Privacy protection method and system based on Internet of Things data encryption transmission

Abstract

The invention discloses a privacy protection method and system based on Internet of Things data encryption transmission, and the method comprises the following steps: a data collection step: collecting original data from at least one Internet of Things device, the Internet of Things device including but not limited to a sensor, an actuator, a smart home device or an industrial Internet of Things node, the original data comprises one or more of environment data, user behavior data and equipment state data, the acquisition frequency is dynamically adjusted according to equipment types and application scenes, the range is one time per second to one time per hour, and a tamper-proof mechanism is adopted in the acquisition process to ensure the data integrity; the core advantage of the method is that refined and quantitative evaluation of the data privacy level is realized by introducing an innovative privacy score calculation formula.

Description

Technical Field

[0001] This invention relates to encrypted data transmission in the Internet of Things (IoT), and more specifically to a privacy protection method and system based on encrypted data transmission in the IoT. Background Technology

[0002] With the rapid development of IoT technology, billions of smart devices have been deployed in various fields such as smart homes, industrial automation, and smart cities, continuously generating massive amounts of environmental data, device status data, and user behavior data. This data is aggregated and analyzed on cloud or edge servers to provide intelligent services. However, IoT data often contains a large amount of sensitive information, such as personal habits, industrial production line operating parameters, and the status of critical infrastructure, making privacy and security issues during transmission increasingly prominent.

[0003] Currently, there are several traditional technical solutions for encrypted data transmission in the Internet of Things (IoT). The most common is the use of standard encryption algorithms, such as Advanced Encryption Standard (AES) or Transport Layer Security (TLS) protocols, to provide end-to-end protection for the data channel. However, these traditional solutions have significant shortcomings in addressing the unique challenges of the IoT: First, they lack flexibility. Most solutions employ a "one-size-fits-all" static encryption strategy, configuring the same encryption strength for all data, ignoring the differences in the privacy sensitivities of the data itself. Using high-strength encryption for low-sensitivity temperature and humidity data results in unnecessary computational and energy consumption; while using weak encryption for highly sensitive human health data poses serious security risks. Second, they are inefficient in terms of resources. IoT terminal devices typically have strict limitations in computing power, storage space, and battery energy. Using consistently high-strength encryption algorithms (such as RSA-2048) will quickly deplete device power, affecting network lifespan and making large-scale deployment difficult in resource-constrained IoT environments. Third, they lack adaptability. Existing solutions cannot adjust security measures in real time according to dynamically changing data content, network environment, or security policies, lacking intelligent adaptive capabilities.

[0004] Furthermore, some improved solutions attempt to selectively encrypt data based on its simple type (such as text or images), but fail to deeply quantify the inherent privacy sensitivity of the data. Their judgment criteria are too coarse-grained and cannot meet the refined privacy protection needs. Therefore, there is an urgent need in this field for a method and system that can intelligently assess the privacy level of data and dynamically adjust encryption strategies accordingly, ensuring security while also taking into account the overall energy efficiency and operational efficiency of the IoT system. Summary of the Invention

[0005] The purpose of this invention is to provide a privacy protection method and system based on encrypted data transmission in the Internet of Things (IoT) to address the aforementioned shortcomings in the prior art.

[0006] To achieve the above object, the present invention provides the following technical solutions: A privacy protection method for encrypted transmission of Internet of Things data, comprising the following steps: Data collection step: Collecting raw data from at least one Internet of Things device, the Internet of Things device including but not limited to sensors, actuators, smart home devices or industrial Internet of Things nodes, the raw data including one or more of environmental data, user behavior data, and device status data, the collection frequency being dynamically adjusted according to the device type and application scenario, ranging from once per second to once per hour, and an anti-tampering mechanism being adopted during the collection process to ensure data integrity, for example, through digital signature or hash verification; Data preprocessing step: Preprocessing the collected raw data, including data cleaning, format standardization, outlier detection and data compression, where data cleaning removes redundant and noisy data, format standardization converts the data into a unified format such as JSON or XML, outlier detection uses statistical methods such as the Z-score algorithm to identify and process outliers, and data compression uses a lossless compression algorithm such as LZ77 to reduce the transmission load, and the preprocessed data is stored in a temporary buffer for subsequent processing; Privacy level evaluation step: Evaluating the privacy level of the preprocessed data, and determining the sensitivity of the data by calculating a privacy score Pscore. The privacy score Pscore is calculated using the following formula: Pscore = α×Σ{i = 1}^{n}wi×Ii + β×log(1 + C) where α and β are weight coefficients, with a value range of 0 to 1, and α + β = 1, n is the number of data attributes, wi is the weight of the i-th attribute, Ii is the sensitivity indicator of the i-th attribute, with a value of 0 or 1, and C is the data content complexity index, calculated based on information entropy; in this formula, α and β are dynamically optimized through a machine learning model to adapt to different application scenarios, and the value range of the privacy score Pscore is 0 to 10, and the higher the score, the stronger the privacy sensitivity; the calculated Pscore is used to adaptively adjust the encryption parameters; Adaptive encryption step: Dynamically selecting an encryption algorithm and key parameters according to the privacy score Pscore, including the following sub-steps: If Pscore ≤ 3, use a lightweight encryption algorithm such as XXTEA, with a key length of 128 bits; If 3 < Pscore ≤ 7, use a standard encryption algorithm such as AES-256, with a key length of 256 bits; If Pscore > 7, use a strengthened encryption algorithm such as RSA-2048 or elliptic curve-based encryption, with a key length not less than 2048 bits; During the encryption process, the session key K is generated by K=H(DID||T||S), where H is the SHA-256 hash function, DID is the device unique identifier, T is the current timestamp, and S is the pre-shared secret. The encryption operation is completed at the device or edge node to ensure that the data is encrypted at the source. Secure transmission steps: The encrypted data is transmitted to the cloud server or edge computing node via wireless or wired network. The transmission protocol adopts a secure version such as MQTT over TLS or CoAP with DTLS. Random delay and frequency hopping mechanisms are added to combat traffic analysis attacks. The transmission channel uses two-way authentication to prevent man-in-the-middle attacks. Data decryption steps: At the receiving end, the data is decrypted using the corresponding decryption algorithm. The decryption key is obtained through a secure key negotiation protocol such as Diffie-Hellman or a pre-set key store. After decryption, the data is verified for integrity to ensure that it has not been tampered with. Privacy protection implementation steps: Implement privacy protection strategies for decrypted data, including data anonymization, pseudonymization, or access control. Anonymization uses k-anonymity or differential privacy technology, and access control is based on a role-based access control model, allowing only authorized users to access the data.

[0007] Preferably, in the data acquisition step, the IoT device includes a low-power wide-area network (LPWAN) device that uses LoRaWAN or NB-IoT protocols for data acquisition. The acquisition frequency is adaptively adjusted based on the device's battery level and network conditions, specifically achieved through the following methods: The device has a built-in power management module that monitors the battery level in real time. When the battery level is below a threshold, the acquisition frequency is automatically reduced to once per minute or less. Simultaneously, network conditions are evaluated using a signal strength indicator. If the signal strength is below -100dBm, the device switches to power-saving mode. Timestamps and geographic location information are added to the acquired data to ensure data traceability. During the acquisition process, a redundant acquisition mechanism is employed, where multiple devices simultaneously acquire the same data. Errors are eliminated through a majority voting algorithm, improving data reliability. Furthermore, the data acquisition module integrates a hardware security module to generate a unique device identifier (DID). This identifier is generated based on the device's hardware fingerprint and a random number to prevent forgery. The acquired data is temporarily stored in an encrypted cache, the size of which is dynamically adjusted based on the device's memory, with a maximum of 1MB to avoid memory overflow.

[0008] Preferably, in the data preprocessing step, data cleaning employs an adaptive filtering algorithm based on machine learning, specifically including: using a pre-trained classification model to identify noisy data; the model is trained based on historical data, with the input being a data feature vector and the output being a noise probability; if the probability exceeds 0.7, it is marked as noise and removed; format standardization converts the data uniformly into AVRO binary format to reduce storage space and improve parsing efficiency; outlier detection uses the Isolation Forest algorithm, setting an outlier score threshold of 0.6 to automatically isolate outliers; data compression uses adaptive dictionary encoding, selecting the optimal compression algorithm based on the data type, such as using Huffman encoding for text data and JPEG-LS for image data; after preprocessing, metadata tags are added to the data, including data source, timestamp, and version number, to facilitate subsequent processing.

[0009] Preferably, in the privacy level assessment step, the weight coefficients α and β are dynamically optimized through an online learning algorithm, specifically using a reinforcement learning model with privacy protection effect and system overhead as the reward function, and the weights are updated every 24 hours; the data attribute sensitivity indicator Ii is predefined based on industry standards such as GDPR or NIST guidelines, for example, Ii is set to 1 for health data and 0 for environmental data; the content complexity index C is obtained by calculating the information entropy of the data, and the higher the entropy value, the greater the complexity; the calculation result of the privacy score Pscore is cached in a local database, and the cache validity period is set to 5 minutes to reduce redundant calculations; user preference input is added to the assessment process, allowing users to set privacy level preferences through the APP and automatically adjust the Pscore; these limitations make the privacy assessment more in line with actual needs and improve personalization and adaptability.

[0010] Preferably, in the adaptive encryption step, the lightweight encryption algorithm XXTEA is modified to enhance security, specifically by increasing the number of rounds to 64 and adding a random salt value to prevent pattern attacks; the standard encryption algorithm AES-256 uses the Galois / Counter mode to provide authentication encryption; the enhanced encryption algorithm is based on the elliptic curve digital signature algorithm, with curve parameters selected from secp256r1, and the key negotiation uses the ECDH protocol; in the generation of the session key K, the timestamp T has a precision of milliseconds, and the pre-shared secret S is rotated periodically with a rotation cycle of 7 days; the encryption operation is performed in a trusted execution environment to prevent side-channel attacks; this limitation improves encryption strength through algorithm optimization and security management, making it suitable for high-security scenarios such as finance or healthcare.

[0011] Preferably, in the secure transmission steps, the transmission protocol adopts the QUIC protocol to reduce latency, adds a forward security mechanism, and uses a temporary key for each session; the random latency range is set to 100-500 milliseconds, the frequency hopping mode is generated based on a pseudo-random sequence, and the sequence seed is derived from the device ID and timestamp; the transmission channel uses certificate pinning technology to fix the server certificate and prevent certificate forgery; data is transmitted in blocks, each block size not exceeding 1KB, and a sequence number and CRC check are added; when transmission fails, a retransmission mechanism is initiated, with a maximum of 3 retries.

[0012] Preferably, in the data decryption step, the decryption key is managed through an identity-based encryption scheme, the user's identity is bound to the public key, and biometric authentication is required during decryption; integrity verification uses the HMAC algorithm, and the hash function is SHA-3; after decryption, the data is first written into a security sandbox for malicious code scanning, and can only be output after the scan is passed; the decryption module supports parallel processing and utilizes multi-core CPUs to accelerate the decryption operation.

[0013] Preferably, in the privacy protection implementation steps, data anonymization employs differential privacy technology, adding Laplace noise with the noise level adjusted according to the Pscore to ensure that the privacy budget ε does not exceed 1.0; pseudonymization uses a reversible encrypted mapping table with entries updated periodically; access control is based on attribute-based encryption, with user attributes including role, time, and location, and policies are automatically executed by smart contracts; and implementation process audit logs are recorded, with the logs encrypted and stored on the blockchain to ensure immutability.

[0014] An IoT data encryption transmission privacy protection system includes: a data acquisition module integrated into IoT devices for collecting and initially processing data; a privacy assessment module deployed at edge nodes for calculating privacy scores; an encryption module located at the device or edge for performing adaptive encryption; a transmission module responsible for secure communication; a decryption module located on a cloud server for decryption and verification; and a privacy enhancement module for implementing anonymization and access control. All modules are connected via API interfaces, and the system supports elastic scaling, dynamically adjusting resources according to load. The system as a whole adopts a microservice architecture to improve maintainability and reliability.

[0015] In the above technical solutions, the present invention provides a privacy protection method and system based on IoT data encryption transmission. (1) The core advantage of the present invention is that by introducing an innovative privacy score (P_score) calculation formula, it realizes a refined and quantitative assessment of the data privacy level. This mechanism breaks the rigid mode of traditional "one-size-fits-all" encryption, enabling the system to adaptively select the most suitable encryption algorithm and key strength according to the actual sensitivity of the data. For low-sensitivity routine monitoring data, the system adopts lightweight encryption, which greatly reduces the computing load and energy consumption of terminal devices and extends the overall lifespan of the IoT network; for high-sensitivity key data, high-strength encryption is automatically enabled to ensure foolproof security. This "on-demand encryption" mode achieves the optimal dynamic balance between security protection and resource consumption in complex and diverse IoT application scenarios (such as from agricultural sensors to medical monitoring equipment), and solves the inherent contradiction between resource-constrained devices and high-strength security requirements; (2) The privacy assessment model described in the present invention has a high degree of intelligence and adaptability. First, the privacy score formula comprehensively considers data attributes (through weights w_i and sensitivity indicators I_i) and the content itself (through the complexity index C calculated by information entropy), making the evaluation dimensions more comprehensive and in-depth. Second, the key weight coefficients (α and β) in the formula can be dynamically optimized through online learning algorithms, enabling the model to continuously adapt to new data patterns, threat situations, or compliance requirements (such as GDPR). This design makes the privacy protection level no longer a pre-static setting, but an intelligent process that dynamically adjusts with changes in data content and context, providing a level of refined and personalized protection far exceeding traditional fixed strategies; (3) The method and system of this invention adopt a modular design, with each functional module (data acquisition, privacy assessment, adaptive encryption, etc.) coupled through a standard interface, exhibiting good flexibility and scalability. The privacy assessment model and encryption algorithm library can be updated and expanded independently without reconstructing the entire system. For example, when a new encryption standard emerges or new IoT protocols need to be adapted, only specific modules need to be upgraded. This architecture enables the present invention to easily meet the challenges of future IoT technology development and diverse application scenarios. Whether it is a small-scale smart home network or a large-scale smart city infrastructure, the system can be effectively deployed by adjusting parameters and strategies, demonstrating excellent universality and scalability. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0017] Figure 1 This is a schematic diagram illustrating the steps of a privacy protection method and system embodiment based on IoT-based encrypted data transmission according to the present invention. Detailed Implementation

[0018] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0019] like Figure 1 As shown in the figure, an embodiment of the present invention provides a privacy protection method for encrypted data transmission based on the Internet of Things, comprising the following steps: Data acquisition steps: Collect raw data from at least one IoT device, including but not limited to sensors, actuators, smart home devices or industrial IoT nodes. The raw data includes one or more of environmental data, user behavior data, and device status data. The acquisition frequency is dynamically adjusted according to the device type and application scenario, ranging from once per second to once per hour. Anti-tampering mechanisms are used during the acquisition process to ensure data integrity, such as through digital signatures or hash verification. Data preprocessing steps: The collected raw data is preprocessed, including data cleaning, format standardization, outlier detection, and data compression. Data cleaning removes redundant and noisy data, format standardization converts the data into a uniform format such as JSON or XML, outlier detection uses statistical methods such as the Z-score algorithm to identify and handle anomalies, and data compression uses lossless compression algorithms such as LZ77 to reduce transmission load. The preprocessed data is stored in a temporary buffer, awaiting subsequent processing. Privacy level assessment steps: Assess the privacy level of the preprocessed data by calculating a privacy score, Pscore, to determine the sensitivity of the data. The privacy score, Pscore, is calculated using the following formula: Pscore=α×Σ{i=1}^{n}wi×Ii+β×log(1+C) In this formula, α and β are weighting coefficients, ranging from 0 to 1, and α + β = 1; n is the number of data attributes; wi is the weight of the i-th attribute; Ii is the sensitivity indicator of the i-th attribute, taking a value of 0 or 1; and C is the data content complexity index, calculated based on information entropy. In this formula, α and β are dynamically optimized through a machine learning model to adapt to different application scenarios. The privacy score Pscore ranges from 0 to 10, with a higher score indicating stronger privacy sensitivity. The calculated Pscore is used to adaptively adjust encryption parameters. Adaptive encryption steps: Dynamically select the encryption algorithm and key parameters based on the privacy score Pscore, including the following sub-steps: If Pscore ≤ 3, use a lightweight encryption algorithm such as XXTEA with a key length of 128 bits; If 3 < Pscore ≤ 7, use a standard encryption algorithm such as AES-256 with a key length of 256 bits; If Pscore > 7, use a strengthened encryption algorithm such as RSA-2048 or elliptic curve-based encryption with a key length of no less than 2048 bits; During the encryption process, the session key K is generated by K = H(DID||T||S), where H is the SHA-256 hash function, DID is the device unique identifier, T is the current timestamp, and S is the pre-shared secret. The encryption operation is completed on the device side or the edge node to ensure data encryption at the source; Steps for secure transmission: Transmit the encrypted data to the cloud server or the edge computing node via a wireless or wired network. The transmission protocol adopts a secure version such as MQTT over TLS or CoAP with DTLS. Add a random delay and a frequency hopping mechanism to counter traffic analysis attacks. The transmission channel uses two-way authentication to prevent man-in-the-middle attacks; Steps for data decryption: At the receiving end, decrypt the data using the corresponding decryption algorithm. The decryption key is obtained through a secure key negotiation protocol such as Diffie-Hellman or a pre-set key library. Perform integrity verification on the decrypted data to ensure it has not been tampered with; Steps for privacy protection implementation: Implement privacy protection policies on the decrypted data, including data anonymization, pseudonymization, or access control. Anonymization adopts k-anonymity or differential privacy techniques. Access control is based on a role-based permission model, and only authorized users can access the data.

[0020] Preferably, in the data acquisition step, the IoT device includes a low-power wide-area network (LPWAN) device that uses LoRaWAN or NB-IoT protocols for data acquisition. The acquisition frequency is adaptively adjusted based on the device's battery level and network conditions, specifically achieved through the following methods: The device has a built-in power management module that monitors the battery level in real time. When the battery level is below a threshold, the acquisition frequency is automatically reduced to once per minute or less. Simultaneously, network conditions are evaluated using a signal strength indicator. If the signal strength is below -100dBm, the device switches to power-saving mode. Timestamps and geographic location information are added to the acquired data to ensure data traceability. During the acquisition process, a redundant acquisition mechanism is employed, where multiple devices simultaneously acquire the same data. Errors are eliminated through a majority voting algorithm, improving data reliability. Furthermore, the data acquisition module integrates a hardware security module to generate a unique device identifier (DID). This identifier is generated based on the device's hardware fingerprint and a random number to prevent forgery. The acquired data is temporarily stored in an encrypted cache, the size of which is dynamically adjusted based on the device's memory, with a maximum of 1MB to avoid memory overflow. This limitation further optimizes the reliability and energy efficiency of data acquisition, making it suitable for remote areas or mobile IoT scenarios.

[0021] Preferably, in the data preprocessing step, data cleaning employs an adaptive filtering algorithm based on machine learning, specifically including: using a pre-trained classification model to identify noisy data; the model is trained based on historical data, with the input being a data feature vector and the output being a noise probability; if the probability exceeds 0.7, it is marked as noise and removed; format standardization converts the data into AVRO binary format to reduce storage space and improve parsing efficiency; outlier detection uses the Isolation Forest algorithm, setting an outlier score threshold of 0.6 to automatically isolate outliers; data compression uses adaptive dictionary encoding, selecting the optimal compression algorithm based on the data type, such as using Huffman encoding for text data and JPEG-LS for image data; after preprocessing, metadata tags are added to the data, including data source, timestamp, and version number, for easy subsequent processing; this limitation improves preprocessing accuracy and speed and reduces transmission latency through intelligent algorithms, making it particularly suitable for high-throughput IoT applications such as video surveillance or real-time sensor networks.

[0022] Preferably, in the privacy level assessment step, the weight coefficients α and β are dynamically optimized through an online learning algorithm, specifically using a reinforcement learning model with privacy protection effect and system overhead as the reward function, and the weights are updated every 24 hours; the data attribute sensitivity indicator Ii is predefined based on industry standards such as GDPR or NIST guidelines, for example, Ii is set to 1 for health data and 0 for environmental data; the content complexity index C is obtained by calculating the information entropy of the data, and the higher the entropy value, the greater the complexity; the calculation result of the privacy score Pscore is cached in a local database, and the cache validity period is set to 5 minutes to reduce redundant calculations; user preference input is added to the assessment process, allowing users to set privacy level preferences through the APP and automatically adjust the Pscore; these limitations make the privacy assessment more in line with actual needs and improve personalization and adaptability.

[0023] Preferably, in the adaptive encryption step, the lightweight encryption algorithm XXTEA is modified to enhance security, specifically by increasing the number of rounds to 64 and adding a random salt value to prevent pattern attacks; the standard encryption algorithm AES-256 uses the Galois / Counter mode to provide authentication encryption; the enhanced encryption algorithm is based on the elliptic curve digital signature algorithm, with curve parameters selected from secp256r1, and the key negotiation uses the ECDH protocol; in the generation of the session key K, the timestamp T has a precision of milliseconds, and the pre-shared secret S is rotated periodically with a rotation cycle of 7 days; the encryption operation is performed in a trusted execution environment to prevent side-channel attacks; this limitation improves encryption strength through algorithm optimization and security management, making it suitable for high-security scenarios such as finance or healthcare.

[0024] Preferably, in the secure transmission steps, the transmission protocol adopts the QUIC protocol to reduce latency, adds a forward security mechanism, and uses a temporary key for each session; the random latency range is set to 100-500 milliseconds, the frequency hopping mode is generated based on a pseudo-random sequence, and the sequence seed is derived from the device ID and timestamp; the transmission channel uses certificate pinning technology to fix the server certificate and prevent certificate forgery; data is transmitted in blocks, each block size not exceeding 1KB, and a sequence number and CRC check are added; when transmission fails, a retransmission mechanism is initiated, with a maximum of 3 retries; this limitation optimizes transmission reliability and anti-attack capability, and is suitable for unstable network environments.

[0025] Preferably, in the data decryption step, the decryption key is managed through an identity-based encryption scheme, the user's identity is bound to the public key, and biometric authentication is required during decryption; integrity verification uses the HMAC algorithm, and the hash function is SHA-3; after decryption, the data is first written into a security sandbox for malicious code scanning, and can only be output after the scan is passed; the decryption module supports parallel processing and utilizes multi-core CPUs to accelerate the decryption operation; this limitation enhances decryption security and efficiency, and is suitable for multi-tenant cloud environments.

[0026] Preferably, in the privacy protection implementation steps, data anonymization employs differential privacy technology, adding Laplace noise with the noise level adjusted according to the Pscore to ensure the privacy budget ε does not exceed 1.0; pseudonymization uses a reversible encrypted mapping table with entries updated periodically; access control is based on attribute-based encryption, with user attributes including role, time, and location, and policies are automatically executed by smart contracts; the implementation process is audited and logged, with the logs encrypted and stored on the blockchain to ensure immutability; this limitation provides verifiable privacy protection, meeting compliance requirements such as HIPAA.

[0027] An IoT data encryption transmission privacy protection system includes: a data acquisition module integrated into IoT devices for collecting and initially processing data; a privacy assessment module deployed at edge nodes for calculating privacy scores; an encryption module located at the device or edge for performing adaptive encryption; a transmission module responsible for secure communication; a decryption module located on a cloud server for decryption and verification; and a privacy enhancement module for implementing anonymization and access control. All modules are connected via API interfaces, and the system supports elastic scaling, dynamically adjusting resources according to load. The system as a whole adopts a microservice architecture to improve maintainability and reliability. Example 1

[0028] This invention provides a privacy protection method for encrypted data transmission based on the Internet of Things, comprising the following steps: Data acquisition steps: Collect raw data from at least one IoT device, including but not limited to sensors, actuators, smart home devices or industrial IoT nodes. The raw data includes one or more of environmental data, user behavior data, and device status data. The acquisition frequency is dynamically adjusted according to the device type and application scenario, ranging from once per second to once per hour. Anti-tampering mechanisms are used during the acquisition process to ensure data integrity, such as through digital signatures or hash verification. Data preprocessing steps: The collected raw data is preprocessed, including data cleaning, format standardization, outlier detection, and data compression. Data cleaning removes redundant and noisy data, format standardization converts the data into a uniform format such as JSON or XML, outlier detection uses statistical methods such as the Z-score algorithm to identify and handle anomalies, and data compression uses lossless compression algorithms such as LZ77 to reduce transmission load. The preprocessed data is stored in a temporary buffer, awaiting subsequent processing. Privacy level assessment steps: Assess the privacy level of the preprocessed data by calculating a privacy score, Pscore, to determine the sensitivity of the data. The privacy score, Pscore, is calculated using the following formula: Pscore=α×Σ{i=1}^{n}wi×Ii+β×log(1+C) Among them, α and β are weight coefficients, with a value range of 0 to 1, and α + β = 1, n is the number of data attributes, wi is the weight of the i-th attribute, Ii is the sensitivity indicator of the i-th attribute, taking values of 0 or 1, C is the data content complexity index, calculated based on information entropy; in this formula, α and β are dynamically optimized through a machine learning model to adapt to different application scenarios, and the privacy score Pscore has a value range of 0 to 10. The higher the score, the stronger the privacy sensitivity; the calculated Pscore is used to adaptively adjust the encryption parameters; Adaptive encryption steps: Dynamically select the encryption algorithm and key parameters according to the privacy score Pscore, including the following sub-steps: If Pscore ≤ 3, use a lightweight encryption algorithm such as XXTEA with a key length of 128 bits; If 3 < Pscore ≤ 7, use a standard encryption algorithm such as AES-256 with a key length of 256 bits; If Pscore > 7, use a strengthened encryption algorithm such as RSA-2048 or elliptic curve-based encryption with a key length not less than 2048 bits; During the encryption process, the session key K is generated by K = H(DID||T||S), where H is the SHA-256 hash function, DID is the device unique identifier, T is the current timestamp, and S is the pre-shared secret. The encryption operation is completed on the device side or the edge node to ensure data encryption at the source; Secure transmission steps: Transmit the encrypted data to the cloud server or edge computing node through a wireless or wired network. The transmission protocol uses a secure version such as MQTT over TLS or CoAP with DTLS, adding a random delay and frequency hopping mechanism to counter traffic analysis attacks. The transmission channel uses two-way authentication to prevent man-in-the-middle attacks; Data decryption steps: At the receiving end, use the corresponding decryption algorithm to decrypt the data. The decryption key is obtained through a secure key negotiation protocol such as Diffie-Hellman or a pre-set key library. The integrity of the decrypted data is verified to ensure that it has not been tampered with; Privacy protection implementation steps: Implement privacy protection policies on the decrypted data, including data anonymization, pseudonymization, or access control. Anonymization uses k-anonymity or differential privacy techniques, and access control is based on a role-based permission model, and only authorized users can access the data. Example 2

[0029] An IoT data encryption transmission privacy protection system includes: a data acquisition module integrated into IoT devices for collecting and initially processing data; a privacy assessment module deployed at edge nodes for calculating privacy scores; an encryption module located at the device or edge for performing adaptive encryption; a transmission module responsible for secure communication; a decryption module located on a cloud server for decryption and verification; and a privacy enhancement module for implementing anonymization and access control. All modules are connected via API interfaces, and the system supports elastic scaling, dynamically adjusting resources according to load. The system as a whole adopts a microservice architecture to improve maintainability and reliability.

[0030] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A privacy protection method for encrypted data transmission based on the Internet of Things, characterized in that, It includes the following steps: Data collection step: Collect raw data from at least one Internet of Things device, which includes but is not limited to sensors, actuators, smart home devices or industrial Internet of Things nodes. The raw data includes one or more of environmental data, user behavior data, and device status data. The collection frequency is dynamically adjusted according to the device type and application scenario, ranging from once per second to once per hour. A tamper-proof mechanism is adopted during the collection process to ensure data integrity; Data preprocessing step: Preprocess the collected raw data, including data cleaning, format standardization, outlier detection and data compression. Among them, data cleaning removes redundant and noisy data, format standardization converts the data into a unified format such as JSON or XML, outlier detection uses statistical methods such as the Z-score algorithm to identify and process outliers, and data compression uses a lossless compression algorithm such as LZ77 to reduce the transmission load. The preprocessed data is stored in a temporary buffer and waits for subsequent processing; Privacy level assessment step: Evaluate the privacy level of the preprocessed data, and determine the sensitivity of the data by calculating the privacy score Pscore. The privacy score Pscore is calculated using the following formula: Pscore = α×Σ{i = 1}^{n}wi×Ii + β×log(1 + C) where, α and β are weight coefficients, with a value range of 0 to 1, and α + β = 1, n is the number of data attributes, wi is the weight of the i-th attribute, Ii is the sensitivity indicator of the i-th attribute, with a value of 0 or 1, and C is the data content complexity index, calculated based on information entropy; in this formula, α and β are dynamically optimized through a machine learning model to adapt to different application scenarios. The value range of the privacy score Pscore is 0 to 10, and the higher the score, the stronger the privacy sensitivity; the calculated Pscore is used to adaptively adjust the encryption parameters; Adaptive encryption step: Dynamically select the encryption algorithm and key parameters according to the privacy score Pscore, including the following sub-steps: If Pscore ≤ 3, use a lightweight encryption algorithm such as XXTEA with a key length of 128 bits; If 3 < Pscore ≤ 7, use a standard encryption algorithm such as AES-256 with a key length of 256 bits; If Pscore > 7, use a strengthened encryption algorithm such as RSA-2048 or elliptic curve-based encryption with a key length not less than 2048 bits; During the encryption process, the session key K is generated by K = H(DID||T||S), where H is the SHA-256 hash function, DID is the device unique identifier, T is the current timestamp, and S is the pre-shared secret. The encryption operation is completed on the device side or the edge node to ensure data encryption at the source; Secure transmission step: Transmit the encrypted data to the cloud server or edge computing node through a wireless or wired network. The transmission protocol uses a secure version such as MQTT over TLS or CoAP with DTLS. Add a random delay and frequency hopping mechanism to counter traffic analysis attacks. The transmission channel uses two-way authentication to prevent man-in-the-middle attacks; Data decryption steps: At the receiving end, the data is decrypted using the corresponding decryption algorithm. The decryption key is obtained through a secure key negotiation protocol such as Diffie-Hellman or a pre-set key store. After decryption, the data is verified for integrity to ensure that it has not been tampered with. Privacy protection implementation steps: Implement privacy protection strategies for decrypted data, including data anonymization, pseudonymization, or access control. Anonymization uses k-anonymity or differential privacy technology, and access control is based on a role-based access control model, allowing only authorized users to access the data.

2. The privacy protection method for encrypted data transmission based on the Internet of Things according to claim 1, characterized in that, In the data acquisition step, the IoT devices include low-power wide-area network (LPWAN) devices that use LoRaWAN or NB-IoT protocols for data acquisition. The acquisition frequency is adaptively adjusted based on the device's battery level and network conditions, specifically achieved through the following methods: The device has a built-in power management module that monitors the battery level in real time. When the battery level is below a threshold, the acquisition frequency is automatically reduced to once per minute or less. Simultaneously, network conditions are assessed using a signal strength indicator. If the signal strength is below -100dBm, the device switches to power-saving mode. Timestamps and geographic location information are added to the acquired data to ensure data traceability. During the acquisition process, a redundant acquisition mechanism is employed, where multiple devices simultaneously acquire the same data. Errors are eliminated through a majority voting algorithm, improving data reliability. Furthermore, the data acquisition module integrates a hardware security module to generate a unique device identifier (DID). This identifier is generated based on the device's hardware fingerprint and a random number to prevent forgery. The acquired data is temporarily stored in an encrypted cache, the size of which is dynamically adjusted based on the device's memory, with a maximum of 1MB to avoid memory overflow.

3. The privacy protection method for encrypted data transmission based on the Internet of Things according to claim 2, characterized in that, In the data preprocessing steps, data cleaning employs an adaptive filtering algorithm based on machine learning, specifically including: using a pre-trained classification model to identify noisy data. The model is trained based on historical data, with the input being a data feature vector and the output being a noise probability. If the probability exceeds 0.7, it is marked as noise and removed; format standardization converts the data into AVRO binary format to reduce storage space and improve parsing efficiency; outlier detection uses the Isolation Forest algorithm, setting an outlier score threshold of 0.6 to automatically isolate outliers; and data compression uses adaptive dictionary encoding, selecting the optimal compression algorithm based on the data type.

4. The privacy protection method for encrypted data transmission based on the Internet of Things according to claim 3, characterized in that, In the privacy level assessment step, the weight coefficients α and β are dynamically optimized through an online learning algorithm, specifically using a reinforcement learning model with privacy protection effect and system overhead as the reward function, and the weights are updated every 24 hours; the data attribute sensitivity indicator Ii is predefined based on industry standards such as GDPR or NIST guidelines.

5. A privacy protection method for encrypted data transmission based on the Internet of Things according to claim 1, characterized in that, In the adaptive encryption steps, a modified version of the lightweight encryption algorithm XXTEA is used to enhance security. Specifically, this includes increasing the number of rounds to 64 and adding a random salt value to prevent pattern attacks. The standard encryption algorithm AES-256 uses the Galois / Counter mode to provide authentication encryption. The enhanced encryption algorithm is based on the elliptic curve digital signature algorithm, with curve parameters selected from secp256r1, and the key negotiation uses the ECDH protocol. In the generation of the session key K, the timestamp T has a precision of milliseconds, and the pre-shared secret S is rotated periodically with a rotation cycle of 7 days. The encryption operation is performed in a trusted execution environment to prevent side-channel attacks.

6. The privacy protection method for encrypted data transmission based on the Internet of Things according to claim 1, characterized in that, In the secure transmission steps, the transmission protocol adopts the QUIC protocol to reduce latency, adds a forward security mechanism, and uses a temporary key for each session; the random latency range is set to 100-500 milliseconds, the frequency hopping mode is generated based on a pseudo-random sequence, and the sequence seed is derived from the device ID and timestamp; the transmission channel uses certificate pinning technology to fix the server certificate and prevent certificate forgery. Data is transmitted in blocks, each no larger than 1KB, with a sequence number and CRC checksum added. In case of transmission failure, a retransmission mechanism is initiated, with a maximum of 3 retries.

7. A privacy protection method for encrypted data transmission based on the Internet of Things according to claim 5, characterized in that, In the data decryption step, the decryption key is managed through an identity-based encryption scheme, with the user's identity bound to the public key, and biometric authentication is required during decryption; integrity verification uses the HMAC algorithm, and the hash function is SHA-3; after decryption, the data is first written into a security sandbox for malicious code scanning, and can only be output after the scan is successful; the decryption module supports parallel processing, utilizing multi-core CPUs to accelerate the decryption operation.

8. A privacy protection method for encrypted data transmission based on the Internet of Things according to claim 7, characterized in that, In the privacy protection implementation steps, data anonymization uses differential privacy technology, adding Laplace noise with the noise scale adjusted according to Pscore to ensure that the privacy budget ε does not exceed 1.0; pseudonymization uses a reversible encrypted mapping table with entries updated periodically; access control is based on attribute-based encryption, with user attributes including role, time, and location, and the policy is automatically executed by the smart contract. The implementation process is audited and logged, and the logs are encrypted and stored on the blockchain to ensure they cannot be tampered with.

9. A privacy protection system for encrypted data transmission in the Internet of Things (IoT) implementing the method as described in any one of claims 1 to 8, characterized in that, include: The data acquisition module, integrated into IoT devices, is used to collect and initially process data; the privacy assessment module, deployed at edge nodes, calculates privacy scores. An encryption module, located on the device or edge, performs adaptive encryption; The transmission module is responsible for secure communication; The decryption module, located on the cloud server, performs decryption and verification. The privacy enhancement module implements anonymization and access control; Each module is connected via API interfaces, and the system supports elastic expansion, dynamically adjusting resources according to load. The system as a whole adopts a microservice architecture, which improves maintainability and reliability.

Images