CN-121037020-B - Intelligent watch data safety transmission method and system based on artificial intelligence

CN121037020BCN 121037020 BCN121037020 BCN 121037020BCN-121037020-B

Abstract

The invention discloses an intelligent watch data safety transmission method and system based on artificial intelligence, which relate to the technical field of data transmission and comprise the steps of collecting multi-source type data through an intelligent watch sensor to form a type data set; determining a disclosure probability according to log access and disclosure matching times of type data, analyzing sensitivity based on acquisition times, summarizing and calculating privacy scores, analyzing importance scores of the type data according to user requests, determining real-time scores according to delay analysis of the data, introducing an entropy weight method to dynamically calculate weights, and calculating a comprehensive sensitivity score. According to the method, the weight of each parameter is dynamically adjusted through the entropy weight method through information entropy, so that the influence on the comprehensive sensitivity score is enhanced, the influence on the parameters with weaker distinguishing capability is reduced, and the distribution characteristics of the comprehensive sensitivity are reflected by the node weights through the introduction of the distribution proportion of the priority components.

Inventors

ZHANG WEN
LIU QINGZHI
WANG DEHUI

Assignees

舟海智能(深圳)有限公司

Dates

Publication Date: 20260508
Application Date: 20250731

Claims (10)

1. An artificial intelligence based intelligent watch data security transmission method is characterized by comprising the following steps: collecting multi-source type data through an intelligent watch sensor to form a type data set; Determining a public probability according to the access times of each type of data in an access log and the public matching times in a user public database, calculating a dynamic adjustment coefficient of potential sensitivity according to the total collection times of each type of data, calculating a privacy score according to the dynamic adjustment coefficient and the public probability, counting the service function coverage rate of each type of data, determining an audit index according to the use times of a user request, calculating the importance score of the type of data based on the coverage rate and the audit index, determining a real-time score according to the delay analysis of the data, introducing an entropy weight method to dynamically calculate a weight, and calculating a comprehensive sensitivity score; Using a k-means clustering algorithm to perform preliminary grouping on all comprehensive sensitivity scores to obtain a plurality of initial clusters, determining the initial cluster number through an elbow method, calculating the profile coefficient of each cluster, introducing a dynamic balance weight factor to perform weighted optimization on the profile coefficient, determining the optimized cluster number, re-clustering according to the optimized cluster number to obtain priority grouping and the central value of each priority group, constructing a graph model composed of intelligent watch equipment and each data transmission node in a data transmission path of the intelligent watch equipment, aiming at each data transmission node, calculating the global weight of the node according to the data quantity ratio of each priority group on the node and the central value of each priority group, calculating the edge weight of each link according to the transmission delay and the maximum load of each link and the global weight of nodes at two ends, and executing a path selection decision by using a Dijkstra algorithm with the maximum total path weight as a target, thereby determining the optimal transmission path; classifying based on the priority group, respectively determining alternative encryption algorithms, and performing data slicing processing on type data in the priority group; and carrying out storage verification on the encrypted data fragments according to priority classification, constructing an LSTM model to predict link attributes, and carrying out anomaly analysis according to prediction errors.
2. The method for secure transmission of artificial intelligence based smart watch data as claimed in claim 1, wherein the introducing entropy weight method dynamically calculates weights, calculates a composite sensitivity score, comprising, Counting data access frequency data of the collected type data, determining log access times according to access logs of each type of data, and simultaneously acquiring public matching times of each type of data according to a user publishing database to calculate a public probability; Analyzing a dynamic adjustment coefficient of potential sensitivity according to the total collection times of each type of data; Calculating privacy scores by combining potential sensitivity and public probability, counting service function coverage rates of different sensor types in the type data, and calculating importance scores by considering the data of the number of times of user request use to define audit indexes; for each type of data Analyzing a real-time score according to the time stamp data and the delay of the current time; and carrying out normalization processing according to the real-time score, the importance score and the privacy score of the type data, introducing an entropy weight method to dynamically calculate weights, and calculating the comprehensive sensitivity score.
3. The method for securely transmitting data of an artificial intelligence based smart watch according to claim 2, wherein the calculating the global weight of the node according to the data amount ratio of each priority group on the node and the central value of each priority group, calculating the edge weight of each link according to the transmission delay of the link, the maximum load and the global weights of the nodes at both ends, and performing a path selection decision using Dijkstra algorithm with the maximum of the total weight of the path as a goal to determine an optimal transmission path comprises, Carrying out distribution analysis according to the comprehensive sensitivity score sets of different types of data, calculating the standard deviation of the comprehensive sensitivity distribution, and judging the bias attribute of the data according to the distribution bias coefficient skewness; The k-means clustering algorithm is used for carrying out preliminary grouping on all comprehensive sensitivity scores according to the comprehensive sensitivity score set, the total intra-group level Fang Wucha is calculated, an elbow curve is drawn according to the total intra-group level error and the preliminary grouping number, and the elbow point is observed, namely the preliminary grouping number is the position at which the decrease of the total intra-group level error is obviously slowed down, so that the value of an initial cluster is determined; carrying out weighted contour coefficient verification according to the clustering result, calculating a contour coefficient, carrying out weighted optimization to obtain an optimized centroid number, clustering the comprehensive sensitivity score data to be used as a priority group, and attaching a central value of a band group; Defining intelligent watch equipment as equipment nodes, setting the intelligent watch equipment as intermediate nodes and terminal nodes according to a data transmission gateway and a server respectively, calculating distribution proportion according to the distribution of type data collected by the nodes in a priority group, and defining global weight of the nodes according to clustering grouping data aiming at each node; According to the transmission links of the nodes, calculating edge weights according to the time delay of the link transmission and the global weight of the maximum load; setting an optimal target according to the edge weight, and selecting a path with the maximum total weight as an optimal path; And executing a path selection decision by using a Dijkstra algorithm, selecting the node with the highest priority from adjacent nodes of the initial node, traversing all the edges connected with the nodes, updating the path weight of the target node for each edge according to the accumulated edge weight, updating the path source of the target node if the accumulated weight is larger, adding the target node into a queue, stopping iterating the backtracking path when the terminal node dequeues, and generating the optimal path.
4. The method for secure transmission of artificial intelligence based smart watch data as claimed in claim 3, wherein the classifying based on the priority group and determining the alternative encryption algorithm respectively, and performing data slicing processing on the type data in the priority group comprises, Based on the comprehensive sensitivity score mean value of the priority group, carrying out priority sorting, and based on the sum of the historical score mean value and the double standard deviation as a priority threshold value, marking the priority group to which the score mean value larger than the priority threshold value belongs as a high priority group, and marking the priority group to which the score mean value smaller than or equal to the priority threshold value belongs as a low priority group; Respectively determining alternative encryption algorithm sets according to the high priority group and the low priority group and the high delay link; Determining a corresponding encryption algorithm according to each side link of the optimal path, and performing data slicing processing on the type data in the priority group, wherein the slicing size u is dynamically adjusted based on the link load and the bandwidth capacity; each fragment is encrypted according to the encryption algorithm of the path and the corresponding link.
5. The method for securely transmitting data to an artificial intelligence based smart watch according to claim 4, wherein said storing and verifying the encrypted pieces of data according to the priority class comprises, The method comprises the steps of designing a differential chain storage rule for encrypted data fragments, wherein the differential chain storage rule comprises the steps of adopting data full-scale encryption storage for high-priority data and managing private keys; and aiming at nodes with different priority levels for storing data, verifying the integrity of the uploading data fragment chain step by step.
6. The method for securely transmitting data of an artificial intelligence based smartwatch according to claim 5, wherein constructing the LSTM model to predict the link attribute and to analyze anomalies based on the prediction error comprises, Constructing a deep learning model based on an LSTM neural network, wherein the deep learning model comprises an input layer, an LSTM layer and an output layer, the input layer inputs a time sequence which comprises transmission delay, bandwidth utilization rate, link packet loss rate and retransmission times of a link, the LSTM layer extracts a time sequence relation of the time sequence and outputs a predicted value of the next moment through the output layer; training the model through calibrated training set data, selecting a cross entropy loss function to calculate the difference between the model predicted category probability and the actual label, using an Adam optimizer to perform gradient descent optimization, updating the parameters of the model, stopping iteration if the model loss is not obviously reduced in the continuous iteration process, and outputting the model parameters; And acquiring real link state data, outputting and calculating a predicted Euclidean distance according to a model, taking the predicted Euclidean distance as a prediction error, taking the sum of the average value and the double standard deviation of a historical prediction error value as a detection threshold, judging that the data is abnormal if the prediction error is greater than or equal to the detection threshold, and carrying out path selection again.
7. The method for securely transmitting data of an artificial intelligence based smart watch according to claim 1, wherein the collecting multi-source type data by the smart watch sensor to form a type data set comprises, Collecting multi-source type data by a smart watch sensor, and composing a type data set for each type data, wherein each type data Including sensor type, data value, and time stamp.
8. An artificial intelligence based intelligent watch data security transmission system, based on the artificial intelligence based intelligent watch data security transmission method of any one of claims 1-7, characterized in that it comprises, The type data management module is used for collecting multi-source type data of the intelligent watch sensor, and aggregating collected log information, wherein the collected log information comprises access logs and public matching times of each data type; The comprehensive scoring module is used for calculating a comprehensive sensitivity score of each type of data according to the comprehensive privacy score, the importance score and the real-time score; the prioritization module analyzes the distribution characteristics of the comprehensive sensitivity scores, performs preliminary grouping on the comprehensive sensitivity scores, and optimizes the priority group based on the distribution bias coefficients and the group center value thereof; the path planning module is used for constructing a graph module, applying Dijkstra algorithm and selecting a path with the maximum edge weight accumulation as an optimal data transmission path; The hierarchical data storage module analyzes, encrypts and performs data slicing processing according to the priority, and performs data hierarchical storage; And the anomaly detection module is used for constructing an LSTM neural network, extracting a time sequence relation, predicting the link state attribute at the next moment and carrying out anomaly detection.
9. A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the computer program is characterized in that the processor realizes the steps of the artificial intelligence-based intelligent watch data security transmission method according to any one of claims 1-7 when executing the computer program.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when executed by a processor, implements the steps of the artificial intelligence based smart watch data security transmission method as claimed in any one of claims 1 to 7.

Description

Intelligent watch data safety transmission method and system based on artificial intelligence Technical Field The invention relates to the technical field of data transmission, in particular to an intelligent watch data safety transmission method and system based on artificial intelligence. Background The intelligent watch is used as important equipment, is not only limited to providing basic functions such as time display and the like, but also becomes an important data acquisition tool in the fields of personal health management, positioning tracking, behavior analysis and the like, and the multi-sensor architecture of the intelligent watch enables the intelligent watch to acquire various types of data in real time, including but not limited to heart rate, step number, blood oxygen concentration, geographic position and environmental data, and the large-scale and multi-source heterogeneous data provides data support for medical treatment, motion analysis and behavior data mining application; However, the data are also subjected to serious security and privacy threat in the process of generation, transmission and storage, if the extremely sensitive data (such as heart rate and position information) of a user are maliciously accessed or leaked, serious personal privacy infringement can be possibly caused, although some researches about data transmission optimization of a smart watch are carried out in recent years, such as encryption improvement of high-sensitivity data and exploration of partial path optimization algorithm, the prior art still has the defects that firstly the prior scheme lacks multi-dimensional dynamic analysis on data privacy protection, the priority distribution of the data cannot be comprehensively calculated from the aspects of privacy, instantaneity and importance synergy, the high-sensitivity data cannot be fully protected, meanwhile, unnecessary encryption of the low-sensitivity data occupies additional resources, so that the transmission efficiency is reduced, and secondly, in terms of path selection, the network transmission efficiency and reliability are difficult to consider due to the fact that the time delay, the load state and the node sensitive data distribution of a dynamic link cannot be flexibly adjusted. Disclosure of Invention The present invention has been made in view of the above-described problems occurring in the prior art. The invention provides an intelligent watch data safety transmission method and system based on artificial intelligence, which solve the problems that the prior scheme lacks multidimensional dynamic analysis for data privacy protection, cannot comprehensively calculate the priority distribution of data from the aspects of privacy, real-time and importance cooperation, so that high-sensitivity data cannot be fully protected, unnecessary encryption of low-sensitivity data occupies extra resources to reduce transmission efficiency, and the time delay, load state and node-sensitivity data distribution of a dynamic link cannot be flexibly adjusted in the aspect of path selection, so that the network transmission efficiency and reliability are difficult to consider. In order to solve the technical problems, the invention provides the following technical scheme: in a first aspect, the invention provides an artificial intelligence-based smart watch data security transmission method, which comprises the steps of acquiring multi-source type data through a smart watch sensor to form a type data set; determining a disclosure probability according to log access and disclosure matching times of type data, analyzing sensitivity based on acquisition times, summarizing and calculating privacy scores, determining a real-time score according to importance scores of user request analysis type data and delay analysis of the data, introducing an entropy weight method to dynamically calculate weights, and calculating a comprehensive sensitivity score; Carrying out distribution analysis on type data according to the comprehensive sensitivity score, determining a distribution bias state coefficient, carrying out preliminary grouping on all comprehensive sensitivity scores by using a k-means clustering algorithm, determining the value of an initial cluster by an elbow method, determining a dynamic weight according to the distribution bias state coefficient in combination with the comprehensive sensitivity score, carrying out weighted optimization on the value of the cluster to obtain an optimized priority group, defining a global weight based on the data proportion of the priority group by a graph model formed by data transmission nodes of intelligent watch equipment, calculating an edge weight, and executing a path selection decision by a Dijkstra algorithm to determine an optimal path; classifying based on the priority group, respectively determining alternative encryption algorithms, and performing data slicing processing on type data in the priority group; and carry