A real-time encryption method for 5G networks based on artificial intelligence

Abstract

The present invention discloses a 5G network real-time encryption technology based on artificial intelligence, which relates to the field of information security and encryption technology. It includes: adjusting the initial encryption parameters according to the security threats and network environment to obtain adaptive encryption parameters; verifying users and devices through blockchain technology, and managing the generation, distribution and update of quantum keys through smart contracts, and updating quantum keys through distributed nodes; determining the final encryption protocol and algorithm through the encryption parameters and the updated quantum keys. The present invention uses an AI algorithm to analyze network traffic patterns in real time, accurately identify potential security threats, and adjust encryption parameters in a timely manner. At the same time, it manages the generation, distribution and update of quantum keys through blockchain technology and smart contracts to ensure the security and non-tamperability of quantum keys, and enhance the security of data transmission.

Description

Technical Field

[0001] The present invention relates to the field of information security and encryption technology, and specifically to a 5G network real-time encryption technology based on artificial intelligence. Background Art

[0002] With the rapid deployment of the fifth generation of mobile communication (5G) technology, it not only brings unprecedented high speed, low latency and high reliability, but also provides possibilities for emerging applications such as smart cities, self-driving cars, industrial Internet of Things, etc. However, these advances are also accompanied by new security challenges. In 5G networks, the security of data transmission is crucial because it is directly related to user privacy protection, service continuity and the overall security of the system.

[0003] Traditional encryption technologies such as symmetric encryption (such as AES-256), asymmetric encryption (such as RSA-2048) and hybrid encryption systems have played an important role in 4G and earlier network environments. However, facing the large-scale connection requirements, extremely high throughput and ultra-low latency characteristics of 5G networks, traditional encryption methods are unable to cope with them. They often require fixed key management and long computing time, which is unacceptable in the dynamically changing 5G network environment.

[0004] The Chinese invention patent with announcement number CN114071461A discloses a 5G communication module based on quantum key encryption, including a 5G module, a secure storage module and a quantum key management and control module. The 5G module is provided with a secure storage module and a quantum key management and control module, and the secure storage module communicates with the quantum key management and control module to complete the identity authentication and quantum key distribution of the 5G module using a quantum certificate, and use the distributed quantum key to complete the encrypted transmission of 5G service data. This invention adds a secure storage module and a quantum key management and control module to the 5G module, and uses quantum keys to solve the identity authentication and transmission encryption problems of the 5G module device, which can prevent the risk of key leakage, reduce the cost of key maintenance, and resist attacks from quantum computing and quantum algorithms, and can provide secure wireless network communications for the Internet of Things using 5G.

[0005] The above and similar 5G communication technologies mainly rely on specific quantum certificate issuing authorities in actual operation. Although this implementation can combine quantum secure communication technology with 5G communication technology, it can ensure the security of data transmission while enjoying the high bandwidth and low latency of 5G. However, this implementation method has the risk of single point failure. Once the quantum certificate issuing authority is attacked, the other devices that rely on the quantum certificate will not be able to operate normally. Summary of the invention

[0006] The purpose of the present invention is to provide a 5G network real-time encryption technology based on artificial intelligence to solve the problems raised in the above background technology.

[0007] To achieve the above purpose, the present invention provides the following technical solution: a 5G network real-time encryption technology based on artificial intelligence, comprising:

[0008] Adjust encryption parameters: Analyze network traffic patterns through AI algorithms to obtain initial encryption parameters. At the same time, identify security threats by identifying real-time network traffic patterns. Adjust the initial encryption parameters based on the security threats and network environment to obtain adaptive encryption parameters.

[0009] Update quantum keys: Use blockchain technology to verify users and devices, and use smart contracts to manage the generation, distribution, and update of quantum keys. At the same time, quantum keys are updated through distributed nodes.

[0010] Determining encryption strength: determining the final encryption protocol and algorithm through the encryption parameters and the updated quantum key;

[0011] Update the quantum key as follows:

[0012] Verify users and devices through the blockchain technology, and build a quantum key management system based on the verification results and smart contracts;

[0013] Obtaining a communication delay change rate between nodes in the quantum key management system, and comparing the communication delay change rate with a change rate threshold, when the communication delay change rate is greater than the change rate threshold, the node corresponding to the communication delay change rate is incorrect, and the next step is executed; otherwise, the node corresponding to the communication delay change rate is correct, and the update frequency of the quantum key remains unchanged;

[0014] The time series data of network traffic is obtained, and a prediction model is constructed to obtain the probability of a security incident. At the same time, the probability of a security incident is compared with a risk threshold. When the probability of a security incident is greater than the risk threshold, the update frequency of the quantum key is increased. Otherwise, the update frequency of the quantum key remains unchanged.

[0015] Furthermore, adaptive encryption parameters are obtained as follows:

[0016] Acquire network traffic data, and preprocess the network traffic data to acquire feature data;

[0017] A traffic identification model is constructed through the characteristic data, and initial encryption parameters are obtained through the traffic identification model;

[0018] By using the traffic identification model, an anomaly score corresponding to the real-time network traffic is obtained, and the anomaly score is compared with a preset threshold, and a security threat is determined based on the comparison result;

[0019] According to the security threat and network environment, the initial encryption parameters are adjusted to obtain adaptive encryption parameters.

[0020] Furthermore, based on the comparison results, security threats are identified, specifically:

[0021] When the anomaly score is greater than a preset threshold, the real-time network traffic corresponding to the anomaly score has a security threat, otherwise, the real-time network traffic corresponding to the anomaly score has no security threat.

[0022] Going further, the method of obtaining anomaly scores is as follows:

[0023] M1: Reconstruct the real-time network traffic through the autoencoder to obtain the reconstructed network traffic;

[0024] M2: According to the reconstructed network traffic and the real-time network traffic, a reconstruction error corresponding to the real-time network traffic is obtained. The reconstruction error is the anomaly score corresponding to the real-time network traffic. The formula for obtaining the anomaly score is specifically:

[0025] ,

[0026] in: For the abnormality score, is the total amount of real-time network traffic, is the i-th real-time network traffic, is the reconstructed network traffic corresponding to the i-th real-time network traffic, An index of real-time network traffic.

[0027] Furthermore, the method for obtaining the preset threshold is as follows:

[0028] N1: According to the network traffic data, the true positive rate and false positive rate of the network traffic data under different thresholds are obtained, specifically:

[0029] ,

[0030] in: is the true positive rate, is the false positive rate, The number of network traffic data with security threats correctly identified. The amount of network traffic data that has not been identified as a security threat. The amount of network traffic data that was falsely identified as a security threat. The number of network traffic data correctly identified as not containing security threats;

[0031] N2: According to the true positive rate and the false positive rate, obtain the gap index, specifically:

[0032] ,

[0033] in: is the gap index, is the true positive rate, is the false positive rate;

[0034] N3: Determine the maximum value gap index according to the gap index corresponding to each threshold value, and the threshold value corresponding to the maximum value gap index is the preset threshold value.

[0035] Furthermore, the initial encryption parameters are adjusted as follows:

[0036] W1: According to the anomaly score corresponding to each real-time network flow, the average value of all the anomaly scores is obtained to determine the security threat intensity;

[0037] W2: Determine the comprehensive network performance indicators based on the network environment, specifically:

[0038] ,

[0039] in: It is a comprehensive indicator of network performance. is the weight corresponding to the bandwidth, is the bandwidth corresponding to the real-time network environment, is the weight corresponding to the delay, The delay corresponding to the real-time network environment;

[0040] W3: Obtain adjustment parameters corresponding to the initial encryption parameters through the security threat intensity and network performance comprehensive index, specifically:

[0041] ,

[0042] in: To adjust the parameters, is the coefficient of security threat intensity, is the security threat intensity, is the coefficient of the comprehensive index of network performance, It is a comprehensive indicator of network performance;

[0043] W4: According to the adjustment parameter, the initial encryption parameter is adjusted, specifically:

[0044] ,

[0045] in: is the adaptive encryption parameter, To adjust the parameters, is the initial encryption parameter.

[0046] Going further, we build a prediction model to obtain the probability of a security incident, specifically:

[0047] ,

[0048] in: for The probability of a security incident occurring at time t, is the characteristic data, for The transpose of is the bias term, is the weight vector.

[0049] Going further, the risk thresholds are:

[0050] ,

[0051] in: is the risk threshold, is the average probability distribution, is the coefficient, is the standard deviation of the probability distribution.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] First, the present invention uses AI algorithms to analyze network traffic patterns in real time, so as to accurately identify potential security threats and adjust encryption parameters in a timely manner. At the same time, blockchain technology and smart contracts are used to manage the generation, distribution and update of quantum keys, which further ensures the security and non-tamperability of quantum keys, thereby enhancing the security of data transmission.

[0054] Second: The present invention dynamically adjusts the encryption strength according to changes in the network environment, data sensitivity and predicted security risks, thereby reducing encryption complexity to reduce latency under high load conditions and enhancing encryption strength when processing highly sensitive data, thereby ensuring data security while optimizing performance;

[0055] Third: The present invention verifies the identity of users and devices through blockchain technology, and automatically manages the generation, distribution and update process of quantum keys through smart contracts, which not only improves the efficiency of key management, but also reduces the risk of single point failure in the key management system;

[0056] Fourthly, the present invention predicts and evaluates the probability of future security incidents by constructing a prediction model, so that relevant measures can be taken in advance according to the prediction results, so that it can more proactively defend against potential security threats instead of relying on post-event reactions. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a schematic diagram of the process of the 5G network real-time encryption technology of the present invention;

[0058] Figure 2 A schematic diagram of a process for obtaining adaptive encryption parameters according to the present invention;

[0059] Figure 3 A schematic diagram of a process for obtaining security threats according to the present invention;

[0060] Figure 4 A schematic diagram of a process of obtaining adaptive encryption parameters through anomaly scoring according to the present invention;

[0061] Figure 5 A schematic diagram of a process for determining the quantum key update frequency according to the present invention. DETAILED DESCRIPTION

[0062] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] The existing 5G communication technology mainly relies on a specific quantum certificate authority in actual operation. Although it can combine quantum secure communication technology with 5G communication technology, there is a risk of single point failure in the specific implementation process. Once the quantum certificate authority is attacked, the remaining devices that rely on the quantum certificate will not be able to operate normally. The technical solution of this application uses AI to analyze network traffic to identify potential security threats and adjust the initial encryption parameters accordingly. At the same time, blockchain technology and smart contracts are used to verify the identity of users and devices, ensure the security and non-tamperability of quantum key distribution, and adjust the update frequency of quantum keys according to the probability of security events calculated by the prediction model to determine the final encryption protocol and algorithm, thereby achieving security protection for data transmission in the 5G network environment.

[0064] refer to Figure 1-Figure 5 This embodiment provides a 5G network real-time encryption technology based on artificial intelligence, and the 5G network real-time encryption technology includes the following steps:

[0065] Step S1: Adjust encryption parameters. That is, the network traffic pattern is analyzed by the AI ​​algorithm to obtain the initial encryption parameters. At the same time, the security threats are obtained from the historical data and real-time network traffic patterns, and the initial encryption parameters are adjusted according to the security threats and different network environments to obtain adaptive encryption parameters. The details are as follows:

[0066] Step S1.1: Obtain network traffic data from the 5G network, perform corresponding preprocessing on the obtained network traffic data, and obtain feature data therefrom. In this embodiment, network traffic data is obtained from the 5G network, including but not limited to source IP address, destination IP address, port number, protocol type, packet size, timestamp and other information. At the same time, the obtained network traffic data is preprocessed accordingly, that is, data cleaning, standardization and normalization are performed. And feature data, namely statistical features, behavioral features and timing features, are extracted from the preprocessed network traffic data. Specifically, the statistical features are the average packet size and the amount of data transmitted per second, the behavioral features are the traffic patterns of specific applications or services, and the timing features are the time periods of the day.

[0067] Step S1.2: Construct a traffic identification model through machine learning (such as decision tree, random forest, support vector machine (SVM), K-nearest neighbor (KNN)) or deep learning algorithm (such as convolutional neural network (CNN), recurrent neural network (RNN), long short-term memory network (LSTM), autoencoder). And obtain the initial encryption parameters through the constructed traffic identification model.

[0068] In this embodiment, a traffic recognition model is constructed through a convolutional neural network (CNN) and feature data (such as statistical features, behavioral features, and timing features) extracted in step S1.1, and traffic patterns (i.e., normal traffic patterns and abnormal traffic patterns) are recognized through the traffic recognition model.

[0069] Furthermore, a set of initial encryption parameters can be determined through the output results of the constructed traffic identification model. It is worth noting that the initial encryption parameters can be determined based on the input characteristics. For example, different encryption algorithms (such as AES, RSA), key lengths, and initialization vectors (such as IV) can be selected according to different traffic patterns. That is, the initial encryption parameters corresponding to the traffic pattern are the encryption algorithm, key length, and initialization vector.

[0070] Step S1.3: Analyze the real-time network traffic through the traffic identification model constructed in step S1.2, and determine whether there is a security threat. Specifically, obtain the anomaly score corresponding to the real-time network traffic through the traffic identification model, compare the anomaly score with the preset threshold, and determine whether the real-time network traffic is a security threat based on the comparison result. Specifically, when the anomaly score corresponding to the real-time network traffic is greater than the preset threshold, it indicates that there is a security threat to the real-time network traffic. Otherwise, there is no security threat to the real-time network traffic.

[0071] In this embodiment, the anomaly score corresponding to the real-time network traffic is obtained through the autoencoder, as follows:

[0072] Step M1: According to the real-time network traffic, the real-time network traffic is reconstructed through an autoencoder to obtain the reconstructed network traffic.

[0073] Step M2: According to the reconstructed network traffic and the real-time network traffic in step M1, a reconstruction error between the two is obtained, and the reconstruction error is the anomaly score corresponding to the real-time network traffic. In this embodiment, the formula for obtaining the anomaly score is specifically:

[0074] ,

[0075] in: For the abnormality score, is the total amount of real-time network traffic, is the i-th real-time network traffic, is the reconstructed network traffic corresponding to the i-th real-time network traffic, An index of real-time network traffic.

[0076] In this embodiment, the network traffic data obtained in step S1.1 is used to determine the true positive rate and false positive rate corresponding to the network traffic data. At the same time, the specific size of the preset threshold is determined based on the obtained true positive rate and false positive rate. The details are as follows:

[0077] Step N1: According to the network traffic data obtained in step S1.1, the true positive rate and false positive rate of the network traffic data under different thresholds are obtained, specifically:

[0078] ,

[0079] in: is the true positive rate, is the false positive rate, The number of network traffic data with security threats correctly identified. The amount of network traffic data that has not been identified as a security threat. The amount of network traffic data that was falsely identified as a security threat. The number of network traffic data correctly identified as not containing security threats.

[0080] Step N2: According to the true positive rate and false positive rate obtained in step N1, determine the difference index between the two, specifically:

[0081] ,

[0082] in: is the gap index, is the true positive rate, is the false positive rate.

[0083] Step N3: Obtain the gap index corresponding to each threshold value and from all the gap indices The largest gap index is determined in , that is, the maximum gap index The corresponding threshold is the preset threshold that is finally determined.

[0084] That is, when the anomaly score When it is greater than the preset threshold that is finally determined, it indicates that there is a security threat in the real-time network traffic.

[0085] In the specific implementation process, the following table data is set (the 0 corresponding to the real label in the table is the network traffic data without security threats, and 1 is the network traffic data with security threats):

[0086] Anomaly Scoring 0.1 0.4 0.6 0.8 0.9 True label 0 0 1 1 1

[0087] That is, when the preset threshold is 0.5, the number of network traffic data with security threats correctly identified is 3, the number of network traffic data that are incorrectly identified as security threats 0, the number of network traffic data with no identified security threats 0, the number of network traffic data that are correctly identified as not posing a security threat is 2. In other words, the true positive rate The false positive rate is 1. is 0, the gap index is 1.

[0088] When the preset threshold is 0.7, the number of network traffic data with security threats correctly identified 2, the number of network traffic data that are incorrectly identified as security threats 0, the number of network traffic data with no identified security threats 1, the number of network traffic data that are correctly identified as not posing a security threat is 2. In other words, the true positive rate The false positive rate is 0.67. is 0, the gap index It is 0.67.

[0089] When the preset threshold is 0.95, the number of network traffic data with security threats correctly identified 1, the number of network traffic data that are incorrectly identified as security threats 0, the number of network traffic data with no identified security threats =2, the number of network traffic data that are correctly identified as not posing security threats is 2. In other words, the true positive rate The false positive rate is 0.33. is 0, the gap index It is 0.33.

[0090] That is to say, in this embodiment, the preset threshold finally determined is 0.5. When it is greater than 0.5, it indicates that there is a security threat in the real-time network traffic.

[0091] Step S1.4: According to the security threats identified in step S1.3 and the network environment obtained in real time, the initial encryption parameters obtained in step S1.2 are adjusted accordingly to obtain the final adaptive encryption parameters. The details are as follows:

[0092] Step W1: Obtain the anomaly score corresponding to each real-time network flow through the anomaly score acquisition formula in step M2, and determine the corresponding security threat intensity based on all the obtained anomaly scores. It is worth noting that the security threat intensity is the average value of all the anomaly scores.

[0093] During the implementation process, a series of network traffic samples were obtained, and their corresponding anomaly scores were [0.2, 0.4, 0.7, 0.9], that is, their corresponding security threat intensity was 0.55.

[0094] Step W2: According to the network environment obtained in real time, the corresponding comprehensive network performance index is obtained, which is specifically:

[0095] ,

[0096] in: It is a comprehensive indicator of network performance. is the weight corresponding to the bandwidth, is the bandwidth corresponding to the real-time network environment, is the weight corresponding to the delay, The delay corresponding to the real-time network environment.

[0097] Step W3: According to the security threat intensity obtained in step W1 and the comprehensive network performance index obtained in step W2, the adjustment parameters corresponding to the initial encryption parameters are determined, specifically:

[0098] ,

[0099] in: To adjust the parameters, is the coefficient of security threat intensity, is the security threat intensity, is the coefficient of the comprehensive index of network performance, It is a comprehensive indicator of network performance.

[0100] Step W4: According to the adjustment parameters determined in step W3 , the initial encryption parameters obtained in step S1.2 are used to obtain the final adaptive encryption parameters, which are specifically:

[0101] ,

[0102] in: is the adaptive encryption parameter, To adjust the parameters, is the initial encryption parameter.

[0103] It is worth noting that the encryption parameters that can be adjusted include but are not limited to key length (such as 128 bits, 192 bits or 256 bits in the AES algorithm), encryption algorithm selection (such as AES, RSA and other encryption algorithms of different strengths), hash function selection (such as SHA-256, SHA-3, etc.), security parameters (such as initialization vector (IV), number of iterations, etc.).

[0104] In the process of specific implementation, the key length of AES encryption is taken as an example for specific explanation. The initial value of the key length of AES encryption in this embodiment is 128 bits. At the same time, the abnormal score corresponding to the network traffic sample monitored in this embodiment is [0.2, 0.4, 0.7, 0.9], so the corresponding security threat intensity is At the same time, through the above implementation process, it can be known that the adjusted length of the key length is 64 bits, so the final key length is 192 bits.

[0105] Step S2: Update the quantum key. That is, verify the legitimacy of the user identity and device through blockchain technology, manage the generation, distribution and update of the quantum key through smart contracts, and dynamically update the quantum key through distributed nodes. The details are as follows:

[0106] Step S2.1: Verify the legitimacy of the user identity and the device through blockchain technology, and build a quantum key management system based on the verification results and smart contracts. In this embodiment, each user and device will generate a pair of public and private keys when registering. It is worth noting that in this embodiment, the generated public key is recorded in the ledger on the blockchain, and the ledger is associated with the user identifier (such as user name or device ID).

[0107] That is to say, during the verification process, the user can initiate a request and sign the request with a private key. At the same time, the recipient (such as a server or other user) can verify the signature through the public key stored on the blockchain, that is, check whether its private key can match the public key. If the two match successfully, it means that the user or device has passed the verification, otherwise, it means that the user or device has failed the verification.

[0108] Furthermore, rules are defined through smart contracts to trigger the quantum key distribution protocol, which means that only users and devices that have passed the above verification can trigger the quantum key distribution protocol.

[0109] Step S2.2: Obtain the communication delay change rate between nodes in the quantum key management system constructed in step S2.1, specifically:

[0110] ,

[0111] in: is the communication delay change rate, is the communication delay at time t, is the communication delay at time t-△t, is the time interval.

[0112] Furthermore, the obtained communication delay change rate Compared with the preset change rate threshold, when the obtained communication delay change rate When it is greater than the preset change rate threshold, it indicates that the communication delay change rate If the corresponding node is wrong, the next step S2.3 is executed. Otherwise, it indicates that the communication delay change rate The corresponding nodes are correct and can operate normally, that is, the update frequency of the quantum key remains unchanged.

[0113] It is worth noting that the preset change rate threshold in this embodiment is determined based on statistical distribution. That is, the expected value of the communication delay change rate under normal circumstances is obtained through statistical data, that is, the average change rate over a period of time is obtained for setting. Therefore, it can be specifically set according to specific data in actual use, so in this embodiment, it will not be specifically explained.

[0114] Step S2.3: Obtain time series data of network traffic, including but not limited to bandwidth usage, request frequency, response time, etc., as well as all detected security event logs. At the same time, based on the acquired time series data, a prediction model is constructed through machine learning algorithms (such as logistic regression, random forest or neural network, etc.) to predict the probability of security events, specifically:

[0115] ,

[0116] in: for The probability of a security incident occurring at time t, is the characteristic data, for The transpose of is the bias term, is the weight vector.

[0117] That is, when step S2.2 obtains the communication delay change rate When a problem occurs in the corresponding node, the probability of failure is further obtained through step S2.3, and the update frequency of the quantum key established in step S2.1 is adjusted according to the probability of failure.

[0118] In this embodiment, the feature data The probability of a security incident occurring at time t The size of the quantum key is compared with the preset risk threshold, and based on the comparison result, it is determined whether the update frequency of the quantum key needs to be adjusted. Specifically:

[0119] When the characteristic data The probability of a security incident occurring at time t When it is greater than the preset risk threshold, the update frequency of the quantum key is increased; otherwise, the update frequency of the quantum key remains unchanged.

[0120] It is worth noting that the risk threshold in this embodiment can be set based on the acquired time series data. Specifically, the statistical characteristics of the time series data (such as the mean, standard deviation, minimum, maximum, and percentile) are obtained, and the probability distribution of the prediction model output under normal circumstances is obtained, and the average probability distribution within the time period is determined. In other words, the risk threshold is determined by the average probability distribution, which is specifically:

[0121] ,

[0122] in: is the risk threshold, is the average probability distribution, is the coefficient, is the standard deviation of the probability distribution.

[0123] Specifically, the coefficients in this embodiment are The size of is set to 2 or 3. Here, it is set to 3 in this embodiment.

[0124] In the process of specific implementation, it can be known from the above prediction model that under normal circumstances, the mean of the probability distribution output by the prediction model is 0.05 and the standard deviation is 0.02. Therefore, the risk threshold in this embodiment is: 0.05+3*0.02=0.11.

[0125] That is, when the feature data The probability of a security incident occurring at time t When it is greater than 0.11, it indicates that the update frequency of the quantum key needs to be adjusted, that is, the update frequency of the quantum key needs to be increased. Otherwise, the update frequency of the quantum key can remain unchanged.

[0126] Step S3: Determine the encryption strength. That is, the final adaptive encryption parameters are obtained through step W4. , the update frequency of the quantum key determined in step S2.2 and step S2.3, and the final encryption protocol and algorithm are intelligently selected, that is, the encryption strength is adaptively adjusted to meet different levels of security requirements.

[0127] In this embodiment, the final adaptive encryption parameters are obtained through step W4. , adjust the complexity of the encryption algorithm. Specifically, reduce the encryption strength under high load to reduce decryption delay, and increase the encryption strength when processing highly sensitive data to ensure data security.

[0128] Furthermore, we can select the corresponding encryption protocol according to the needs of different application scenarios. For example, on mobile devices, lightweight and efficient encryption schemes are preferred. Similarly, for highly sensitive data, we can choose the symmetric encryption algorithm AES-256 combined with the asymmetric encryption algorithm RSA or ECC for hybrid encryption.

[0129] That is to say, by continuously monitoring the operating status of the system and adjusting the adaptive encryption parameters according to the latest network conditions and security threats, the complexity of the encryption algorithm can be adjusted according to actual needs. At the same time, according to the update frequency of the quantum key, the quantum key is updated in real time during data transmission. As a result, the 5G network real-time encryption technology in this embodiment can not only dynamically adjust the encryption strength according to changes in the network environment, data sensitivity and predicted security risks, but also select the best encryption protocol and algorithm according to actual needs.

[0130] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is limited by the attached embodiments and their equivalents.

Claims

1. A 5G network real-time encryption method based on artificial intelligence, characterized in that: Included are: Adjust encryption parameters: Analyze network traffic patterns through AI algorithms to obtain initial encryption parameters. At the same time, identify security threats by identifying real-time network traffic patterns. Adjust the initial encryption parameters based on the security threats and network environment to obtain adaptive encryption parameters, as follows: Acquire network traffic data, and preprocess the network traffic data to acquire feature data; A traffic identification model is constructed through the characteristic data, and initial encryption parameters are obtained through the traffic identification model; By using the traffic identification model, an anomaly score corresponding to the real-time network traffic is obtained, and the anomaly score is compared with a preset threshold, and a security threat is determined based on the comparison result; According to the security threat and network environment, the initial encryption parameters are adjusted to obtain adaptive encryption parameters, as follows: W1: According to the anomaly score corresponding to each real-time network flow, the average value of all the anomaly scores is obtained to determine the security threat intensity; W2: Determine the comprehensive network performance indicators based on the network environment, specifically: ， in: It is a comprehensive indicator of network performance. is the weight corresponding to the bandwidth, is the bandwidth corresponding to the real-time network environment, is the weight corresponding to the delay, The delay corresponding to the real-time network environment; W3: Obtain adjustment parameters corresponding to the initial encryption parameters through the security threat intensity and network performance comprehensive index, specifically: ， in: To adjust the parameters, is the coefficient of security threat intensity, is the security threat intensity, is the coefficient of the comprehensive index of network performance, It is a comprehensive indicator of network performance; W4: According to the adjustment parameter, the initial encryption parameter is adjusted, specifically: ， in: is the adaptive encryption parameter, To adjust the parameters, is the initial encryption parameter; Update quantum keys: Use blockchain methods to verify users and devices, and use smart contracts to manage the generation, distribution, and update of quantum keys. At the same time, quantum keys are updated through distributed nodes. Determining encryption strength: determining the final encryption protocol and algorithm through the encryption parameters and the updated quantum key; Update the quantum key as follows: Verify users and devices through the blockchain method, and build a quantum key management system based on the verification results and smart contracts; Obtaining a communication delay change rate between nodes in the quantum key management system, and comparing the communication delay change rate with a change rate threshold, when the communication delay change rate is greater than the change rate threshold, the node corresponding to the communication delay change rate is incorrect, and the next step is executed; otherwise, the node corresponding to the communication delay change rate is correct, and the update frequency of the quantum key remains unchanged; Obtaining time series data of network traffic, and building a prediction model to obtain the probability of a security incident, and comparing the probability of a security incident with a risk threshold. When the probability of a security incident is greater than the risk threshold, increasing the frequency of updating the quantum key, otherwise, keeping the frequency of updating the quantum key unchanged; Build a prediction model to obtain the probability of a security incident, specifically: ， in: for The probability of a security incident occurring at time t, is the characteristic data, for The transpose of is the bias term, is the weight vector; Risk thresholds, specifically: ， in: is the risk threshold, is the average probability distribution, is the coefficient, is the standard deviation of the probability distribution.

2. According to the 5G network real-time encryption method based on artificial intelligence according to claim 1, it is characterized in that: Based on the comparison results, security threats are identified, specifically: When the anomaly score is greater than a preset threshold, the real-time network traffic corresponding to the anomaly score has a security threat, otherwise, the real-time network traffic corresponding to the anomaly score has no security threat.

3. A 5G network real-time encryption method based on artificial intelligence according to claim 1 or 2, characterized in that: The method for obtaining anomaly scores is as follows: M1: Reconstruct the real-time network traffic through the autoencoder to obtain the reconstructed network traffic; M2: According to the reconstructed network traffic and the real-time network traffic, a reconstruction error corresponding to the real-time network traffic is obtained. The reconstruction error is the anomaly score corresponding to the real-time network traffic. The formula for obtaining the anomaly score is specifically: ， in: For the abnormality score, is the total amount of real-time network traffic, is the i-th real-time network traffic, is the reconstructed network traffic corresponding to the i-th real-time network traffic, An index of real-time network traffic.

4. A 5G network real-time encryption method based on artificial intelligence according to claim 1 or 2, characterized in that: The method for obtaining the preset threshold is as follows: N1: According to the network traffic data, the true positive rate and false positive rate of the network traffic data under different thresholds are obtained, specifically: ， in: is the true positive rate, is the false positive rate, The number of network traffic data with security threats correctly identified. The amount of network traffic data that has not been identified as a security threat. The amount of network traffic data that was falsely identified as a security threat. The number of network traffic data correctly identified as not containing security threats; N2: According to the true positive rate and the false positive rate, obtain the gap index, specifically: ， in: is the gap index, is the true positive rate, is the false positive rate; N3: Determine the maximum value gap index according to the gap index corresponding to each threshold value, and the threshold value corresponding to the maximum value gap index is the preset threshold value.

Images