Electric power 5G eSIM terminal and secondary authentication method thereof

Abstract

According to the electric power 5G eSIM terminal and the secondary authentication method thereof provided by the invention, by introducing a dynamic behavior analysis algorithm and a dynamic encryption algorithm based on a pseudo tag, compared with the prior art, the security and data transmission efficiency of the electric power 5G terminal are remarkably improved. And the dynamic behavior analysis algorithm monitors the behaviors of the power 5G terminal in real time and identifies abnormal activities, so that the security of identity verification is enhanced. And the dynamic encryption algorithm dynamically generates an encryption key and selects an encryption algorithm according to context information, so that the security and adaptability of data transmission are improved, particularly, the performance is outstanding under the conditions of data scarcity and complex network environment, and therefore, on the premise of not sacrificing efficiency, the security and adaptability of data transmission are improved. And a more flexible and highly-adaptive security guarantee is provided for an electric power 5G management and control platform and a state grid secret service platform.

Description

Technical Field

[0001] The present invention relates to the field of power communication technology, and in particular to a power 5G eSIM terminal and a secondary authentication method thereof. Background Art

[0002] With the rapid development of information technology, the power industry is undergoing a digital transformation. The introduction of 5G communication technology provides strong support for the intelligence and automation of power systems. 5G lightweight communication terminals can achieve higher data transmission rates and lower latency, facilitating real-time monitoring, fault diagnosis and remote control of power equipment. The widespread use of these terminals not only improves the security and reliability of the power grid, but also provides power companies with flexible network solutions that support a variety of application scenarios. However, as the application scenarios of power private networks become increasingly complex, the industry faces multiple security challenges.

[0003] (1) Data security issues: As a critical infrastructure, the power industry faces the risk of cyber attacks such as data leakage and service interruption. Traditional identity authentication methods have the possibility of being exploited by attackers and are difficult to effectively prevent identity forgery. In addition, the need for confidentiality and integrity protection during data transmission is also an important challenge. Traditional encryption methods have insufficient performance when processing large amounts of data, and key management is complex.

[0004] (2) Communication design issues of 5G terminals: In the power communication network, it is generally necessary to process the transmission of small data packets, with medium to high bandwidth requirements and extremely strict requirements on latency and reliability. Although the flexible deployment of 5G terminals reduces the cost of communication construction and improves the control response speed of the power grid and the monitoring of power equipment, in actual applications, the current 5G modules have only two 7816 interfaces at most, which cannot meet the free switching options of China Mobile eSIM, China Unicom / China Telecom eSIM, and USIM cards at the same time, which brings challenges to the communication design of 5G eSIM terminals.

[0005] (3) Operator switching issue: At present, due to the protection of their own card number resources by domestic operators, it is impossible to realize the function of switching between multiple operators based on one eSIM card, which poses certain challenges to the power industry.

[0006] (4) Problems with secondary authentication: Traditional identity authentication is usually a one-time static authentication, which is difficult to cope with security challenges in a dynamic environment and is vulnerable to impersonation or attack. After a single authentication, the system may be affected by undetected vulnerabilities or malicious code, resulting in the risk of the terminal being exploited in subsequent operations. Currently, authentication usually does not have a real-time monitoring and feedback mechanism, and cannot immediately identify security vulnerabilities or abnormal behavior. Summary of the invention

[0007] In view of this, the purpose of the present invention is to provide a power 5G eSIM terminal and a secondary authentication method to improve the flexibility, security and reliability of the power private network and effectively ensure the security and flexibility of the power communication terminal.

[0008] To achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a power 5G eSIM terminal, the power 5GeSIM terminal includes a 5G module including a built-in eSIM, an external eSIM module and a physical USIM card; wherein the built-in eSIM is responsible for providing the communication number of China Unicom or China Telecom, and the external eSIM module is responsible for providing the communication number of China Mobile, China Unicom and China Telecom, and the physical USIM card can communicate with the numbers of China Mobile, China Unicom and China Telecom; when the physical USIM card is inserted, the external eSIM module is disabled, and the built-in eSIM is not affected; and when the physical USIM card is pulled out, the external eSIM module is enabled and powered on, and the built-in eSIM can also take effect, ensuring that the USIM cards and eSIMs of the three major operators can communicate on the 5G terminal.

[0009] The present invention also provides a secondary authentication method based on a power 5G eSIM terminal, which adopts the power 5G eSIM terminal and includes the following steps:

[0010] Step 1: eSIM pre-set certificate; the power 5G control platform generates an eUICC certificate before the eSIM chip leaves the factory, and the eUICC certificate is integrated with the platform information and loaded into the chip; when the device activates the eSIM service for the first time, this pre-set certificate is automatically verified to ensure the legitimacy of the device identity and network access rights;

[0011] Step 2: Code number application; the power 5G management and control platform obtains code numbers in batches from operators, and the operator returns the encrypted information of the code numbers, including ICCID, IMSI, KI, OPC and MSISDN;

[0012] Step 3: Secondary authentication and code number download; When the power 5G terminal is connected to the Internet for the first time, the security agent SDK and eSIM chip encryption application are used to complete the two-way identity authentication with the power 5G control platform and the State Grid secret service platform; After the authentication is passed, the power 5G terminal will use the services of the power 5G control platform and the eSIM chip to automatically download and configure the network code number for fast access; After the power 5G terminal successfully accesses the 5G network, if it needs to access a specific power private network, the terminal must perform secondary authentication; The secondary authentication is performed between the terminal and the authentication server, and the authentication server feedbacks the relevant QoS parameters, MAC address, IP address and user billing information based on the terminal information;

[0013] Step 4: After completing the secondary authentication, the power 5G terminal can start uploading key data; the terminal encrypts important business data through the embedded security chip and sends the encrypted data to the business system.

[0014] In a preferred embodiment, step 3 monitors and analyzes the behavior of the power 5G terminal in real time based on a pseudo-label dynamic behavior analysis algorithm to build its behavior model, which specifically includes the following steps:

[0015] Step 31: Data preparation;

[0016] 1) Collect labeled data

[0017] First, collect and organize well-labeled data and corresponding labels to form a labeled dataset D labled :

[0018] D labled ={(X i ,y i )},y i For the label (y i =1 means normal; y i =0 indicates abnormality);

[0019] Among them, X i is the behavior feature vector of power 5G terminal i (i=1~k), X i ={x1,x2,···,x n}, x1 is the first behavior feature of the power 5G terminal i, x2 is the second behavior feature of the power 5G terminal i, x n is the nth behavior characteristic of power 5G terminal i;

[0020] 2) Collect unlabeled data: Collect a large number of unlabeled data sets at the same time to form an unlabeled data set D unlabled , extract the feature vector F of the unlabeled data j , used for subsequent pseudo-label generation;

[0021] D unlabeled ={F j};

[0022] Among them, F j is the behavioral feature vector of the unlabeled power 5G terminal j (j=1~m),

[0023] F j ={x′1,x′2,···,x′ n}, x′1 is the first behavior feature of the power 5G terminal j, x′2 is the second behavior feature of the power 5G terminal j, and x′ nis the nth behavior characteristic of power 5G terminal j;

[0024] Step 32: Pre-training model;

[0025] Train the initial model: Use the labeled dataset D labeled , train a preliminary power 5G terminal behavior classification model f:

[0026] f * = arg min L(D labeled ,f)

[0027] The loss function adopts the improved binary cross entropy loss function, which is expressed as:

[0028]

[0029] Among them, X i is the feature vector of the i-th power 5G terminal, w1 is the weight of the positive class (normal behavior), w0 is the weight of the negative class (abnormal behavior), f(X i ) is the model’s predicted probability that the i-th sample belongs to the positive class; weights w1 and w0 can be set according to the ratio of positive and negative samples to balance the contribution of the two types of samples to the loss function;

[0030] Step 33: Generate pseudo labels

[0031] Apply pre-trained model: Use the trained model f * Applied to unlabeled dataset D unlabeled , generate a pseudo label set

[0032]

[0033] Step 34: Construct a new training set

[0034] Merge dataset: Combine the generated pseudo labels with the labeled data to form a new mixed training set:

[0035]

[0036] Step 35: Model Retraining

[0037] Train the model on a new dataset: Use the new mixed dataset D new Retrain the model;

[0038] f ** =argminL(D new ,f)

[0039] Step 36: Keep iterating

[0040] Feedback and update: In actual operation, the performance of the model is continuously monitored; if new labeled samples are obtained, they can be included in the training set to further optimize the model;

[0041] Step 37: Dynamic Threshold Adjustment

[0042] Dynamically adjust the threshold θ based on real-time data:

[0043]

[0044] in, σ P is the mean and standard deviation of the most recent N historical prediction probabilities, k1 is an adjustable parameter;

[0045] Step 38: Anomaly Detection

[0046] When the power 5G terminal performs secondary authentication, the real-time data X real-time (Feature vector of the power 5G terminal) is input into the trained model f ** , perform anomaly detection, the probability P that the current power 5G terminal behavior belongs to normal behavior:

[0047] P(y i =1|X real-time )=f ** (X real-time )

[0048] According to the threshold value θ, we can make a judgment:

[0049] if P(y i =1||X real-time )>θ:y i =1 (normal behavior)

[0050] else:y i =0 (abnormal behavior)

[0051] When abnormal behavior is detected, the system will trigger additional security measures, such as requiring the power 5G terminal to perform additional identity authentication to ensure the authenticity of the power 5G terminal when accessing power private network services.

[0052] In a preferred embodiment, step 4 includes a dynamic encryption algorithm, dynamically generating an encryption key and selecting an encryption algorithm according to context information; specifically, the following steps are included:

[0053] Step 41: context information collection: collect context information that affects encryption, including the power 5G terminal identity U, network status N, and data sensitivity S;

[0054] Step 42: Key generation: Generate dynamic key K based on context information:

[0055] K=Hash(U||N||S||timestamp)

[0056] Among them, || indicates the splicing operation, and timestamp is the timestamp;

[0057] Step 43: Encryption algorithm selection: Select encryption algorithm A based on data sensitivity:

[0058]

[0059] For highly sensitive data, strong encryption algorithms are used to ensure high security during data transmission;

[0060] For ordinary data, choose a lightweight encryption algorithm to increase processing speed while ensuring security;

[0061] Step 44: Data encryption: Encrypt the data Data using the dynamically generated key K and the selected encryption algorithm:

[0062] C=A(K,Data)

[0063] Among them, C is the ciphertext;

[0064] Step 45: Decryption process; the receiving end uses the same context information to generate a key for decryption:

[0065] Data = A -1 (K,C).

[0066] Compared with the prior art, the present invention has the following beneficial effects:

[0067] 1. Based on the pseudo-label dynamic behavior analysis algorithm as a supplementary layer, it provides real-time monitoring and anomaly detection to help identify potential security threats. Combined with dynamic behavior analysis, the flexibility and adaptability of the system can be improved. The algorithm effectively combines labeled data and unlabeled data, thereby improving the training effect and generalization ability of the model.

[0068] 2. Dynamic encryption algorithm dynamically generates encryption keys and selects encryption algorithms based on context information to improve the security of data transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 This is a schematic diagram of a 5G eSIM terminal communication architecture according to a preferred embodiment of the present invention;

[0070] Figure 2 This is a schematic diagram of an eSIM platform according to a preferred embodiment of the present invention;

[0071] Figure 3 The figure is a schematic diagram of the overall authentication process of a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0072] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0073] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0074] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application; as used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or their combinations.

[0075] refer to Figure 1-3 In most power distribution automation terminal application scenarios, the flexible deployment of 5G terminals reduces the cost of communication construction and improves the control response speed of the power grid and the monitoring intensity of power equipment. However, the current 5G module has only two 7816 interfaces at most, which cannot meet the free switching selection of mobile eSIM, China Unicom / China Telecom eSIM, and USIM card at the same time. To solve this problem, an external eSIM module is introduced, and the automatic switching mechanism between USIM and eSIM is realized by hardware switch to ensure that mobile eSIM and USIM card can be switched freely. At the same time, China Unicom / China Telecom eSIM is built into the module through another 7816 port or externally placed outside the module (built-in eSIM includes 7816 interface and SPI interface). Specifically, when the physical USIM card is inserted, the eSIM module is disabled; and when the physical USIM card is pulled out, the eSIM module is enabled and powered on. This design can flexibly meet different communication needs and ensure the effective switching of SIM card types in various scenarios. In terms of number management, mobile eSIM can only write numbers on the external network air interface through the mobile eSIM platform after presetting the eSIM seed number;

[0076] Telecom / Unicom eSIM can support the writing of numbers on the power eSIM control platform. After the eSIM pre-sets the private network seed number, the number can be written through the power intranet platform, or the 5G terminal communicates with the intranet platform through the USIM card, and the eSIM chip writes the number through the SPI interface. For Telecom / Unicom eSIM, secondary authentication is achieved by embedding the State Grid secret service certificate and algorithm in the eSIM chip; Mobile eSIM uses quantum T card (SDIO to USB) or quantum soft key for quantum communication authentication.

[0077] This design enables the switching of eSIM cards between multiple operators. Through the design of hardware switches and external eSIM modules, combined with various interfaces and functions, the ability to flexibly manage and switch different eSIM cards on terminal devices is achieved. This switching method provides users with greater freedom and flexibility, allowing them to choose services from different operators according to their needs without having to replace physical SIM cards, thereby improving user experience and convenience.

[0078] Secondary authentication includes the following steps:

[0079] Step 1: eSIM provisioning certificate

[0080] The power 5G control platform generates an eUICC certificate before the eSIM chip leaves the factory. The certificate is integrated with the platform information and loaded into the chip. When the device activates the eSIM service for the first time, this pre-set certificate is automatically verified to ensure the legitimacy of the device identity and network access rights.

[0081] Arrangement 2: Code number application

[0082] The power 5G control platform obtains code numbers in batches from operators, and the operators return the encrypted information of the code numbers, including ICCID, IMSI, KI, OPC, and MSISDN. This process ensures the security and privacy protection of the code numbers.

[0083] Step 3: Secondary authentication and code download

[0084] When the power 5G terminal is connected to the Internet for the first time, it will complete two-way identity authentication with the power 5G control platform and the State Grid's encryption service platform through the security agent SDK and eSIM chip encryption application. After the authentication is passed, the power 5G terminal will use the services of the power 5G control platform and the eSIM chip to automatically download and configure the network code number for fast access.

[0085] Once the electric power 5G terminal successfully accesses the 5G network, if it needs to access a specific electric power private network, the terminal must undergo secondary authentication. This process requires the terminal to carry specific user credentials, which are issued by the electric power private network, rather than the credentials provided by the operator. Authentication is performed between the terminal and the authentication server, which feeds back relevant QoS parameters, MAC address, IP address, and user billing information based on the terminal information. This mechanism ensures the authenticity of the identity of the electric power 5G terminal when accessing the electric power private network service, and provides the necessary secure connection information for the 5G network.

[0086] Specifically, although the national secret algorithm provides a high level of security, the pseudo-label-based dynamic behavior analysis algorithm can serve as a supplementary layer to provide real-time monitoring and anomaly detection to help identify potential security threats. Combined with dynamic behavior analysis, the flexibility and adaptability of the system can be improved. For example, when the behavior of the power 5G terminal changes, the system can automatically adjust the security measures instead of relying solely on static identity authentication mechanisms. Therefore, before the two-way identity authentication with the power 5G management and control platform and the State Grid secret service platform. In order to improve the security of the authentication process, a pseudo-label-based dynamic behavior analysis algorithm is introduced to monitor the power 5G terminals in real time and identify abnormal activities to enhance identity authentication and system security. Through the self-supervised learning method, the algorithm effectively combines labeled data and unlabeled data, thereby improving the training effect and generalization ability of the model, especially when data is scarce. The algorithm mainly includes the following steps:

[0087] (1) Data preparation

[0088] 1) Collect labeled data

[0089] First, collect and organize a well-annotated dataset and corresponding labels. For example:

[0090] D labled ={(X i ,y i )}where y i Label(y i =1 means normal; y i =0 indicates abnormality);

[0091] Among them, X i is the characteristic vector of power 5G terminal i (i=1~k), X i ={x1,x2,···,x n} is the behavioral characteristics of the power 5G terminal i (including login time, device type, IP address, geographic location, operation type, network delay, device power, etc.);

[0092] 2) Collect unlabeled data: Collect a large number of unlabeled data sets at the same time to form an unlabeled data set D unlabled , extract the feature vector F of the unlabeled data j , used for subsequent pseudo-label generation;

[0093] D unlabeled ={F j};

[0094] Among them, F j is the feature vector of the unlabeled power 5G terminal j (j=1~m),

[0095] Fj ={x′1,x′2,···,x′ n} is the behavioral characteristics of the power 5G terminal j (including login time, device type, IP address, geographic location, operation type, network delay, device power, etc.);

[0096] (2) Pre-trained Model

[0097] 1) Train the initial model: Use a labeled dataset D labeled , train a preliminary power 5G terminal behavior classification model f:

[0098] f * = arg min L(D labeled ,f)

[0099] 2) The loss function adopts the improved binary cross entropy loss function, which is expressed as:

[0100]

[0101] Among them, X i is the feature vector of the i-th power 5G terminal, w1 is the weight of the positive class (normal behavior), w0 is the weight of the negative class (abnormal behavior), f(X i ) is the model's predicted probability that the i-th sample belongs to the positive class. The weights w1 and w0 can be set according to the ratio of positive and negative samples to balance the contribution of the two types of samples to the loss function;

[0102] (3) Generate pseudo labels

[0103] Apply pre-trained model: Use the trained model f * Applied to unlabeled dataset D unlabeled , generate a pseudo label set

[0104]

[0105] (4) Construct a new training set

[0106] Merge dataset: Combine the generated pseudo labels with the labeled data to form a new mixed training set:

[0107]

[0108] (5) Model retraining

[0109] Train the model on a new dataset: Use the new mixed dataset D new Retrain the model;

[0110] f ** =argminL(D new,f)

[0111] (6) Continuous iteration

[0112] Feedback and update: In actual operation, the performance of the model is continuously monitored; if new labeled samples are obtained, they can be included in the training set to further optimize the model;

[0113] (7) Dynamic Threshold Adjustment

[0114] Dynamically adjust the threshold θ based on real-time data:

[0115]

[0116] in, σ P is the mean and standard deviation of the most recent N historical prediction probabilities, k1 is an adjustable parameter;

[0117] (8) Anomaly Detection

[0118] When the power 5G terminal performs secondary authentication, the real-time data X real-time (Feature vector of the power 5G terminal) is input into the trained model f ** , perform anomaly detection, the probability P that the current terminal behavior belongs to normal behavior:

[0119] P(y i =1|X real-time )=f ** (X real-time )

[0120] According to the threshold value θ, we can make a judgment:

[0121] if P(y i =1||X real-time )>θ:y i =1 (normal behavior)

[0122] else:y i =0 (abnormal behavior)

[0123] When abnormal behavior is detected, the system will trigger additional security measures, such as requiring the power 5G terminal to perform additional identity authentication to ensure the authenticity of the power 5G terminal when accessing power private network services.

[0124] After authentication, the power 5G terminal will use the services of the power 5G management and control platform and the eSIM chip to automatically download and configure the network code number for fast access.

[0125] Step 4: Encrypted upload and download of key data

[0126] After completing the secondary authentication, the power 5G terminal can start uploading key data. The terminal encrypts important business data through the embedded security chip and sends the encrypted data to the business system. After receiving the ciphertext, the business system will call the unified password service platform for decryption and return the plaintext data. Similarly, when the business system needs to send key data, it first calls the unified password service platform to encrypt the plaintext data, and then sends the encrypted data to the power 5G communication terminal. After receiving the data, the terminal decrypts it through the embedded security chip to obtain the required plaintext data.

[0127] The terminal encrypts important business data through the built-in security chip and sends the encrypted data to the business system. To improve data security and transmission efficiency, the system introduces a dynamic encryption algorithm to dynamically generate encryption keys and select encryption algorithms based on context information to improve the security of data transmission.

[0128] Specific implementation steps:

[0129] (1) Context Information Collection

[0130] Collect context information that affects encryption, including the power 5G terminal identity U, network status N, and data sensitivity S.

[0131] (2) Key Generation

[0132] Generate dynamic key K based on context information:

[0133] K=Hash(U||N||S||timestamp)

[0134] Among them, || indicates the concatenation operation, and timestamp is the timestamp.

[0135] (3) Encryption selection

[0136] Select an encryption algorithm based on data sensitivity:

[0137]

[0138] For highly sensitive data (such as control instructions), strong encryption algorithms (such as AES-256) are used to ensure high security of data during transmission.

[0139] For common data (such as status monitoring information), you can choose a lightweight encryption algorithm (such as AES-128) to increase processing speed while ensuring security.

[0140] (4) Data encryption

[0141] Encrypt the data Data using the dynamically generated key K and the selected encryption algorithm:

[0142] C=A(K,Data)

[0143] Among them, C is the ciphertext.

[0144] (5) Decryption process

[0145] The receiving end uses the same context information to generate the key for decryption:

[0146] Data = A -1 (K,C)

[0147] Through the proposed dynamic encryption algorithm, the power 5G terminal can effectively improve the security and transmission efficiency of data during the upload and download of key data. This mechanism not only ensures the confidentiality and integrity of the data, but also can adapt to the changing network environment and enhance the flexibility and responsiveness of the system.

[0148] Compared with the prior art, the present invention significantly improves the security and data transmission efficiency of power 5G terminals by introducing a dynamic behavior analysis algorithm and a dynamic encryption algorithm based on pseudo-labels. The dynamic behavior analysis algorithm monitors the behavior of power 5G terminals in real time and identifies abnormal activities, enhancing the security of identity authentication. The dynamic encryption algorithm dynamically generates encryption keys and selects encryption algorithms based on context information, improving the security and adaptability of data transmission, especially in cases where data is scarce and the network environment is complex. This provides a more flexible and adaptable security guarantee for the power 5G management and control platform and the State Grid secret service platform without sacrificing efficiency.

Claims

1. A power 5G eSIM terminal, characterized in that: The electric power 5G eSIM terminal includes a 5G module with a built-in eSIM, an external eSIM module and a physical USIM card; the built-in eSIM is responsible for providing the communication number of China Unicom or China Telecom, and the external eSIM module is responsible for providing the communication number of China Mobile. The physical USIM card can communicate with the numbers of China Mobile, China Unicom and China Telecom; when the physical USIM card is inserted, the external eSIM module is disabled and the built-in eSIM is not affected; and when the physical USIM card is pulled out, the external eSIM module is enabled and powered on, and the built-in eSIM can also take effect, ensuring that the USIM cards and eSIMs of the three major operators can communicate on the 5G terminal.

2. A secondary authentication method based on a power 5G eSIM terminal, characterized in that: The electric power 5G eSIM terminal described in claim 1 above is adopted, comprising the following steps: Step 1: eSIM pre-set certificate; the power 5G control platform generates an eUICC certificate before the eSIM chip leaves the factory, and the eUICC certificate is integrated with the platform information and loaded into the chip; when the device activates the eSIM service for the first time, this pre-set certificate is automatically verified to ensure the legitimacy of the device identity and network access rights; Step 2: Code number application; the power 5G management and control platform obtains code numbers in batches from operators, and the operator returns the encrypted information of the code numbers, including ICCID, IMSI, KI, OPC and MSISDN; Step 3: Secondary authentication and code number download; When the power 5G terminal is connected to the Internet for the first time, the security agent SDK and eSIM chip encryption application are used to complete the two-way identity authentication with the power 5G control platform and the State Grid secret service platform; After the authentication is passed, the power 5G terminal will use the services of the power 5G control platform and the eSIM chip to automatically download and configure the network code number for fast access; After the power 5G terminal successfully accesses the 5G network, if it needs to access a specific power private network, the terminal must perform secondary authentication; The secondary authentication is performed between the terminal and the authentication server, and the authentication server feedbacks the relevant QoS parameters, MAC address, IP address and user billing information based on the terminal information; Step 4: After completing the secondary authentication, the power 5G terminal can start uploading key data; the terminal encrypts important business data through the embedded security chip and sends the encrypted data to the business system.

3. According to claim 2, a secondary authentication method based on a power 5G eSIM terminal is characterized in that: The step 3 monitors and analyzes the behavior of the power 5G terminal in real time based on the pseudo-label dynamic behavior analysis algorithm to build its behavior model, which specifically includes the following steps: Step 31: Data preparation; 1) Collect labeled data First, collect and organize well-labeled data and corresponding labels to form a labeled dataset D labled : D labled ={(X i ,y i )},y i For the label (y i =1 means normal; y i =0 indicates abnormality); Among them, X i is the behavior feature vector of power 5G terminal i (i=1~k), X i ={x1,x2,···,x n }, x1 is the first behavior feature of the power 5G terminal i, x2 is the second behavior feature of the power 5G terminal i, x n is the nth behavior characteristic of power 5G terminal i; 2) Collect unlabeled data: Collect a large number of unlabeled data sets at the same time to form an unlabeled data set D unlabled , extract the feature vector F of the unlabeled data j , used for subsequent pseudo-label generation; D unlabeled ={F j };; Among them, F j is the behavioral feature vector of the unlabeled power 5G terminal j (j=1~m), F j ={x′1,x′2,···,x′ n }, x′1 is the first behavior feature of the power 5G terminal j, x′2 is the second behavior feature of the power 5G terminal j, and x′ n is the nth behavior characteristic of power 5G terminal j; Step 32: Pre-training model; Train the initial model: Use the labeled dataset D labeled , train a preliminary power 5G terminal behavior classification model f: f * =arg min L(D labeled ,f) The loss function adopts the improved binary cross entropy loss function, which is expressed as: Among them, X i is the feature vector of the i-th power 5G terminal, w1 is the weight of the positive class (normal behavior), w0 is the weight of the negative class (abnormal behavior), f(X i ) is the model’s predicted probability that the i-th sample belongs to the positive class; weights w1 and w0 can be set according to the ratio of positive and negative samples to balance the contribution of the two types of samples to the loss function; Step 33: Generate pseudo labels Apply pre-trained model: Use the trained model f * Applied to unlabeled dataset D unlabeled , generate a pseudo label set Step 34: Construct a new training set Merge dataset: Combine the generated pseudo labels with the labeled data to form a new mixed training set: Step 35: Model Retraining Train the model on a new dataset: Use the new mixed dataset D new Retrain the model; f ** =argminL(D new ,f) Step 36: Keep iterating Feedback and update: In actual operation, the performance of the model is continuously monitored; if new labeled samples are obtained, they can be included in the training set to further optimize the model; Step 37: Dynamic Threshold Adjustment Dynamically adjust the threshold θ based on real-time data: in, σ P is the mean and standard deviation of the most recent N historical prediction probabilities, k1 is an adjustable parameter; Step 38: Anomaly Detection When the power 5G terminal performs secondary authentication, the real-time data X real-time (Feature vector of the power 5G terminal) is input into the trained model f ** , perform anomaly detection, the probability P that the current power 5G terminal behavior belongs to normal behavior: P(y i =1|X real-time )=f ** (X real-time ) According to the threshold value θ, we can make a judgment: if P(y i =1|X real-time )>θ:y i =1 (normal behavior) else:y i =0 (abnormal behavior) When abnormal behavior is detected, the system will trigger additional security measures, such as requiring the power 5G terminal to perform additional identity authentication to ensure the authenticity of the power 5G terminal when accessing power private network services.

4. According to claim 2, a secondary authentication method based on a power 5G eSIM terminal is characterized in that: The step 4 includes a dynamic encryption algorithm, dynamically generating an encryption key and selecting an encryption algorithm based on context information; The specific steps include: Step 41: context information collection: collect context information that affects encryption, including the power 5G terminal identity U, network status N, and data sensitivity S; Step 42: Key generation: Generate dynamic key K based on context information: K=Hash(U||N||S||timestamp) Among them, || indicates the splicing operation, and timestamp is the timestamp; Step 43: Encryption algorithm selection: Select encryption algorithm A based on data sensitivity: For highly sensitive data, strong encryption algorithms are used to ensure high security during data transmission; For ordinary data, choose a lightweight encryption algorithm to increase processing speed while ensuring security; Step 44: Data encryption: Encrypt the data Data using the dynamically generated key K and the selected encryption algorithm: C=A(K,Data) Among them, C is the ciphertext; Step 45: Decryption process; the receiving end uses the same context information to generate a key for decryption: Data=A -1 (K,C)。

Images