtain a performance estimation value of the target network at each time point; and performing network switching management on the target user terminal according to the performance estimation value of the target network at each time point.

[0136] As an optional implementation, the electronic device can exist in the form of a mobile terminal, a computer terminal or a similar computing device. Figure 4 A hardware structure block diagram of an electronic device for implementing the network adaptive handover method is shown. As shown in the figure, Figure 4 The electronic device 40 can comprise one or more processors 402 (the processor 402 can comprise, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 404 for storing data, and a transmission device 406 for communication functions. In addition, it can also include a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which can be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. Those skilled in the art can understand that, Figure 4 The structure shown in the figure is only a schematic, which does not limit the structure of the above-mentioned electronic device. For example, the electronic device 40 can also include more or fewer components than those shown in the figure, or have a different configuration from that shown in the figure. Figure 4 Figure 4 ​​

[0137] It should be noted that the one or more processors 402 and / or other data processing circuitry described above can be referred to herein generically as "data processing circuitry". The data processing circuitry can be embodied in whole or in part as software, hardware, firmware, or any combination thereof. Furthermore, the data processing circuitry can be a single standalone processing module, or it can be incorporated in whole or in part within any one of the other elements of the electronic device 40. As referred to in the embodiments herein, the data processing circuitry acts as a processor to control, for example, the selection of the variable resistance terminal path connected to the interface.

[0138] The memory 404 can be used to store software programs of application software and modules, such as program instructions / data storage means corresponding to the network adaptive handover method in the embodiments of the present application. The processor 402 can execute various functional applications and data processing by running the software programs and modules stored in the memory 404, i.e. implement the vulnerability detection method of the application program described above. The memory 404 can include a high-speed random access memory, and can also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some examples, the memory 404 can further include a memory remotely arranged with respect to the processor 402, which can be connected to the electronic device 40 through a network. Examples of the network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0139] The transmission device 406 is used to receive or send data via a network. The specific examples of the network can include a wireless network provided by a communication provider of the electronic device 40. In one example, the transmission device 406 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 406 can be a radio frequency (Radio Frequency, RF) module, which is used to communicate with the Internet in a wireless manner.

[0140] The display can be, for example, a touch screen type liquid crystal display (LCD), which can enable a user to interact with the user interface of the electronic device 40.

[0141] The serial numbers of the embodiments of the present application described above are only for description, and do not represent the advantages and disadvantages of the embodiments.

[0142] In the above-described embodiments of the present application, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0143] In several embodiments provided in the present application, it should be understood that the disclosed technology can be implemented by other ways. Among them, the above-described device embodiments are only schematic, for example, the division of units can be a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed units can be indirect coupling or communication connection through some interfaces, units or modules, which can be electrical or other forms.

[0144] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e. they can be located in one place or distributed to multiple units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0145] In addition, the functional units in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0146] If the integrated unit is realized in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the essential part or all or part of the related technology that make contributions can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the embodiments of the present application. The aforementioned storage medium includes: a U disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0147] The above is only the preferred embodiment of the present application, and it should be pointed out that for ordinary skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, which should be considered as the protection scope of the present application.

Claims

1. A network adaptive handover method, characterized by, include: The network anomaly characteristics of the target network when a network anomaly occurs within a preset time period and the network parameter set of the target network for processing target user services of target user terminals within the preset time period are obtained. The network parameter set includes: a subset of network parameters at multiple times within the preset time period, and the subset of network parameters includes network parameters under at least one network attribute. The method for selecting at least one target network attribute affecting the target user's service from multiple network attributes within the network parameter set based on the network anomaly characteristics includes: calculating a first mutual information between each network attribute in the network parameter set and the network anomaly characteristics, and a second mutual information between each network attribute in the network parameter set; analyzing the first mutual information between each network attribute and the network anomaly characteristics and the second mutual information between each network attribute using the maximum correlation minimum redundancy (mRMR) algorithm to select multiple key network attributes affecting the target user's service from multiple network attributes within the network parameter set; constructing an objective function with the optimization objective of maximizing the first mutual information between the key network attributes and the network anomaly characteristics and minimizing the second mutual information between each key network attribute, and solving the objective function using a preset optimization algorithm to select at least one target network attribute from the multiple key network attributes; The network parameters of each target network attribute at various times within a preset time period are integrated to obtain the performance estimate of the target network at each time point. This includes: for each time point within the preset time period, determining the zero-mean score noise of each target network attribute at that time point, and determining the weight of each target network attribute based on the zero-mean score noise of each target network attribute at that time point; calculating the sum of the zero-mean score noise of each target network attribute at that time point and the network parameters, and using the sum of the products of the sum of the zero-mean score noise of each target network attribute at that time point and the weight of the target network attribute as the performance estimate of the target network at that time point. Network switching management is performed on the target user terminal based on the performance estimates of the target network at various times.

2. The method according to claim 1, characterized in that, The service types of the target user services include at least: low-latency communication, wide-connectivity communication, or high-bandwidth communication; The types of network attributes include at least one of the following: loopback delay, network signal strength, network bandwidth, and packet loss rate.

3. The method according to claim 1, characterized in that, The objective function is: ; In the formula Indicates the indicator values ​​of the aforementioned key network attributes, and The indicated value This is used to reflect whether the i-th key network attribute is selected as the target network attribute; d represents the number of target network attributes; A correlation matrix representing the multiple key network attributes and the network anomaly features; represents the redundancy matrix among the multiple key network attributes; p represents the number of target network attributes that affect the target user's service selected from the multiple network attributes.

4. The method according to claim 1, characterized in that, Determining the weights of each target network attribute based on the zero-mean score noise of each target network attribute at the given time includes: Determine an unbiased estimator of the variance of the zero-mean score noise for each of the target network attributes at the stated time. Based on the unbiased estimator of the variance of the zero-mean score noise of each of the target network attributes at the stated time, the weights of each of the target network attributes are calculated using the following formula: ; In the formula This represents the weight of the k-th target network attribute at time t. Let d represent the unbiased estimator of the variance of the zero-mean score noise of the k-th target network attribute at time t, and let d represent the number of the target network attributes.

5. The method according to claim 1, characterized in that, Network handover management of the target user terminal is performed based on the performance estimates of the target network at various times, including: The number of times the estimated performance value of the target network is lower than a preset threshold value within a preset time period is counted. If the number of times exceeds a preset threshold, the target user terminal will be vertically switched from the target network to another heterogeneous network.

6. A network adaptive handover device, characterized in that, include: The acquisition module is used to acquire network anomaly characteristics of the target network when a network anomaly occurs within a preset time period and the network parameter set of the target network processing target user services of the target user terminal within the preset time period. The network parameter set includes: a subset of network parameters at multiple times within the preset time period, and the subset of network parameters includes network parameters under at least one network attribute. The determination module is used to filter at least one target network attribute affecting the target user's service from multiple network attributes in the network parameter set based on the network anomaly characteristics. This includes: calculating a first mutual information between each network attribute in the network parameter set and the network anomaly characteristics, and a second mutual information between each network attribute in the network parameter set; analyzing the first mutual information between each network attribute and the network anomaly characteristics and the second mutual information between each network attribute using the maximum correlation minimum redundancy (mRMR) algorithm, to filter out multiple key network attributes affecting the target user's service from multiple network attributes in the network parameter set; constructing an objective function with the optimization objective of maximizing the first mutual information between the key network attributes and the network anomaly characteristics and minimizing the second mutual information between each key network attribute, and solving the objective function using a preset optimization algorithm, to filter out at least one target network attribute from the multiple key network attributes. The evaluation module is used to integrate the network parameters of each target network attribute at various times within a preset time period to obtain the performance estimate of the target network at each time. This includes: for each time within the preset time period, determining the zero-mean score noise of each target network attribute at that time, and determining the weight of each target network attribute based on the zero-mean score noise; calculating the sum of the zero-mean score noise and network parameters of each target network attribute at that time, and using the sum of the products of the sum of the zero-mean score noise and the weights of the target network attributes as the performance estimate of the target network at that time. The management module is used to manage network switching for the target user terminal based on the performance estimates of the target network at various times.

7. A non-volatile storage medium, characterized in that, The non-volatile storage medium stores a computer program, wherein the device containing the non-volatile storage medium executes the network adaptive handover method according to any one of claims 1 to 5 by running the computer program.

8. A network device, characterized in that, include: A memory and a processor, the processor being configured to run a program stored in the memory, wherein the program, when running, executes the network adaptive switching method according to any one of claims 1 to 5.

Images