Local data updating method and system in local database

Abstract

The invention discloses a local data updating method and system in a local database, relates to the technical field of data management and security, and realizes user identity authentication by acquiring electrocardiogram and environment data through an edge computing node to generate a dynamic permission token. And then, the system verifies and authorizes a database updating request, and performs data fusion analysis by adopting federated learning and a deep neural network to ensure the security and accuracy of data updating. In the data updating process, the system uses chaotic encryption and a trusted execution environment to guarantee the confidentiality and integrity of data. And after updating, broadcasting an operation record through a block chain network, and performing auditing and rollback control by using zero-knowledge proof and a smart contract. And finally, ensuring the integrity and non-tampering property of data recovery by adopting a hash chain technology. Through multi-technology fusion, the security, reliability and efficiency of the data updating process are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data management and security, and in particular to a local database partial data updating method and system. BACKGROUND

[0002] With the development of Internet of Things and big data technology, the local database partial data updating method is also evolving from the early centralized updating to the current distributed and intelligent updating strategy, aiming to improve the efficiency and security of data updating.

[0003] The existing related technologies mainly have the following deficiencies: on the one hand, the traditional updating method mostly relies on a single authentication mechanism, lacks comprehensive authentication of user identity and environmental information, resulting in insufficient security, especially in the face of complex Internet of Things environment, which is vulnerable to malicious attacks; on the other hand, the existing technology lacks effective checking of data integrity and consistency in the data updating process, and cannot ensure the matching of updated data and environmental information, which is prone to data inconsistency or error updating; moreover, the audit and rollback mechanism after data updating is mostly complex and inefficient, which cannot quickly restore the original data under the premise of ensuring data security, especially in the face of large-scale data updating, which is more prominent. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides a local database partial data updating method to solve the problems of insufficient security, data updating errors and efficient and secure data updating and rollback in the existing local database partial data updating method.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] The present application provides a local database partial data updating method, characterized by comprising the following steps:

[0008] S1, when the system starts, the edge computing node collects electrocardiogram and environmental temperature and humidity data, generates a multi-dimensional authentication tuple and transmits it to the authority management system for verification, generates a dynamic authority token and stores it, completes user identity authentication and authority initialization;

[0009] S2, when a database updating request is detected, the authority management system verifies the validity of the dynamic authority token, and after verification, the edge computing node broadcasts a data checking request to the surrounding Internet of Things devices, and the surrounding devices extract environmental feature vectors and gather them to the edge node to form a checking matrix;

[0010] S3, the edge processor utilizes federated learning technology to analyze the check matrix, generates a check report, triggers the update of the authorization level, and starts the physiological signal gating mechanism;

[0011] S4, the user authenticates through the physiological signal, compares the authentication data with the dynamic permission token, and verifies the environmental feature fingerprint at the same time, after the matching is passed, the edge processor allows the execution of the data update operation, the data is stored in the form of chaotic encryption during the update process and the replacement and structure adjustment are completed by using the trusted execution environment, the new data multi-dimensional feature fingerprint is generated and the data state record in the blockchain account is updated;

[0012] S5, the edge computing node broadcasts the update operation record through the blockchain network, the distributed audit node verifies the compliance and data consistency of the update operation according to the smart contract rules, the validity of the update operation is verified by using the zero-knowledge proof technology during the audit process, if the abnormal or illegal operation is found, the system immediately starts the state rollback mechanism;

[0013] S6, read the historical data state record in the blockchain account, restore the original data by using the trusted execution environment, ensure the integrity and non-tamperability of the data recovery, the audit result is stored in the blockchain audit chain by using the hash chain technology, and the non-tamperable update operation record is formed.

[0014] As a preferred scheme of the local database partial data update method, in step S1, the generation of the multi-dimensional authentication tuple includes wavelet transform processing of the electrocardiogram signal to extract the feature vector, and the environmental data is processed by the Kalman filter algorithm to remove noise and smooth, the processed physiological and environmental feature vectors are combined into a multi-dimensional authentication tuple, and the generation process is carried out in the trusted execution environment of the edge computing node to enhance the security and integrity of the data.

[0015] As a preferred scheme of the local database partial data update method, in step S2, when the edge computing node broadcasts the data check request to the surrounding Internet of Things devices, a priority sorting mechanism based on device location and function is adopted, so that the surrounding devices respond to the request and extract the environmental feature vector according to the priority order, and the distributed hash table technology is used to optimize the convergence path of the feature vector, so as to improve the efficiency and response speed of the data check.

[0016] As a preferred scheme of the local data updating method in the local database according to the present application, wherein: in step S3, when the edge processor performs fusion analysis by using the federated learning technology, a deep neural network architecture is adopted, a multi-layer perceptron model is constructed to perform feature extraction and fusion analysis on the check matrix, a check report containing data credibility score and environmental consistency is generated, and different levels of update authorization are automatically triggered by the smart contract according to the preset credibility threshold, so as to realize fine-grained permission control.

[0017] As a preferred scheme of the local data updating method in the local database according to the present application, wherein: in step S4, when the user performs physiological signal authentication, a multi-modal biometric feature fusion technology is adopted to combine electrocardiogram waveform features and skin electrical response patterns for comprehensive authentication, a chaotic encryption algorithm is used to encrypt the authentication data, and a dynamic permission token is used for secure comparison, and a non-volatile storage technology based on a memristor array is used in the data updating operation to ensure the reliability and persistence of data updating.

[0018] As a preferred scheme of the local data updating method in the local database according to the present application, wherein: in step S5, when the edge computing node broadcasts the update operation record through the blockchain network, a zero-knowledge proof protocol is used to verify the validity of the data, the preset rules in the smart contract of the blockchain are used to automatically judge the compliance and data consistency of the update operation, and if an abnormality is found, the state rollback mechanism is triggered through the smart contract, and the original data segments required for rollback are redundantly stored in multiple nodes by using distributed storage technology, so as to improve the reliability and efficiency of data recovery.

[0019] As a preferred scheme of the local data updating method in the local database according to the present application, wherein: in step S6, when the system starts the state rollback mechanism, a data recovery instruction is generated by using the historical data state record in the blockchain ledger, the rollback engine in the trusted execution environment is used to recover the original data step by step according to the instruction, the recovery process is verified for integrity by using the hash chain technology, the integrity and non-tamperability of data recovery are ensured, the audit result is stored in the blockchain audit chain to form an unalterable update operation record, and the entire data updating process is fully traced and monitored.

[0020] The beneficial effects of this invention are as follows: By collecting user electrocardiogram and environmental temperature and humidity data through edge computing nodes, multi-dimensional authentication tuples are generated and dynamic permission tokens are generated using deep learning feature similarity measurement and spatiotemporal consistency verification of environmental data. This significantly improves the security and accuracy of user identity authentication. Simultaneously, by broadcasting data verification requests to surrounding IoT devices and using federated learning technology to fuse and analyze the verification matrix, a verification report is generated to trigger an update of the authorization level, effectively ensuring the accuracy and environmental adaptability of data updates. During the data update phase, chaotic encryption and a trusted execution environment are used to complete data replacement and structural adjustments, and the data state record is updated using a blockchain ledger, further enhancing the confidentiality and integrity of the data update process. Furthermore, by broadcasting update operation records through the blockchain network and using zero-knowledge proof technology to verify the validity of update operations, efficient and secure distributed auditing is achieved. If an update anomaly or violation is detected, the system immediately initiates a state rollback mechanism, reading historical data state records from the blockchain ledger and using the trusted execution environment to restore the original data, ensuring the integrity and immutability of the data recovery. Finally, the audit results are stored in the blockchain audit chain using hash chain technology, forming an immutable record of update operations, enabling comprehensive traceability and monitoring of the entire data update process. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart of the local data update method in the local database in Example 1. Detailed Implementation

[0023] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0024] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0025] Second, the "one embodiment" or "an embodiment" referred to herein means a specific feature, structure, characteristic, or combination of features and / or characteristics described herein that can be included in at least one implementation of the present application. The various appearances of "in one embodiment" or "an embodiment" in the specification do not all refer to the same embodiment, although they can.

[0026] Embodiment 1, refer to Figure 1 For the first embodiment of the present application, the embodiment provides a local data update method in a local database, characterized in that it comprises the following steps:

[0027] S1, when the system starts, the edge computing node collects electrocardiogram and environmental temperature and humidity data of the user, generates a multi-dimensional authentication tuple and transmits it to the authority management system for verification, generates a dynamic authority token and stores it, completes user identity authentication and authority initialization;

[0028] S2, when a database update request is detected, the authority management system verifies the validity of the dynamic authority token, and after verification, the edge computing node broadcasts a data verification request to the surrounding Internet of Things devices, and the surrounding devices extract environmental feature vectors and converge to the edge node to form a verification matrix;

[0029] S3, the edge processor uses federated learning technology to analyze and fuse the verification matrix, generates a verification report, triggers the update authorization level, and starts the physiological signal gating mechanism;

[0030] S4, the user authenticates through the physiological signal, the authentication data is compared with the dynamic authority token, and the environmental feature fingerprint is verified at the same time, and after matching, the edge processor allows the execution of the data update operation, the data is stored in the form of chaotic encryption during the update process and is replaced and structurally adjusted using the trusted execution environment, and a new data multi-dimensional feature fingerprint is generated and the data state record in the blockchain account is updated;

[0031] S5, the edge computing node broadcasts the update operation record through the blockchain network, and the distributed audit node verifies the compliance and data consistency of the update operation according to the smart contract rules, and uses zero-knowledge proof technology to verify the validity of the update operation during the audit process. If abnormal or illegal operation is found, the system immediately starts the state rollback mechanism;

[0032] S6, read the historical data state record in the blockchain account, restore the original data using the trusted execution environment, ensure the integrity and non-tamperability of the data recovery, and store the audit result in the blockchain audit chain using the hash chain technology, forming an unalterable update operation record.

[0033] It should be noted that step 1:

[0034] When the system starts, the edge computing node simultaneously collects electrocardiogram data of the user and temperature and humidity data of the surrounding environment. The electrocardiogram data is obtained by a flexible biosensor at high resolution, while the environmental temperature and humidity data is provided by an Internet of Things environmental sensor. The edge processor performs wavelet transform processing on the electrocardiogram signal to extract feature vectors; the environmental data is processed by a Kalman filter algorithm to remove noise and smooth. The processed physiological and environmental feature vectors are combined into a multi-dimensional authentication tuple. The tuple is transmitted to the permission management system through a secure channel and compared with the pre-registered user feature template for verification. The comparison algorithm is based on deep learning feature similarity measurement, combined with spatial and temporal consistency verification of environmental data, and finally generates a dynamic permission token containing biological feature hash value and environmental feature fingerprint. The token is stored in the local security module and the blockchain account book, completing user identity authentication and permission initialization.

[0035] Through the generation and verification of the multi-dimensional authentication tuple in step 1, dual identity authentication based on biological features and environmental features is realized. This multi-modal authentication method significantly improves the security and accuracy of user identity authentication, effectively preventing the risk of single authentication factor being cracked. At the same time, combined with deep learning feature similarity measurement and spatial and temporal consistency verification, it can adapt to user authentication needs in different environments, ensuring that only legitimate users can initiate data update operations, thereby providing a reliable identity basis for subsequent data update processes, enhancing the security and credibility of the entire system.

[0036] Step 2:

[0037] When the system detects a database update request, the permission management system first verifies the validity of the dynamic permission token. After verification, the edge computing node broadcasts a data verification request to the surrounding Internet of Things devices. Peripheral devices extract environmental feature vectors related to the update data according to their own functions and locations. For example, temperature sensors extract current environmental temperature features, pressure sensors extract pressure features, image acquisition devices extract visual features, etc. These feature vectors are aggregated into the edge node through a secure multi-hop communication protocol to form a verification matrix containing spatial and temporal context information, providing basic data support for subsequent data verification and fusion analysis.

[0038] Through the broadcast data verification request and feature vector aggregation in step 2, multi-dimensional perception and data collection of the data update environment are realized. This distributed data collection method can fully utilize the advantages of Internet of Things devices to obtain more comprehensive and accurate environmental information, providing a rich data foundation for subsequent data verification and fusion analysis. At the same time, the use of secure multi-hop communication protocol ensures the security and reliability of data transmission, preventing data from being tampered with or stolen during transmission, thereby improving the accuracy and reliability of data verification and providing strong data support and environmental protection for subsequent data update operations.

[0039] Step 3:

[0040] After receiving the check matrix aggregated by the peripheral devices, the edge processor performs fusion analysis on the check matrix using federated learning techniques. Specifically, a deep neural network architecture is adopted to construct a multi-layer perceptron model for feature extraction and fusion analysis in the check matrix, generating a check report containing data credibility scores and environmental consistency. The report automatically triggers different levels of update authorization based on the preset credibility threshold through the smart contract of the blockchain. At the same time, the physiological signal gating mechanism is started, preparing for subsequent user physiological signal authentication.

[0041] Through the federated learning fusion analysis in step 3 and the smart contract triggered authorization, intelligent evaluation and fine-grained permission control of data update requests are achieved. Federated learning techniques fully utilize distributed data resources while protecting the privacy of each Internet of Things device, improving the accuracy and reliability of data verification. Deep neural network models can deeply mine data features to generate more valuable check reports. The automatic triggering mechanism of the smart contract flexibly adjusts the update authorization level based on data credibility, achieving automated, data quality-based permission management, improving the intelligence level of the system and the security of data updates.

[0042] Step 4:

[0043] The user needs to pass the physiological signal authentication again, and the edge computing node collects real-time electrocardiogram waveforms and skin electricity reaction data. The physiological feature data generated during the authentication process is compared with the dynamic permission token generated in step 1 through a secure hash algorithm, while verifying the environmental feature fingerprint. After matching, the edge processor allows the actual data update operation to be performed. During the data update process, data is stored in a temporary secure storage area in a chaotic encryption form, and the trusted execution environment is used to complete data item replacement and structure adjustment. After the update is completed, the system automatically generates multi-dimensional feature fingerprints for the new data and updates the data state record in the blockchain ledger.

[0044] Through the physiological signal gating update execution mechanism in step 4, high security, reliability, and non-repudiation of data update operations are achieved. The second physiological signal authentication further confirms the legitimacy of the user's identity, preventing identity theft and illegal operations. The combination of chaotic encryption and trusted execution environment ensures the confidentiality and integrity of data during the update process, preventing data leakage and tampering. The generation of multi-dimensional feature fingerprints provides accurate data identification for subsequent data auditing and rollback, while the update of the blockchain ledger ensures the transparency and non-tamperability of the data state, making the data update operation highly credible and traceable.

[0045] Step 5:

[0046] The edge computing node broadcasts the update operation record through the blockchain network, including the update time, update content, data characteristics before and after the update, and other information. After receiving the broadcast record, the distributed audit node verifies the compliance and data consistency of the update operation according to the rules in the smart contract. During the audit process, zero-knowledge proof technology is used to verify the validity of the update operation without exposing sensitive data. If abnormal updates or illegal operations are found, the system immediately starts the state rollback mechanism, reads the historical data state record in the blockchain ledger, and restores the original data using the trusted execution environment.

[0047] Through the blockchain broadcast and zero-knowledge proof audit of step 5, the transparency, compliance, and efficiency of data update operations are achieved. The broadcast mechanism of the blockchain network ensures the openness and tamper resistance of the update operation record, so that all related nodes can obtain update information in a timely manner. The zero-knowledge proof technology effectively verifies the validity and compliance of the update operation while protecting data privacy, avoiding the need to expose sensitive data in traditional audit methods. The timely activation of the state rollback mechanism can quickly restore data to a normal state, reducing the risk of abnormal updates and improving the stability and reliability of the system.

[0048] Step 6:

[0049] When the system starts the state rollback mechanism, it reads the historical data state record in the blockchain ledger and generates detailed data recovery instructions based on the record. Using the rollback engine in the trusted execution environment, the original data is gradually restored according to the instructions. During the recovery process, the hash chain technology is used to verify the integrity of the recovered data, ensuring that the recovered data is consistent with the historical record and has not been tampered with. Finally, the audit result is stored in the blockchain audit chain through the hash chain technology, forming an unalterable update operation record and completing the comprehensive tracing and monitoring of the entire data update process.

[0050] Through the blockchain audit chain and hash chain technology of step 6, comprehensive tracing, monitoring, and unalterable records of the data update process are achieved. The blockchain audit chain provides complete and unalterable operation records, providing a reliable audit basis for the system's data management. The hash chain technology ensures the integrity and consistency of data during data recovery and audit, preventing data from being tampered with during the recovery process. The rollback engine of the trusted execution environment improves the reliability and security of data recovery, ensuring that data can be accurately restored to the state before the update. This provides strong support for the system's data security and compliance management, enhancing the system's recovery ability and fault tolerance when facing data anomalies.

[0051] Specifically, in step S1, the generation of the multi-dimensional authentication tuple includes wavelet transform processing of the electrocardiogram signal to extract a feature vector, and the environmental data is subjected to noise removal and smoothing processing by a Kalman filtering algorithm. The processed physiological and environmental feature vectors are combined into a multi-dimensional authentication tuple, and the generation process is performed in a trusted execution environment of an edge computing node to enhance the security and integrity of the data.

[0052] It should be noted that in step S1, when the system starts, the edge computing node synchronously starts the data acquisition module, and first acquires the electrocardiogram of the user. The flexible biosensor closely adheres to the user's skin to capture subtle changes in the electrocardiogram signal with high resolution, ensuring the accuracy and integrity of the signal. The acquired electrocardiogram signal is amplified by an amplifier and then sent to an analog-to-digital converter for digital processing. At the same time, the edge computing node also collects the temperature and humidity data of the surrounding environment, and the Internet of Things environment sensor senses the environmental changes in real time and converts the temperature and humidity information into electrical signals. These electrical signals also undergo preprocessing steps such as amplification and filtering to improve signal quality. The preprocessed physiological and environmental signals are sent to the signal processing unit, where the electrocardiogram signal is analyzed in the time-frequency domain through wavelet transform, and the feature vectors containing R-wave, P-wave and T-wave and other key features are extracted. These feature vectors can effectively represent the morphology and periodic changes of the electrocardiogram signal, providing reliable physiological feature basis for subsequent identity authentication. The environmental data is processed using Kalman filter algorithm to filter out noise interference and smooth the data curve, thereby obtaining stable temperature and humidity feature sequences, ensuring that the environmental data can accurately reflect the current actual environmental conditions. The processed physiological feature vectors and environmental feature sequences are integrated into a multi-dimensional authentication tuple, which contains the user's unique physiological features and the current environmental feature information. The multi-dimensional authentication tuple is then transmitted to the authority management system through a secure channel. During transmission, encryption algorithms are used to encrypt the data to prevent data leakage and malicious tampering, ensuring the security of the authentication information during transmission. After receiving the multi-dimensional authentication tuple, the authority management system compares and verifies it with the user feature template pre-registered and stored in the secure database. The comparison algorithm is based on deep learning feature similarity measurement, which uses a well-trained neural network model to match the input authentication tuple and template features, while combining the spatio-temporal consistency verification of environmental data to comprehensively judge the legitimacy of the current user's identity and the authority level. After successful verification, the authority management system generates a dynamic authority token containing the biological feature hash value and the environmental feature fingerprint. This token not only contains the user's authentication information, but also integrates the current environmental features, making the authority token dynamic and environmentally adaptive. The dynamic authority token is securely stored in the local security module and the blockchain ledger. The local security module provides hardware-level security storage guarantee to ensure the confidentiality and integrity of the token locally; the blockchain ledger uses its distributed ledger technology to record token information on multiple nodes, further enhancing the security and tamper resistance of the token, laying a solid foundation for subsequent data update operations based on identity authentication and authority.

[0053] The feature vector is extracted by wavelet transform of the electrocardiogram signal in step S1, and the Kalman filter is used to process the environmental data, and a multi-dimensional authentication tuple is generated in the trusted execution environment of the edge computing node, realizing high-precision and high-security user identity authentication and permission initialization. Wavelet transform can accurately capture the key features of electrocardiogram signals, effectively avoiding the problem of incomplete feature extraction in traditional time or frequency domain analysis methods, and can accurately identify user identity even in complex and variable use scenarios, preventing identity impersonation. The Kalman filter algorithm ensures the accuracy and stability of the environmental data, filters out environmental noise interference, so that the environmental features can truly reflect the current environmental conditions, and enhance the credibility of the authentication result. The multi-dimensional authentication tuple combines physiological and environmental features, providing a more three-dimensional and comprehensive authentication dimension compared to single biometric authentication, significantly improving the security and attack resistance of identity authentication. Generation and processing in the trusted execution environment effectively prevent data leakage and malicious tampering during the authentication process, ensuring the confidentiality and integrity of the authentication information. The generation of dynamic permission tokens combines biometric hash values and environmental feature fingerprints, making the token dynamic and environment-aware, and enabling it to adjust permission status in real time according to environmental changes. Compared with traditional static tokens, it is more adaptable to complex use environments, greatly enhancing the flexibility and security of permission management. The dual storage mechanism of the local security module and the blockchain ledger further improves the security and reliability of the dynamic permission token, ensuring high security and stability during the permission initialization phase, providing a solid and reliable foundation for subsequent data update operations, and effectively reducing the security risks caused by identity authentication and permission management vulnerabilities during data update.

[0054] Specifically, in step S2, when the edge computing node broadcasts a data verification request to surrounding Internet of Things devices, a priority sorting mechanism based on device location and function is used to make surrounding devices respond to the request and extract environmental feature vectors in priority order, and distributed hash table technology is used to optimize the convergence path of the feature vectors to improve the efficiency and response speed of data verification.

[0055] It should be noted that in step S2, the edge computing node will start the data verification process after successfully verifying the validity of the dynamic permission token. First, it will generate a verification request data packet containing the key information required for the data verification task, such as the type of data currently needing to be updated, the corresponding environmental parameter requirements, and the priority evaluation criteria for the response of each peripheral Internet of Things device. Then, the data packet is sent to the surrounding pre-registered and active Internet of Things devices in the form of a broadcast through the wireless communication module, ensuring that all relevant devices can receive this verification request. At the same time, the device management module inside the edge computing node quickly retrieves the detailed information table of the surrounding Internet of Things devices from the database, which covers the unique identifier of each device, the specific geographic location coordinates, the intended functional characteristics, and the record of their past response speed, etc.

[0056] Next, the system sorts the surrounding devices according to the internal preset device location weight coefficient and functional adaptation degree parameter. Specifically, the location weight coefficient is dynamically calculated according to the physical distance between the device and the core computing node and the signal strength distribution in the area, and devices with close distance and stable signal will be given higher priority; the functional adaptation degree parameter is determined according to whether the device sensor type is highly matched with the data type required for verification, and the priority of devices with high matching degree is correspondingly improved. The system automatically constructs a list containing the priority scores of all surrounding devices, which is updated in real time to ensure that the priority of the devices can be adjusted in time according to the changes of the field environment and task requirements. Subsequently, according to this priority list, the surrounding devices respond to the data verification request one by one and start extracting environmental feature vectors. For example, a temperature and humidity sensor will start the internal sensing element, collect the current temperature and humidity data of the environment according to the predetermined sampling frequency, and use the built-in data processing chip to perform preliminary linear correction and quantization processing on these raw data, finally forming a standardized environmental feature vector; the image acquisition device will open the camera module, take pictures of the surrounding environment, extract key information such as texture features and edge features of the image through the fast Fourier transform algorithm, and then construct the corresponding environmental feature vector.

[0057] Meanwhile, to optimize the convergence path of feature vectors, the edge computing node deploys a set of routing optimization algorithms based on distributed hash table technology. This algorithm maps the network topology of peripheral Internet of Things devices into a virtual hash space, calculates a corresponding hash value for each device, and constructs an efficient data transmission path based on these hash values. On this path, data will preferentially select nodes with sufficient network bandwidth and low latency for transmission, ensuring that feature vectors can quickly and losslessly converge to the edge computing node. In this process, the system also monitors the status of data transmission in real time, and once it finds that a node has failed or the transmission efficiency is low, it will immediately activate the backup path and re-plan the data convergence route to ensure smooth progress of the entire data verification process. Finally, all peripheral devices extract environmental feature vectors along the optimized path to the edge computing node, forming a complete verification matrix and providing a high-quality data foundation for subsequent federated learning fusion analysis.

[0058] Through the priority sorting mechanism based on device location and function in step S2 and the distributed hash table technology to optimize the convergence path of feature vectors, an efficient, orderly and reliable data verification process is achieved. The priority sorting mechanism ensures that devices with better locations and more matched functions can participate in the data verification task first, avoiding the disorder of device response and resource waste, significantly improving the response speed and efficiency of data verification. At the same time, the data convergence path optimized by the distributed hash table technology effectively reduces the delay and packet loss rate in the data transmission process, ensuring the integrity and accuracy of feature vectors, thereby providing high-quality data input for subsequent data fusion analysis. This optimization mechanism not only improves the performance of the data verification stage, but also enhances the stability and reliability of the entire data update process, ensuring that subsequent data update operations can be based on accurate environmental perception information.

[0059] Specifically, in step S3, when the edge processor performs fusion analysis using federated learning technology, it adopts a deep neural network architecture to construct a multi-layer perceptron model for feature extraction and fusion analysis of the verification matrix, generate a verification report containing data credibility score and environmental consistency, and automatically trigger different levels of update authorization according to the pre-set credibility threshold through the smart contract, to realize fine-grained permission control.

[0060] It should be noted that in step S3, after the edge processor receives the check matrix aggregated by the peripheral devices, the matrix is first preprocessed, including unified conversion of data format, missing value filling, and outlier detection and correction, to ensure the quality and consistency of the data. Subsequently, a decentralized learning network is constructed using federated learning technology, with each peripheral Internet of Things device as a participating node in the network and the edge processor as the coordination center. In the federated learning process, the parameters of a deep neural network model are first initialized, which adopts a multi-layer perceptron architecture, including an input layer, multiple hidden layers, and an output layer. The number of neurons in the input layer matches the feature dimension of the check matrix, the hidden layers use ReLU activation functions to introduce non-linear characteristics, and the output layer outputs data credibility scores and environmental consistency scores as needed. In the training process, the edge processor broadcasts the initialized model parameters to each Internet of Things device node. Each device node trains the model based on the local check data subset, calculates the gradient, and updates the local model parameters. To protect data privacy, each node only feeds back the updated model parameters to the edge processor without the original data. After receiving the model parameters from each node, the edge processor uses a secure aggregation algorithm to fuse and update the global model parameters, and then broadcasts them to each node for the next round of iterative training. This process is repeated until the model converges, generating a final check report. The generated check report contains two key pieces of information: data credibility scores and environmental consistency scores. Data credibility scores are calculated by analyzing the distribution characteristics of each data point in the check matrix, the deviation from historical data, and the similarity to other device data, quantitatively evaluating the reliability of the data. Environmental consistency scores are calculated by analyzing the spatial and temporal consistency of environmental feature vectors from different devices, determining whether the environmental information collected by each device is consistent with the current scene logic. For example, whether temperature sensor data in the same area is within a reasonable error range, whether the environmental features of image acquisition devices match other environmental data such as temperature and humidity, etc. Finally, the check report is compared with the pre-set credibility threshold through a blockchain smart contract. The smart contract defines the trigger conditions for different levels of update authorization, such as triggering the highest level of authorization when the data credibility score is above a certain high threshold and the environmental consistency score meets expectations, and triggering lower levels of authorization or denying authorization when the score is below the threshold. The edge processor determines whether to allow data updates and the scope and depth of the allowed updates based on the execution results of the smart contract, and further refines the permission control by combining the authorization level with the user's dynamic permission token, ensuring that only strictly verified data and legitimate users can perform data update operations at the corresponding level, thereby achieving fine-grained permission management.

[0061] The federated learning technology combined with the deep neural network architecture is adopted in step S3 to analyze and fuse the check matrix, and the update authorization is triggered based on the smart contract, realizing efficient, intelligent and privacy-protected data checking and permission control. The application of federated learning technology enables data checking to be performed while protecting the privacy of Internet of Things device data. Each device can collaborate to complete model training without sharing raw data, effectively avoiding the risk of data leakage. The deep neural network model can deeply mine the complex feature relationships in the check matrix, generating accurate data credibility scores and environment consistency scores. Compared with traditional statistical analysis methods, the evaluation of data quality is more comprehensive and accurate. The automatic triggering mechanism of the smart contract flexibly adjusts the update authorization level according to the preset credibility threshold, realizing automated permission management based on data quality, reducing manual intervention, and improving the efficiency and response speed of permission control. At the same time, the combination of check results and dynamic permission tokens further enhances the granularity and dynamics of permission control, ensuring the rationality and security of data update operations. This permission control mechanism based on data quality and privacy protection effectively improves the intelligence level and security of the entire data update process, providing reliable protection for subsequent data update operations.

[0062] Specifically, in step S4, when the user is authenticated through physiological signals, a multi-modal biometric fusion technology is used to combine electrocardiogram waveform features and galvanic skin response patterns for comprehensive authentication. A chaotic encryption algorithm is used to encrypt the authentication data, which is then securely compared with the dynamic permission token. In data update operations, a non-volatile storage technology based on memristor arrays is used to ensure the reliability and persistence of data updates.

[0063] It should be noted that in step S4, when the user is authenticated by physiological signals, the system first starts the multi-modal biometric feature fusion technology. The edge computing node simultaneously collects electrocardiogram waveforms and galvanic skin response data of the user. The electrocardiogram data acquires high-resolution signals through a flexible biosensor, which closely adheres to the user's skin to capture key feature points in the cardiac cycle, including P waves, QRS complexes, and T waves, etc. These feature points reflect the regularity of cardiac electrophysiological activity. The collected electrocardiogram signals are preprocessed through amplification, filtering, etc. and then enter the feature extraction module, which performs time-frequency domain analysis based on wavelet transform to extract the key feature vectors of electrocardiogram. The galvanic skin response data is obtained through a skin conductance sensor, which monitors the change rate of skin resistance in real time, captures the emotional fluctuations and physiological stress response of the user, and records the data at a high sampling rate while removing environmental noise interference through a differential amplification circuit. After preprocessing, the galvanic skin response data extracts feature parameters reflecting sympathetic nervous activity, such as skin conductance level (SCL) and amplitude and frequency of skin conductance response (SCR), etc. Subsequently, the system combines electrocardiogram waveform features with galvanic skin response patterns to construct a comprehensive feature vector. A multi-modal fusion model based on deep learning is used to jointly model the two biological features, extract the associated features between the two modalities through the network structure of the shared feature layer and the task-specific layer, and achieve more accurate user identity authentication. The fused feature vector generates an authentication digest through a secure hash algorithm, and is compared with the dynamic permission token generated in step S1. The dynamic permission token is stored in the local security module and the blockchain ledger, and the comparison process uses asymmetric encryption technology to ensure the security of the comparison operation. In the data update operation, the system uses a non-volatile storage technology based on a memristor array. The memristor array is composed of multiple storage units with variable resistance characteristics, and each unit can store multiple bits of data. When updating data, the system precisely controls the voltage pulse of the memristor to change its resistance state, thereby realizing data writing and updating. The non-volatile nature of the memristor ensures that data can be stored for a long time even in the event of power failure, reducing the risk of data loss. At the same time, the high-density integration characteristics of the memristor array enable the storage device to provide large-capacity storage in a limited space, improving the efficiency of data storage. The data is stored in the form of chaotic encryption in the memristor array. The encryption process uses the initial value sensitivity and unpredictability of chaotic systems to generate a pseudo-random key stream, which encrypts the original data bit by bit, enhancing the confidentiality of the data. The encrypted data is stored through the write operation of the memristor array, and the data integrity and consistency are verified by the trusted execution environment to ensure the accuracy and reliability of data updates.

[0064] By adopting the multi-modal biometric feature fusion technology and the chaotic encryption algorithm in step S4, and combining with the non-volatile storage technology based on the memristor array, high security, high reliability and high efficiency of data update operation are realized. The multi-modal biometric feature fusion technology combines electrocardiogram and galvanic skin response two biological characteristics, provides more unique and more difficult to counterfeit authentication basis than single biological characteristics, effectively improves the accuracy and security of user identity authentication, prevents the risk of illegal users impersonating legitimate users to update data. The application of chaotic encryption algorithm enhances the confidentiality of authentication data and update data, making it difficult to be cracked in the process of data transmission and storage, further improving the security of the system. The non-volatile storage technology based on the memristor array ensures the reliability and persistence of data update. The non-volatile characteristics and high-density integration characteristics of the memristor make the data can be saved for a long time after power off, at the same time, reduce the volume and power consumption of storage device, improve the storage efficiency. Compared with traditional storage technology, the memristor array has significant advantages in data writing speed and durability, which can quickly complete data update operation and ensure the long-term reliability of data. This multi-technology fusion scheme not only enhances the security and reliability of data update, but also improves the efficiency and storage performance of data update, providing a solid guarantee for the whole data update process.

[0065] Specifically, in step S5, when the edge computing node broadcasts the update operation record through the blockchain network, the zero-knowledge proof protocol is used for data validity verification, and the preset rules in the smart contract of the blockchain are used to automatically judge the compliance and data consistency of the update operation. If abnormality is found, the state rollback mechanism is triggered through the smart contract, and the original data segment required for rollback is redundantly stored in multiple nodes by using distributed storage technology to improve the reliability and efficiency of data recovery.

[0066] It should be noted that in step S5, the edge computing node immediately encapsulates the update operation record into a data packet containing multiple fields, including the update timestamp, the hash value of the updated data, the type of the update operation, the involved data item identifier, and the user identity information performing the update, etc., after completing the data update operation. This data packet is then broadcasted through the blockchain network to all participating nodes, ensuring transparency and tamper resistance of the information. After receiving the broadcasted update record, each node temporarily stores it in its own memory pool, waiting for further verification processing. At the same time, the system initiates a zero-knowledge proof protocol, with the edge computing node as the prover, to prove the validity of the update operation to the distributed audit nodes without revealing the specific content of the updated data. Specifically, the zero-knowledge proof protocol constructs mathematical puzzles to enable the audit nodes to verify the integrity and consistency of the data without obtaining the original data, thereby ensuring the validity of the data. During the verification process, the system automatically judges the compliance and data consistency of the update operation using the pre-set rules in the smart contract of the blockchain. The smart contract defines a series of conditions and constraints, such as whether the updated data meets the predetermined data format, whether the update is performed within the allowed time range, whether the update operation complies with the user's permission level, etc. The smart contract automatically checks whether the broadcasted update record meets these conditions. If an abnormal or illegal operation is found, the smart contract will immediately trigger a state rollback mechanism, sending a rollback instruction to the edge computing node, requiring it to restore to the last legal state. To improve the reliability and efficiency of data recovery, the system uses distributed storage technology to segment the original data segments required for rollback into multiple parts and increases redundant blocks through a redundancy coding algorithm (such as Reed-Solomon code). These data blocks are then stored on multiple nodes, with each node storing a data block accompanied by a digital signature to ensure data integrity and tamper resistance. When data recovery is needed, the system can collect enough data blocks from multiple nodes, use the redundancy coding algorithm for decoding, and reconstruct the original data, thereby achieving fast and efficient data recovery.

[0067] Through the blockchain broadcast in step S5, zero-knowledge proof verification, smart contract triggered rollback, and distributed storage technology, the high transparency, security, and reliability of data update operations are achieved. Blockchain broadcast ensures that all nodes can obtain update information in a timely manner, enhancing the transparency and tamper resistance of the data update process. Zero-knowledge proof verification protects data privacy while effectively ensuring data validity and integrity, avoiding sensitive data leakage. The automatic triggering mechanism of the smart contract can quickly respond to abnormal situations and initiate state rollback in a timely manner, reducing the risk of data update errors or malicious operations and improving system stability and security. Distributed storage technology enhances data fault tolerance and recovery efficiency through redundant coding and multi-node storage. Even if some nodes fail, the original data can be quickly reconstructed to ensure the continuity and reliability of data update operations. The application of this comprehensive technical solution not only improves the security and efficiency of the data update process but also provides a strong guarantee for the long-term stable operation of the system.

[0068] Specifically, in step S6, when the system starts the state rollback mechanism, data recovery instructions are generated from the historical data state records in the blockchain ledger. The rollback engine in the trusted execution environment gradually recovers the original data according to the instructions. At the same time, the integrity of the recovery process is verified using the hash chain technology to ensure the integrity and tamper resistance of data recovery. The audit results are stored in the blockchain audit chain to form tamper-proof update operation records, enabling comprehensive tracing and monitoring of the entire data update process.

[0069] It should be noted that in step S6, when the system starts the state rollback mechanism, it first accesses the blockchain ledger to query the historical data state records related to the current update operation. The blockchain ledger stores data state snapshots before each data update, including the hash value of the data, the timestamp, and the associated update operation record, etc. The system generates detailed data recovery instructions based on these historical records, which specify the data items to be recovered, the target state, and the execution order of the recovery operations. These instructions are then sent to the trusted execution environment of the edge computing node, where the rollback engine receives the recovery instructions and starts executing the data recovery operations. The rollback engine first locks the current database to prevent new update operations from interfering with the recovery process. Then, it retrieves the corresponding historical data snapshots from the blockchain ledger according to the target state in the instructions. These snapshots are stored in encrypted form, and the rollback engine uses pre-shared keys or digital certificates to decrypt them, ensuring the security of the data during transmission and storage. After decryption, the rollback engine compares the historical data snapshots with the current database state to identify the data items and modifications that need to be rolled back. To ensure the integrity and non-tamperability of data recovery, the system uses hash chain technology to verify the integrity of the recovery process. During the recovery process, the rollback engine calculates the hash value of each recovered data block and compares it with the hash value in the historical record. If the hash values match, it means that the data recovery is correct; if they do not match, an alarm is triggered and the recovery operation is suspended, waiting for human intervention. The hash chain links the hash values of each data block to form an unalterable chain, and any modification to the data will cause the hash chain to break, which will be detected by the system. After the data recovery is completed, the system stores the audit results in the blockchain audit chain. The audit results include the time of the recovery operation, the execution node, the range of data recovered, and the hash chain verification results, etc. The blockchain audit chain uses its distributed ledger feature to record the audit results on multiple nodes, ensuring the non-tamperability and traceability of the audit records. Each audit record contains the hash value of the previous record, forming a continuous audit chain. Any modification to the audit record will be immediately discovered. This provides reliable audit evidence for the system's data management, facilitating subsequent compliance checks and problem tracking.

[0070] Through the blockchain ledger query in step S6, the trusted execution environment recovery, the hash chain integrity verification, and the blockchain audit chain storage, high reliability and traceability of data recovery are realized. The blockchain ledger provides complete historical data state records, ensuring that the recovery operation has evidence and avoiding recovery failure caused by missing or inaccurate historical data. The trusted execution environment provides a secure and reliable operating environment for data recovery, preventing data leakage and malicious tampering during the recovery process. The introduction of the hash chain technology enables each data block in the recovery process to be verified, ensuring the integrity and accuracy of the recovered data and enhancing the system's ability to protect data integrity. The blockchain audit chain further enhances the transparency and traceability of data recovery operations, providing strong support for the system's data management and compliance. This multi-level, multi-technology integrated data recovery solution not only improves the success rate and reliability of data recovery, but also provides a solid guarantee for the long-term stable operation and data security of the system.

[0071] In summary, the present application significantly improves the security and accuracy of user identity authentication by collecting user electrocardiogram and environmental temperature and humidity data through edge computing nodes, generating multi-dimensional authentication tuples, and verifying the spatiotemporal consistency of environmental data using deep learning feature similarity measurement. At the same time, by broadcasting data verification requests to surrounding Internet of Things devices and using federated learning technology to analyze and verify the matrix, a verification report is generated to trigger the update of the authorization level, effectively ensuring the accuracy and environmental adaptability of data updates. During the data update phase, chaotic encryption and trusted execution environment are used to complete data replacement and structure adjustment, and blockchain ledger is used to update data state records, further enhancing the confidentiality and integrity of the data update process. In addition, by broadcasting update operation records through the blockchain network and using zero-knowledge proof technology to verify the validity of the update operation, efficient and secure distributed auditing is achieved. If abnormal or illegal operations are found, the system immediately starts the state rollback mechanism by reading the historical data state records in the blockchain ledger and using the trusted execution environment to restore the original data, ensuring the integrity and non-tamperability of data recovery. Finally, the audit results are stored in the blockchain audit chain through hash chain technology, forming an unalterable update operation record and achieving comprehensive traceability and monitoring of the entire data update process.

[0072] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the present application, which should be covered by the scope of the claims of the present application.

Claims

1. A method of updating local data in a local database, characterized by, The method comprises the following steps: S1, when the system starts, the edge computing node collects the electrocardiogram of the user and the environmental temperature and humidity data, generates a multi-dimensional authentication tuple and transmits it to the permission management system for verification, generates a dynamic permission token and stores it, completes user identity authentication and permission initialization; S2, when a database update request is detected, the permission management system verifies the validity of the dynamic permission token, and after verification, the edge computing node broadcasts a data verification request to the surrounding Internet of Things devices, the surrounding devices extract environmental feature vectors and converge to the edge node to form a verification matrix; S3, the edge processor uses federated learning technology to analyze and fuse the verification matrix, generates a verification report, triggers the update authorization level, and starts the physiological signal gating mechanism; S4, the user authenticates through the physiological signal, the authentication data is compared with the dynamic permission token, and the environmental feature fingerprint is verified at the same time, after matching, the edge processor allows the execution of the data update operation, the data is stored in the form of chaotic encryption during the update process and is replaced and structurally adjusted by using the trusted execution environment, and a new data multi-dimensional feature fingerprint is generated and the data state record in the blockchain account book is updated; S5, the edge computing node broadcasts the update operation record through the blockchain network, and the distributed audit node verifies the compliance and data consistency of the update operation according to the smart contract rules, and uses zero-knowledge proof technology to verify the validity of the update operation during the audit process, and if an abnormal or illegal operation is found, the system immediately starts the state rollback mechanism; S6, read the historical data state record in the blockchain account book, restore the original data by using the trusted execution environment, ensure the integrity and non-tamperability of data recovery, and store the audit result in the blockchain audit chain by using the hash chain technology, forming an unalterable update operation record.

2. The local data update method in a local database as described in claim 1, characterized in that: In step S1, the generation of the multi-dimensional authentication tuple includes wavelet transform processing of the electrocardiogram signal to extract the feature vector, and the environmental data is processed by the Kalman filter algorithm to remove noise and smooth the processing. The physiological and environmental feature vectors are combined into a multi-dimensional authentication tuple, and the generation process is carried out in the trusted execution environment of the edge computing node to enhance the security and integrity of the data.

3. The local data update method in a local database as described in claim 2, characterized in that: In step S2, when the edge computing node broadcasts the data verification request to the surrounding Internet of Things devices, a priority sorting mechanism based on device location and function is used to make the surrounding devices respond to the request in priority order and extract the environmental feature vectors, and a distributed hash table technology is used to optimize the convergence path of the feature vectors to improve the efficiency and response speed of data verification.

4. The local data update method in a local database as described in claim 3, characterized in that: In step S3, when the edge processor uses federated learning technology for fusion analysis, a deep neural network architecture is used to construct a multi-layer perceptron model to extract and analyze the features of the verification matrix, generate a verification report containing data credibility score and environmental consistency, and automatically trigger different levels of update authorization according to the preset credibility threshold through the smart contract, to realize fine-grained permission control.

5. The local data update method in a local database as described in claim 4, characterized in that: In step S4, when the user is authenticated by physiological signals, a multi-modal biometric fusion technology is used to combine electrocardiogram waveform features and galvanic skin response patterns for comprehensive authentication. Chaotic encryption algorithm is used to encrypt the authentication data, and dynamic permission tokens are used for secure comparison. Non-volatile storage technology based on memristor array is used in data update operations to ensure the reliability and persistence of data updates.

6. The local data update method in a local database as described in claim 5, characterized in that: In step S5, when the edge computing node broadcasts the update operation record through the blockchain network, zero-knowledge proof protocol is used for data validity verification, and pre-set rules in the blockchain smart contract are used to automatically judge the compliance and data consistency of the update operation. If abnormalities are found, the smart contract triggers the state rollback mechanism, and the original data segments required for rollback are redundantly stored in multiple nodes using distributed storage technology to improve the reliability and efficiency of data recovery.

7. The local data update method in a local database as described in claim 6, characterized in that: In step S6, when the system starts the state rollback mechanism, data recovery instructions are generated from the historical data state records in the blockchain ledger. The rollback engine in the trusted execution environment restores the original data step by step according to the instructions. Hash chain technology is used to verify the integrity of the recovery process, ensuring the integrity and tamper resistance of data recovery. The audit results are stored in the blockchain audit chain to form an unalterable update operation record, enabling comprehensive tracing and monitoring of the entire data update process.

Images