A data encryption method for Beidou short message transmission

Abstract

The present invention relates to the field of Beidou short message technology, and proposes a data encryption method for Beidou short message transmission, comprising the following steps: S1 sets a hybrid homomorphic encryption framework, the hybrid homomorphic encryption framework includes a full homomorphic encryption algorithm and a partial homomorphic encryption algorithm; S2 receives a short message input by a transmitter, and pre-processes the short message; S3 performs a hybrid encryption operation on the pre-processed short message through the full homomorphic encryption algorithm and the partial homomorphic encryption algorithm in the hybrid homomorphic encryption framework, and adds differential privacy noise during the encryption operation to obtain an intermediate file; S4 generates a key according to the intermediate file and distributes it to the transmitter and the receiver. The data encryption method for Beidou short message transmission can work under limited bandwidth and high real-time requirements, can adapt to the security needs of different users, and maintain stable encryption performance in a complex and changeable use environment.

Description

Technical Field

[0001] The present invention relates to the field of Beidou short message technology, and in particular to a data encryption method for Beidou short message transmission. Background Art

[0002] BeiDou Navigation Satellite System (BDS) is a global satellite navigation system independently developed by China, with positioning, navigation, timing and short message communication functions. In the field of short message communication, the BeiDou system can provide users with two-way transmission services, which is particularly important in remote areas without network coverage. However, the characteristic of short message service is that the transmission bandwidth is limited, which makes it difficult to directly apply traditional encryption algorithms and there is a risk of privacy leakage.

[0003] Most of the encryption algorithms in the current Beidou short message encryption require complex mathematical calculations, which will lead to excessive energy consumption and processing time on mobile terminals and satellite devices with limited computing resources, affecting the overall efficiency of the system. At the same time, in order to ensure the security of communication, it is often necessary to add additional encryption overhead, such as introducing randomness and redundant information. This has caused a decrease in the proportion of effective data transmission in the Beidou short message service with extremely limited bandwidth. Summary of the invention

[0004] In view of this, the present invention proposes a data encryption method for Beidou short message transmission, which can work under limited bandwidth and high real-time requirements, can adapt to the security needs of different users, and maintain stable encryption performance in complex and changeable usage environments.

[0005] The technical solution of the present invention is achieved in this way:

[0006] On the one hand, the present invention provides a data encryption method for Beidou short message transmission, comprising the following steps:

[0007] S1 sets a hybrid homomorphic encryption framework, which includes a fully homomorphic encryption algorithm and a partially homomorphic encryption algorithm;

[0008] S2 receives the short message input by the sender and pre-processes the short message;

[0009] S3 uses the fully homomorphic encryption algorithm and the partially homomorphic encryption algorithm in the hybrid homomorphic encryption framework to perform hybrid encryption operations on the pre-processed short messages, and adds differential privacy noise during the encryption operation to obtain the intermediate file;

[0010] S4 generates a key based on the intermediate file and distributes it to the sender and receiver;

[0011] The S5 sender uses the key to encrypt the short message to obtain an encrypted file, and sends the encrypted file to the receiver through the Beidou short message transmission system;

[0012] The S6 receiving end decrypts the encrypted file using the key and performs post-processing to obtain a short message.

[0013] Based on the above technical solution, preferably, in step S3, adding differential privacy noise includes the following sub-steps:

[0014] S31 clusters multiple short messages and associates them with each other to form a message chain;

[0015] S32 calculates the sensitivity score of the short message according to the length of the short message, the length of the content sensitive information and the frequency of the short message in the message chain;

[0016] S33 is based on differential privacy and determines the noise level of short messages according to the sensitivity score, privacy budget, and the correlation between short messages in the message chain;

[0017] S34 adds noise to the short message according to the noise level of the short message.

[0018] Further preferably, the sensitivity score Score is:

[0019] Score = w l ·L+w f ·F+w q Q

[0020] Among them, w1 is the length of the short message, L is the weight of the length, and w f is the frequency of short messages in the message chain, F is the weight of the frequency, and w q is the content-sensitive information length of the short message, and Q is the weight of the content-sensitive information length.

[0021] Further preferably, the noise level M(x) is:

[0022]

[0023] Where ∈ is the privacy budget, q(X) is the original query function, Lap() is the Laplace noise function, and Δq is the sensitivity of the query.

[0024] Further preferably, after step S32, a special score S is also set. When the sensitivity score reaches the special score S, N(S)=α·S+β, wherein N(S) is the noise level after the sensitivity score reaches the special score S, and α and β are adjustment parameters used to adjust the intensity and offset of the noise, respectively.

[0025] On the basis of the above technical scheme, preferably, in step S3, the fully homomorphic encryption algorithm and the partially homomorphic encryption algorithm in the hybrid homomorphic encryption framework are used to perform a hybrid encryption operation on the preprocessed short message, including dynamically converting the fully homomorphic encryption algorithm and the partially homomorphic encryption algorithm to perform a layered encryption operation on the preprocessed short message.

[0026] On the basis of the above technical solution, preferably, the preprocessing in step S2 includes formatting and encoding conversion, and the post-processing in step S6 includes removing privacy noise and data format conversion.

[0027] On the basis of the above technical solutions, preferably, the partially homomorphic encryption algorithm includes at least one of the Paillier algorithm, the ElGamal algorithm and the RSA algorithm, and the fully homomorphic encryption algorithm includes one of the BGV algorithm and the BFV algorithm.

[0028] On the other hand, the present invention provides a data encryption system, which embeds the above-mentioned data encryption method.

[0029] On the other hand, the present invention provides a computer-readable storage medium, on which a data encryption program can be read by a computer, and the data encryption program implements the above-mentioned data encryption method when executed by a processor.

[0030] The data encryption method for Beidou short message transmission of the present invention has the following beneficial effects compared with the prior art:

[0031] (1) By setting up a hybrid homomorphic encryption framework, data can be dynamically converted from partially homomorphic encryption to fully homomorphic encryption or vice versa during encryption operations, aiming to combine the advantages of both, namely the high efficiency of PHE and the extensive computing power of FHE, to adapt to different computing needs and improve the overall performance and flexibility of the system. This conversion process is controlled by a secure key management system to ensure the security of the conversion and prevent data leakage. The key management system is responsible for handling all related key operations to ensure the security of data during the conversion process. The system can use the efficiency of PHE when performing simple operations and use the functions of FHE when complex operations are required, thereby optimizing computing performance and resource utilization to meet the needs of diverse applications;

[0032] (2) By setting sensitivity scores and privacy budgets based on differential privacy, the authenticity and availability of short message data can be maximized while protecting user privacy. This is crucial for scenarios that are sensitive to delays, have limited data volumes, and require high privacy protection, such as instant messaging and certain IoT applications, to ensure that data privacy is not violated while providing sufficient information for correct data analysis and decision support;

[0033] (3) Setting the sensitivity score: By analyzing the length, frequency, and length of sensitive information in short messages in real time, the noise level can be adaptively adjusted to optimize the security and effectiveness of the information. Specifically, for highly sensitive messages, noise can be increased to protect the information by adding random data fragments, slightly modifying the message content, or changing the sending frequency. For low-sensitivity messages, the amount of noise can be reduced to retain more useful information and reduce resource consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 The present invention is a schematic diagram of the steps of the data encryption method for Beidou short message transmission. DETAILED DESCRIPTION

[0036] The following will be combined with 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.

[0037] like Figure 1 As shown, the data encryption method for Beidou short message transmission of the present invention includes steps S1-S6.

[0038] Step S1: Setting a hybrid homomorphic encryption framework, wherein the hybrid homomorphic encryption framework includes a fully homomorphic encryption algorithm and a partially homomorphic encryption algorithm.

[0039] Combine fully homomorphic and partially homomorphic encryption methods to optimize computational efficiency and flexibility of encrypted data. In the field of data security, homomorphic encryption technology provides a new possibility for the processing of encrypted data, which allows operations on encrypted data without decryption. However, existing homomorphic encryption schemes have limitations in compatibility and flexibility, mainly because the computational efficiency of fully homomorphic encryption (FHE) is low, while the scenarios of partially homomorphic encryption (PHE), additive or multiplicative homomorphic encryption are limited. Therefore, fully homomorphic encryption algorithms and partially homomorphic encryption algorithms can be combined to combine the advantages of partially homomorphic encryption technology and fully homomorphic encryption technology to optimize computational efficiency, enhance compatibility and improve flexibility.

[0040] Accordingly, when setting up a hybrid homomorphic encryption framework, it is necessary to select specific fully homomorphic encryption algorithms and partially homomorphic encryption algorithms. The selection of an appropriate homomorphic encryption algorithm should be based on the characteristics of the data and the expected type of calculation. For data that mainly involves addition, additive homomorphic encryption such as the Paillier algorithm is suitable and is suitable for scenarios such as statistical analysis. When multiplication operations are involved, multiplicative homomorphic encryption such as the ElGamal algorithm can be used for product calculations. For data that requires multiple types of calculations or complex calculations, fully homomorphic encryption (such as BGV and BFV algorithms) can perform arbitrary calculations and is suitable for complex applications with high privacy requirements. Partially homomorphic encryption (such as RSA) is optional when the data processing requirements are simple and limited to one operation. For scenarios with low computational requirements and specific scenarios, PHE or algorithms with homomorphic properties of a single operation can be selected. For scenarios that require complex calculations or multiple types of calculations, FHE provides greater flexibility and powerful functions, but usually at the cost of higher computational costs. Taking into account the computational complexity, performance, and security requirements, the best matching homomorphic encryption scheme is selected to optimize computing performance and data security. It should be noted that the above-mentioned algorithms can all be set in the hybrid homomorphic encryption framework to generate encryption schemes based on the characteristics of short messages and each algorithm, and use the corresponding algorithms in the schemes for encryption operations.

[0041] Step S2: receiving the short message input by the sending end, and preprocessing the short message.

[0042] The system receives data from the source of Beidou short message transmission. Due to format and other reasons, the received data may need to be preprocessed, such as formatting, encoding conversion, etc., to adapt to subsequent encryption processing. After the data is preprocessed, subsequent hybrid homomorphic encryption can be performed.

[0043] Step S3: Perform hybrid encryption operation on the preprocessed short message through the fully homomorphic encryption algorithm and the partially homomorphic encryption algorithm in the hybrid homomorphic encryption framework, and add differential privacy noise during the encryption operation to obtain an intermediate file.

[0044] Data privacy protection is particularly important in short message transmission scenarios, such as instant messaging and IoT sensor data reporting. However, the amount of data in these scenarios is usually small, and the standard differential privacy noise addition strategy may be too rough, so that the added noise seriously undermines the practicality of the data.

[0045] In this embodiment, a fully homomorphic encryption algorithm and a partially homomorphic encryption algorithm are dynamically converted to perform layered encryption operations on the preprocessed short messages. Specifically, different levels of encryption strategies are applied for different types of data and computing requirements. The choice of partial homomorphic encryption (PHE) or fully homomorphic encryption (FHE) depends on the complexity and performance requirements of the calculation. PHE is suitable for simple calculations (such as single addition or multiplication). Taking additive homomorphic encryption as an example, let E represent the encryption function, D be the decryption function, and a and b be plaintext values, then: in, represents an operation on encrypted data, which corresponds to an addition operation on plaintext data. Its computational depth is shallow, and its advantage is high efficiency. FHE is applicable to more complex operations with arbitrary computational depth, and supports any type of computation, but at the cost of lower efficiency. For addition and multiplication, let E and D be the encryption and decryption functions respectively, and a and b be two plaintext values, then there is additive homomorphism: Multiplicative Homomorphism: Indicates a multiplication operation on encrypted data. In this implementation, a fixed "computation threshold" is not set to make a direct decision, but a comprehensive consideration is required based on the calculation type, depth, performance and security requirements.

[0046] Dynamically converting data from partially homomorphic encryption (PHE) to fully homomorphic encryption (FHE) or vice versa during the operation process aims to combine the advantages of both, namely the high efficiency of PHE and the extensive computing power of FHE, to adapt to different computing needs and improve the overall performance and flexibility of the system. This conversion process is controlled by a secure key management system to ensure the security of the conversion and prevent data leakage. The key management system is responsible for handling all related key operations to ensure the security of data during the conversion process. This method allows the system to take advantage of the efficiency of PHE when performing simple operations and use the functions of FHE when complex operations are required, thereby optimizing computing performance and resource utilization to meet the needs of diverse applications.

[0047] Before encryption, it is necessary to dynamically adjust the encryption strength and add differential privacy noise according to the sensitivity of the data and the privacy protection requirements to further protect data privacy, which specifically includes steps S31-S34.

[0048] Step S31: clustering multiple short messages and correlating them to form a message chain.

[0049] First, through data clustering and sequence pattern mining technology such as K-means and GSP (Generalized Sequential Pattern), the messages that constitute the information chain are identified and clustered. The K-means algorithm is used to cluster the messages. The clustering process can be expressed by the following simplified formula:

[0050]

[0051] Among them, C i is the i-th cluster, x i represents the message feature vector, μ i is cluster C i The center of n is the total number of messages.

[0052] When using the GSP algorithm to mine message sequence patterns, it is assumed that there is a sequence database D and a sequence s. The goal of sequence pattern mining is to find all sequences s that meet the minimum support min-sup, that is:

[0053]

[0054] These messages are analyzed and associated based on features such as timestamp, sender / receiver, and content keywords. Subsequently, deep natural language processing (NLP) techniques, including text classification, sentiment analysis, named entity recognition, and graph analysis technology, are used to identify key entities and sensitivity assessments in the messages.

[0055] Step S32: Calculate the sensitivity score of the short message according to the length of the short message, the length of the content sensitive information and the frequency of the short message in the message chain.

[0056] The strategy of differentiated noise distribution is designed based on semantic analysis of message content. It relies on advanced natural language processing (NLP) technology, including keyword recognition, context analysis, and named entity recognition (NER). These technologies work together to accurately identify and evaluate sensitive information in messages. In short messages, the sum of the lengths of all sensitive information is the length of content sensitive information. For example, personal identity information and geographic location, and then make detailed noise adjustments based on the sensitivity of the content. Messages containing geographic information confuse location information by adding noise to geocoding; messages containing personal identity information protect this data by replacing or blurring sensitive fields. The function S (content) can evaluate the sensitivity of the message content and return a value between 0 and 1, where 0 means insensitive and 1 means extremely sensitive. The function that determines the noise level N based on the sensitivity score can be expressed as: N = f (S (content)). The function f determines the noise level based on the sensitivity score. For example, for a simple linear model, f is defined as:

[0057] f(S)=N min +(N max -N min )·S

[0058] Among them, N max and N minRespectively represent the maximum noise and the minimum noise, Score max is the maximum sensitivity score. In this model, the higher the sensitivity score, the closer the added noise level is to N. max , and vice versa, for specific information types, such as geographic information and personal identity information, different sensitivity assessment functions S can be defined geo (content) and S id (content), and adjust the corresponding noise level respectively, for example, for geographic information: N geo =f(S geo (content)), for personally identifiable information N id =f(S id This method not only improves information security and prevents the leakage of sensitive information, but also optimizes data transmission efficiency and quality by preserving the usefulness and readability of the message. The strategy of differentiated noise distribution embodies an intelligent approach to seek a balance between ensuring data security and data utility, providing a higher level of protection for data transmission and processing.

[0059] For sensitivity scores, the length indicates the amount of information, the frequency reflects the interactive activity, and the sensitivity assessment of the content is based on the analysis of keywords or data patterns. Combining these features, a sensitivity score is calculated for each message through a weighted scoring system. For example, the sensitivity score Score is:

[0060] Score = w l ·L+w f ·F+w q Q

[0061] Among them, w1 is the length of the short message, L is the weight of the length, and w f is the frequency of short messages in the message chain, F is the weight of the frequency, and w q is the content-sensitive information length of the short message, and Q is the weight of the content-sensitive information length.

[0062] Each weight factor reflects the relative importance of this feature in the sensitivity score, that is, for each message, its sensitivity score is calculated based on its length, frequency and content-sensitive information length.

[0063] According to the sensitivity score, the noise adjustment amount NoiseLevel can be determined. For example, a simple function can be used to indicate that for a high-sensitivity message (high score), the noise amount is close to N max ; For low-sensitivity messages (low scores), the amount of noise is close to N min , the noise adjustment formula is:

[0064]

[0065] Score max It is the highest score and is used to normalize the sensitivity score to the range of 0 and 1.

[0066] Adaptive adjustment of noise amount: For highly sensitive messages, increase noise to protect information, which may be achieved by adding random data fragments, slightly modifying the message content, or changing the sending frequency; for low-sensitivity messages, reduce the amount of noise to retain more useful information and reduce resource consumption. To implement this strategy, it should be noted that the addition of noise should not interfere with the basic functions of the message, and the sensitivity assessment and noise adjustment should dynamically adapt to the changing security requirements to find a balance between protecting privacy and maintaining transmission efficiency.

[0067] Step S33: Based on differential privacy, the noise level of the short message is determined according to the sensitivity score, the privacy budget and the correlation between the short messages in the message chain.

[0068] In differential privacy, a privacy budget ∈ is set for each message, and the dynamic adjustment of the noise level can be expressed as:

[0069]

[0070] Where M(x) is the output after adding noise, ← is the privacy budget, q(x) is the original query function, Lap() is the Laplace noise function, and Δq is the sensitivity of the query.

[0071] By setting a privacy budget for each message and making precise noise adjustments at key points in the information chain, noise coordination among multiple messages is achieved.

[0072] The strategy of dynamically allocating the privacy budget for each short message takes into account the sensitivity of the message content and the correlation between messages to achieve a balance between privacy protection and information availability. The initial privacy budget is set according to the sensitivity of the message and the user's preference for privacy protection, reflecting the amount of resources available for privacy protection of the message. By analyzing the relationship between messages through association rule learning or machine learning technology, the privacy protection needs in the entire communication process can be identified to ensure overall privacy management for coherent data streams. Real-time tracking of the consumption of the privacy budget allows for quantitative management of the privacy impact of each data processing, ensuring timely adjustment of privacy protection measures. The amount of noise added to subsequent messages is dynamically adjusted according to the remaining amount of the privacy budget and the correlation between messages, which not only protects user privacy but also maintains the validity of data and the smoothness of communication.

[0073] In addition, in this embodiment, some special scores S can also be set. When the sensitivity score reaches the special score S, N(S)=α·S+β, where N(S) is the noise level after the sensitivity score reaches the special score S, and α and β are adjustment parameters, which are used to adjust the intensity and offset of the noise, respectively. That is to say, when the special score S is not reached, the noise calculation is performed according to the above formula based on differential privacy, and when the special score S is reached, the calculation is performed according to this formula.

[0074] Step S34: adding noise to the short message according to the noise level of the short message.

[0075] Finally, according to the obtained noise level, noise can be added to the short message during the encryption process. Through the above mechanism, this solution can protect the user's privacy while maximizing the authenticity and availability of the short message data. This is crucial for scenarios that are sensitive to delays, have limited data volume, and require high privacy protection, such as instant messaging and certain IoT applications, to ensure that the privacy of the data is not violated while providing sufficient information for correct data analysis and decision support.

[0076] Step S4: Generate a key based on the intermediate file and distribute it to the sending end and the receiving end.

[0077] Specifically, the system generates the required keys and uses the key management protocol to ensure the secure generation, distribution, storage, use, update and destruction of the keys. The key to the key management protocol is to ensure the secure generation, distribution, storage, use, update and destruction of the keys to enhance the security of the entire data processing process. Its distribution can adopt the latest encryption standards and technologies such as quantum key distribution (QKD) to resist future threats. This improves the credibility, flexibility and scalability of the system, aiming to provide users with solid data security and enhance their trust in the system. Ensure that the use, storage and update of keys meet the highest security standards throughout the data processing process.

[0078] Step S5: The sending end encrypts the short message using the key to obtain an encrypted file, and sends the encrypted file to the receiving end through the Beidou short message transmission system.

[0079] The generated key includes the noise to be added and the calculation method for encryption. The calculation method is to use the homomorphic encryption algorithm in the hybrid homomorphic encryption framework and the partial homomorphic encryption algorithm for combined encryption. Correspondingly, the decryption solution can also be obtained according to the key.

[0080] Step S6: The receiving end decrypts the encrypted file using the key and performs post-processing to obtain a short message.

[0081] After obtaining the key and encrypted file at the receiving end, they can be decrypted using the key, and post-processed such as removing privacy noise and converting data formats to ensure the availability and accuracy of the data.

[0082] In addition, for some special short message usage scenarios, such as federated learning, secure aggregation algorithms are used to process encrypted data from multiple participants. Including secure multi-party computing (SMC) protocol, lightweight encryption mechanism, threshold homomorphic encryption, model compression and sparsity technology, etc. The encrypted and securely processed data can be used for further data processing and analysis. Including data mining, machine learning model training, etc., to extract valuable information and make decisions. Finally, the processed data will be transmitted to the destination that needs this information, such as the Beidou emergency broadcast command center platform or other related systems.

[0083] Specifically, the SMC protocol is adopted to securely aggregate model updates from different participants, ensuring that no individual participant's data is leaked during the entire process.

[0084] Select efficient encryption algorithms suitable for short texts, such as elliptic curve cryptography (ECC) and stream ciphers. Optimizing key management is also necessary, which can be achieved through pre-shared keys and automatic key rotation mechanisms. Simplify the handshake protocol and message format to reduce protocol overhead, and use data compression technology to reduce the amount of data before encryption. Block cipher modes such as CTR or GCM can improve the performance of processing short messages. In terms of privacy protection, the combination of differential privacy technology can dynamically adjust the amount of noise added to the message when necessary. In addition, the use of secure hardware can accelerate the key generation and encryption process, while the use of lightweight communication protocols such as MQTT can help reduce the overhead of the protocol stack. Finally, the use of hash functions with low computational complexity but security can ensure the integrity of the message while maintaining overall computational efficiency. These strategies combined can ensure that security and privacy protection are taken into account while maximizing transmission efficiency.

[0085] Threshold homomorphic encryption allows multiple participants to collaboratively compute the aggregated encryption result without revealing their respective inputs. This provides an effective means of privacy protection in short message transmission scenarios. This threshold refers to the sum of the minimum number of key holders who can collaboratively compute the aggregated result.

[0086] In order to reduce the network burden in short message transmission without sacrificing model quality, the comprehensive use of model compression and sparsity technology is key. First, through model pruning, the weights with the lowest contribution are removed. Then, weight quantization converts floating-point numbers into low-order integers, which not only reduces storage requirements but also improves computational efficiency. Knowledge distillation technology allows a smaller model to learn the behavior of a large model while retaining performance. In addition, the introduction of sparsity constraints prompts the model to use fewer non-zero weights and reduces the number of independent parameters through parameter sharing. The sparse update strategy only uploads weights that have changed significantly, further reducing the amount of data transmitted. Finally, low-rank approximation technology can effectively compress the weight matrix. These measures combined not only reduce the size of model updates, but also ensure the retention of key information, making the model lightweight and robust.

[0087] The data encryption system of the present invention embeds the above-mentioned data encryption method.

[0088] In this system, a special operation library needs to be set up to identify the type of input data, analyze the type and complexity of the operation to be performed on the data, and evaluate the required computing performance, including processing speed, resource consumption and energy consumption. Then, based on the sensitivity and security requirements of the data, a homomorphic encryption algorithm that provides a sufficient level of security is selected, and the computing resources available to the current system are checked, including CPU, memory and storage, etc. At the same time, it is ensured that the selected algorithm is compatible with existing systems and applications, and the algorithm selection and parameter configuration are adjusted according to real-time feedback and system status. Historical performance data and machine learning technology can be used to predict the performance of different algorithms in the current environment and make a choice, and the user's specific preferences and strategies are considered in the process.

[0089] The computer-readable storage medium of the present invention can read a data encryption program on the computer-readable storage medium, and the data encryption program can implement the above-mentioned data encryption method when executed by a processor.

[0090] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A data encryption method for Beidou short message transmission, characterized in that: The following steps are involved: S1 sets a hybrid homomorphic encryption framework, which includes a fully homomorphic encryption algorithm and a partially homomorphic encryption algorithm; S2 receives the short message input by the sender and pre-processes the short message; S3 performs hybrid encryption operations on the preprocessed short messages through the fully homomorphic encryption algorithm and the partially homomorphic encryption algorithm in the hybrid homomorphic encryption framework, and adds differential privacy noise according to the privacy budget during the encryption operation to obtain an intermediate file. The adding of differential privacy noise includes the following sub-steps: S31 clusters multiple short messages and associates them with each other to form a message chain; S32 calculates the sensitivity score of the short message according to the length of the short message, the length of the content sensitive information and the frequency of the short message in the message chain. The sensitivity score Score is: Score=w l ·L+w f ·F+w q ·Q Among them, w l is the length of the short message, L is the weight of the length, w f is the frequency of short messages in the message chain, F is the weight of the frequency, and w q is the length of the content-sensitive information of the short message, and Q is the weight of the length of the content-sensitive information; S33 determines the noise level of the short message based on differential privacy, according to the sensitivity score, privacy budget and the correlation between short messages in the message chain. The noise level M(x) is: Where, ∈ is the privacy budget, q(x) is the original query function, Lap() is the Laplace noise function, and Δq is the sensitivity of the query; S34 adds noise to the short message according to the noise level of the short message; The step S3 further includes setting a special score S after step S32. When the sensitivity score reaches the special score S, N(S)=α·S+β, wherein N(S) is the noise level after the sensitivity score reaches the special score S, and α and β are adjustment parameters for adjusting the intensity and offset of the noise, respectively; S4 generates a key based on the intermediate file and distributes it to the sender and receiver; The S5 sender uses the key to encrypt the short message to obtain an encrypted file, and sends the encrypted file to the receiver through the Beidou short message transmission system; The S6 receiving end decrypts the encrypted file using the key and performs post-processing to obtain a short message.

2. The data encryption method for Beidou short message transmission as claimed in claim 1, characterized in that: In step S3, the hybrid encryption operation is performed on the preprocessed short message through the fully homomorphic encryption algorithm and the partially homomorphic encryption algorithm in the hybrid homomorphic encryption framework, including dynamically converting the fully homomorphic encryption algorithm and the partially homomorphic encryption algorithm to perform layered encryption operations on the preprocessed short message.

3. The data encryption method for Beidou short message transmission as claimed in claim 1, characterized in that: The preprocessing in step S2 includes formatting and encoding conversion, and the postprocessing in step S6 includes removing privacy noise and data format conversion.

4. The data encryption method for Beidou short message transmission as claimed in claim 1, characterized in that: The partially homomorphic encryption algorithm includes at least one of the Paillier algorithm, the ElGamal algorithm and the RSA algorithm, and the fully homomorphic encryption algorithm includes one of the BGV algorithm and the BFV algorithm.

5. A data encryption system, characterized in that: The data encryption method described in any one of claims 1 to 4 is embedded.

6. A computer-readable storage medium, characterized in that: The data encryption program on the computer-readable storage medium implements the data encryption method according to any one of claims 1 to 4 when the data encryption program is executed by the processor.

Images