Private data protection method and device for communication network

Abstract

The invention discloses a privacy data protection method and device oriented to a communication network. The method comprises the following steps: acquiring network operation state information of a communication network in a preset time period, and acquiring a user privacy data set corresponding to a target edge node in the communication network; analyzing the network operation state information and the user privacy data set by using a pre-trained privacy protection decision model to obtain a target privacy budget allocation strategy which enables the communication delay to be the lowest and meets a preset privacy protection requirement, the target privacy budget allocation strategy at least comprises a first target privacy protection budget allocated to the target edge node and a second target privacy protection budget allocated to the center node; and performing privacy protection processing on the user privacy data set according to the target privacy budget distribution strategy. The technical problem that dynamic change network optimization and security analysis requirements are difficult to meet due to the fact that a static privacy budget allocation strategy is adopted in related technologies is solved.

Description

Technical Field

[0001] The present application relates to the field of network information security technology, and more specifically, to a method and device for protecting privacy data in a communication network. Background Art

[0002] Against the backdrop of the rapid development of mobile communications and the Internet of Things (IoT), differential privacy (DP), as a powerful privacy protection technology, has been widely used in data sharing and analysis to protect personal data from direct or indirect leakage risks during statistical analysis. However, existing DP practices mostly rely on static privacy budget allocation strategies, which are gradually revealing their limitations in the highly dynamic 5G network environment, mainly manifested in the following aspects:

[0003] (1) The conflict between static privacy budget and dynamic network environment: 5G networks provide customized Quality of Service (QoS) for different types of services through network slicing technology, but traditional differential privacy methods use a fixed privacy budget and cannot make real-time adjustments based on changes in actual network load and user activity patterns. This means that during high-traffic periods, more noise needs to be injected to maintain the same level of privacy protection, which will seriously damage the accuracy and availability of data, thereby affecting the quality of network optimization decisions.

[0004] (2) Current differential privacy schemes tend to implement data privacy protection on central servers or cloud platforms. While this facilitates centralized resource management and unified standards, it also introduces additional communication delays and computational burdens. With the popularization of 5G networks and the surge in the number of IoT devices, large amounts of real-time data streams converge on central nodes. Traditional centralized privacy processing mechanisms can no longer meet the millisecond-level low-latency communication requirements, becoming a key factor restricting the efficiency of network optimization and security analysis.

[0005] In summary, static privacy budget allocation strategies lack flexibility and adaptability when faced with the high speed, low latency, and massive connectivity characteristics of 5G networks. Therefore, ensuring data privacy while maintaining data availability and network service efficiency has become a key technical issue that needs to be addressed. Summary of the Invention

[0006] The embodiments of the present application provide a privacy data protection method and device for communication networks, so as to at least solve the technical problem that related technologies use static privacy budget allocation strategies and are difficult to meet the dynamically changing network optimization and security analysis needs.

[0007] According to one aspect of an embodiment of the present application, a privacy data protection method for a communication network is provided, including: obtaining network operation status information of the communication network within a preset time period, and obtaining a user privacy data set corresponding to a target edge node in the communication network; using a pre-trained privacy protection decision model to analyze the network operation status information and the user privacy data set to obtain a target privacy budget allocation strategy that minimizes communication latency and meets preset privacy protection requirements, wherein the privacy protection decision model is trained based on a dual deep Q learning algorithm and Stackelberg game theory, and the target privacy budget allocation strategy includes at least: a first target privacy protection budget allocated to the target edge node and a second target privacy protection budget allocated to the central node; and privacy protection processing is performed on the user privacy data set according to the target privacy budget allocation strategy.

[0008] Optionally, obtaining a user privacy dataset corresponding to a target edge node in a communication network includes: determining multiple users within the coverage area of the target edge node; for each user, obtaining the user's privacy dataset, wherein the privacy dataset includes at least: real-time location coordinates and a device identifier; adding spatiotemporal tags to the real-time location coordinates using GeoHash encoding to obtain spatiotemporal location data; and composing the user privacy dataset corresponding to the target edge node from the processed privacy datasets of multiple users within the coverage area of the target edge node.

[0009] Optionally, the training process of the privacy protection decision model includes: constructing an online Q network and a target Q network for solving the target privacy budget allocation strategy, and initializing the network weight parameters of the online Q network and the target Q network; setting an experience pool and determining the capacity of the experience pool; determining a first preset number of iteration cycles, and iteratively solving the privacy budget allocation strategy and network weight parameters through the following steps: in each iteration cycle, initializing the communication network environment state, and determining a second preset number of calculation cycles, wherein the communication network environment state includes at least: the user privacy data set and network operation status information corresponding to the target edge node in the communication network; in each calculation cycle, determining the second initial privacy protection budget of the central node based on the current communication network environment state, and determining the first initial privacy protection budget of the target edge node in combination with the preset total privacy budget, and determining the initial privacy budget allocation strategy based on the first initial privacy protection budget and the second initial privacy protection budget; based on the initial privacy budget allocation strategy, using Stackelberg game theory to conduct multiple rounds of games to obtain multiple privacy budget allocation strategies; inputting the current communication network environment state into the online Q network, and calculating each privacy The predicted Q value corresponding to the budget allocation strategy is obtained, and the target privacy budget allocation strategy corresponding to the maximum predicted Q value is selected based on the greedy strategy, where the target privacy budget allocation strategy is the privacy budget allocation strategy that minimizes the communication delay and meets the preset privacy protection requirements; the corresponding privacy processing is performed according to the target privacy budget allocation strategy, the corresponding reward and the new state of the communication network environment are obtained, and the current communication network environment state, the target privacy budget allocation strategy, the reward and the new state of the communication network environment are taken as a sample, and the sample is then stored in the experience pool; multiple samples randomly sampled from the experience pool are input into the neural network, the probability of each sample being sampled, the mean square error loss function and the loss function weight are calculated, and the target Q value corresponding to the second target privacy protection budget is determined based on the obtained calculation results; based on the predicted Q value and the target Q value corresponding to the target privacy budget allocation strategy, the network weight parameters of the online Q network are updated by the gradient descent method; after a third preset number of calculation cycles, the network weight parameters of the target Q network are updated based on the network weight parameters of the online Q network, where the third preset number is less than the second preset number; after the iteration is completed, the obtained target Q network is used as the privacy protection decision model.

[0010] Optionally, the current communication network environment state is input into the online Q network to calculate the predicted Q values corresponding to various privacy budget allocation strategies, including: for each privacy budget allocation strategy, calculating the data utility corresponding to the privacy budget allocation strategy based on the user privacy data set of the edge node, the network operation status information, and the privacy budget allocation strategy; and determining the predicted Q values corresponding to various privacy budget allocation strategies based on the long-term data utility values corresponding to the various privacy budget allocation strategies.

[0011] Optionally, privacy protection processing is performed on the user privacy dataset according to the target privacy budget allocation strategy, including: dividing the preset time period according to the preset time window to obtain multiple sub-time periods, and determining the spatiotemporal sensitivity of the coverage area of the target edge node corresponding to each sub-time period according to the following formula: Δ T =α·Δ base +β·log(1+N T / N avg ), where Δ T represents the spatiotemporal sensitivity of the coverage area of the target edge node in the sub-time period T, Δ base Indicates the preset temporal and spatial sensitivity reference value, N T N represents the number of concurrent users in the coverage area of the target edge node within the sub-time period T. avg represents the average number of concurrent users in the coverage area of the target edge node within a preset time period, α and β respectively represent weight parameters, and α + β = 1; initial privacy protection processing is performed on the user privacy dataset corresponding to the target edge node based on the first target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period, where the initial privacy protection processing at least includes: noise addition processing and desensitization processing; secondary privacy protection processing is performed on the user privacy dataset corresponding to the target edge node based on the second target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period, where the secondary privacy protection processing at least includes: noise addition processing.

[0012] Optionally, initial privacy protection processing is performed on the user privacy data set corresponding to the target edge node based on the first target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period, including: for each sub-time period, determining the size relationship between the spatiotemporal sensitivity within the sub-time period and the spatiotemporal sensitivity baseline value; when the spatiotemporal sensitivity corresponding to the sub-time period is not less than the spatiotemporal sensitivity baseline value, adding a first truncated Laplace noise to the spatiotemporal location data of each user in the coverage area of the target edge node based on the first target privacy protection budget, and desensitizing the device identifier of each user to obtain a device pseudo-identifier; when the spatiotemporal sensitivity corresponding to the sub-time period is less than the spatiotemporal sensitivity baseline value, adding a second truncated Laplace noise to the spatiotemporal location data of each user in the coverage area of the target edge node based on the first target privacy protection budget, and desensitizing the device identifier of each user to obtain a device pseudo-identifier; wherein, the distribution width of the first truncated Laplace noise is smaller than the distribution width of the second truncated Laplace noise.

[0013] Optionally, desensitizing the device identifier of each user to obtain a device pseudo-identifier includes: for each user in the coverage area of the target edge node, determining a preset first random number, where the value of the first random number is between 0 and 1; and calculating the cumulative probability of different device pseudo-identifiers according to the following formula: Where r represents the user's device identifier, o represents the device pseudo identifier, and ε edge represents the first target privacy protection budget, d(o,r) represents the example measurement of the device pseudo-identifier and the user's device identifier, Z represents the normalization constant, and protocolCheck(o) represents the compliance check function used to perform compliance check on the device pseudo-identifier; the device pseudo-identifier corresponding to the device with a cumulative probability greater than the first random number is used as the device pseudo-identifier after desensitizing the user's device identifier.

[0014] Optionally, secondary privacy protection processing is performed on the user privacy dataset corresponding to the target edge node based on the second target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period, including: for each user in the coverage area of the target edge node, the proportion λ of Gaussian noise added to the user's privacy dataset is calculated according to the following formula: Wherein, γ represents a weight parameter, ACC represents a control parameter, and the value of ACC changes continuously with the convergence process of the privacy-preserving decision model, wherein the value of ACC in the early convergence stage of the privacy-preserving decision model is smaller than the value in the later convergence stage of the privacy-preserving decision model. The relationship between the proportion λ and a preset second random number is determined, wherein the value of the second random number is between 0 and 1. When the proportion λ is not less than the second random number, Gaussian noise corresponding to each sub-time period is generated based on the second target privacy-preserving budget and the spatiotemporal sensitivity corresponding to each sub-time period, and the Gaussian noise corresponding to each sub-time period is added to the user's private dataset in the corresponding sub-time period. When the proportion λ is not less than the second random number, Laplace noise corresponding to each sub-time period is generated based on the second target privacy-preserving budget and the spatiotemporal sensitivity corresponding to each sub-time period, and the Laplace noise corresponding to each sub-time period is added to the user's private dataset in the corresponding sub-time period.

[0015] According to another aspect of an embodiment of the present application, a privacy data protection device for a communication network is also provided, including: an acquisition module, used to obtain network operation status information of the communication network within a preset time period, and obtain a user privacy data set corresponding to a target edge node in the communication network; a decision module, used to use a pre-trained privacy protection decision model to analyze the network operation status information and the user privacy data set to obtain a target privacy budget allocation strategy that minimizes the communication delay and meets the preset privacy protection requirements, wherein the privacy protection decision model is trained based on a dual deep Q learning algorithm and Stackelberg game theory, and the target privacy budget allocation strategy includes at least: a first target privacy protection budget allocated to the target edge node, and a second target privacy protection budget allocated to the central node; the privacy protection module is used to perform privacy protection processing on the user privacy data set according to the target privacy budget allocation strategy.

[0016] According to another aspect of an embodiment of the present application, an electronic device is further provided, which includes: a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to execute the above-mentioned privacy data protection method for a communication network through the computer program.

[0017] In an embodiment of the present application, the system intelligently analyzes the network operation status information of the communication network and the user privacy data set of the target edge node within the network based on a privacy-preserving decision model pre-trained using a dual deep Q-learning algorithm and Stackelberg game theory, and obtains the optimal target privacy budget allocation strategy for privacy-preserving the user privacy data set. Therefore, the embodiment of the present application effectively balances privacy protection and network performance, achieving the goal of protecting user privacy while ensuring network operation efficiency and security. This further solves the technical problem that related technologies use static privacy budget allocation strategies, which are difficult to meet the dynamically changing needs of network optimization and security analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0019] Figure 1 This is a flow chart of an optional method for protecting privacy data in a communication network according to an embodiment of the present application;

[0020] Figure 2 1 is a schematic structural diagram of an optional privacy data protection device for a communication network according to an embodiment of the present application;

[0021] Figure 3It is a schematic structural diagram of an optional electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0023] It should be noted that the terms "first", "second", etc. in the specification, claims, and drawings of the present application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products, or devices.

[0024] In order to better understand the embodiments of the present application, some nouns or terms that appear in the description of the embodiments of the present application are first translated and explained as follows:

[0025] Truncated Laplace mechanism: A variant of the Laplace mechanism, it limits the range of input and output by designing a probability density function, ensuring that the output range of users is consistent under different noise perturbations, achieving poor privacy and bounded privacy loss.

[0026] Hash function: A function that converts data of any size into data of a specific size. The converted data is called a hash value or hash code.

[0027] GeoHash: A geocoding system that encodes geographic coordinates (latitude and longitude) into strings. It recursively divides the Earth's surface into smaller rectangular regions, each with a unique GeoHash code. The length of the GeoHash code directly affects the accuracy of the region it represents. The longer the code, the smaller the region represented and the higher the accuracy.

[0028] The Stackelberg game, also known as the Stackelberg game, is a non-cooperative game model in game theory. In a Stackelberg game, participants are divided into two roles: leader and follower. The leader makes decisions first, while followers react after observing the leader's decisions. Therefore, the leader can be considered the first mover in the game, taking into account the followers' reactions and formulating strategies accordingly. Followers, on the other hand, choose the best decision for themselves based on the leader's strategy.

[0029] Example 1

[0030] According to an embodiment of the present application, a privacy data protection method for a communication network is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0031] Figure 1 This is a flow chart of a privacy data protection method for a communication network provided in accordance with an embodiment of the present application. Figure 1 As shown, the method includes the following steps:

[0032] Step S102: obtaining network operation status information of the communication network within a preset time period and a user privacy data set corresponding to a target edge node in the communication network.

[0033] In the technical solution provided in step S102 above, the network operation status information is used to reflect the network monitoring status of the communication network, while the user privacy dataset includes the privacy datasets of multiple users within the coverage area of the target edge node (base station or gateway), such as the user's spatiotemporal location information, device fingerprint information, network signaling data, etc. Together, these data constitute a real-time status snapshot that describes the network health and user behavior patterns.

[0034] In step S104, the network operation status information and the user privacy data set are analyzed using a pre-trained privacy protection decision model to obtain a target privacy budget allocation strategy that minimizes the communication delay and meets the preset privacy protection requirements.

[0035] In the technical solution provided in step S104 above, the privacy-preserving decision model utilizes an intelligent decision-making framework based on a dual-deep Q-learning algorithm and Stackelberg game theory. This model comprehensively considers real-time network conditions and user data security requirements to generate a privacy budget allocation strategy that both protects privacy and optimizes network performance. This strategy includes at least a first target privacy-preserving budget allocated to specific target edge nodes and a second target privacy-preserving budget allocated to central nodes, which guide subsequent noise injection and data processing operations.

[0036] Step S106: Perform privacy protection processing on the user privacy dataset according to the target privacy budget allocation strategy.

[0037] In the technical solution provided in the above step S106, according to the target privacy budget allocation strategy, that is, the first target privacy protection budget allocated to the target edge node and the second target privacy protection budget allocated to the central node, the corresponding privacy protection strategy is implemented at the target edge node and the central node, thereby achieving accurate protection of the user privacy data set.

[0038] Based on the scheme defined in steps S102 to S106 above, it can be seen that in this embodiment of the present application, the system intelligently analyzes the network operation status information of the communication network and the user privacy dataset of the target edge node within the network based on the privacy-preserving decision model pre-trained by the dual deep Q-learning algorithm and Stackelberg game theory, and obtains the optimal target privacy budget allocation strategy for privacy-preserving processing of the user privacy dataset. Therefore, this embodiment of the present application effectively balances privacy protection and network performance, achieving the goal of protecting user privacy while ensuring network operation efficiency and security.

[0039] The following describes the steps of the privacy data protection method for communication networks in combination with a specific implementation process.

[0040] As an optional implementation, in the technical solution provided in the above step S102, the system can capture the network operation status information of the communication network in real time through built-in sensors or external interfaces, including but not limited to: bandwidth utilization, queue waiting time, attack threat level, historical attack frequency, sensitive field exposure risk, edge CPU utilization, center memory margin, storage I / O pressure, network privacy budget margin, data signal-to-noise ratio, spatiotemporal risk index, etc.

[0041] At the same time, the system can also analyze the geographical distribution and base station configuration information of the communication network, determine the boundaries of the service area of the target edge node, and monitor and record the privacy data sets of all users in the service area of the target edge node, including but not limited to: real-time location coordinates (including longitude, latitude and real-time timestamp), device identifiers (such as media access control MAC address, international mobile equipment identity code IMEI, IP address), network signaling data (reference signal received power RSRP, signal-to-noise ratio SINR, channel quality indication CQI, etc.), these data can be directly obtained on the user's terminal device, or indirectly inferred through base stations and network signaling. Then, because in mobile communication scenarios, there is a close correlation between the user's location, time and network behavior. Therefore, this application adds spatiotemporal labels to the real-time location coordinates of each user, which can facilitate subsequent analysis algorithms to explore these correlations.

[0042] GeoHash encoding is used to discretize each user's real-time location coordinates. This is then combined with a second-level timestamp to form a (longitude, latitude, time) triple, which serves as the spatiotemporal label for the location data. This translates the longitude, latitude, and real-time timestamp into a standardized spatiotemporal feature vector, enabling subsequent processing to accurately correlate the spatial location and temporal sequence of each data point. Finally, the user privacy dataset corresponding to the target edge node is composed of the processed privacy datasets of multiple users within the target edge node's coverage area.

[0043] Specifically, the steps for implementing the above process of adding spatiotemporal tags to real-time location coordinates may include:

[0044] Step 1: Get input parameters (longitude lon, latitude lat, real-time timestamp timestamp);

[0045] Step 2: Divide the real-time timestamp into discrete time units according to the preset time granularity. If the preset time granularity is 5 minutes (i.e. 300 seconds), the discrete time unit can be expressed as: t = timestamp / 300;

[0046] Step 3: Use Geohash encoding to map (longitude lon, latitude lat) into fixed-precision 2D grid coordinates (x, y). The precision of Geohash encoding can be set based on the actual application scenario, such as 6 bits or 8 bits.

[0047] Step 3: Output the code t of the discrete time unit and the two-dimensional grid coordinates (two-dimensional integer coding based on Geohash), that is, output (x, y, t).

[0048] For example, if a user is located at (116.3975°E, 39.9087°N) at 2023-08-01 14:05:03 (the corresponding real-time timestamp is 1690872303), the real-time timestamp is divided into time according to a time granularity of 5 minutes, and is divided into space according to a Geohash code with a precision of 6 bits. The final spatiotemporal position coordinates output are (x=12543, y=28791, t=5636241).

[0049] As an optional implementation, in the technical solution provided in step S104 above, the training process of the privacy-preserving decision model may include:

[0050] First, we construct an online Q-network and a target Q-network (these two networks have the same structure) for solving the target privacy budget allocation strategy, and initialize the network weight parameters of the online Q-network and the target Q-network. The online Q-network is used to evaluate the action value in the current state, while the target Q-network is used to predict the value of the future state. Their initial weights are the same to ensure consistency of the training starting point.

[0051] Next, set up the experience pool and determine its capacity. The experience pool is used to store past state-action-reward-new state sequence samples, and the capacity of the experience pool is pre-set to control the stored sequence samples and ensure that the model can learn from sufficient experience.

[0052] Determine a first preset number of iteration cycles (i.e., the number of iterations for multiple rounds of optimization of the strategy), and iteratively solve the privacy budget allocation strategy and network weight parameters through the following steps:

[0053] In each iteration cycle, the communication network environment state is initialized, and a second predetermined number of calculation cycles (i.e., the number of repetitions of the real-time state evaluation and action selection performed in each iteration) is determined, wherein the communication network environment state includes at least: a user privacy data set corresponding to any edge node in the communication network and network operation status information;

[0054] In each calculation cycle, the following steps are included:

[0055] Step 1: Determine the second initial privacy protection budget of the central node based on the current communication network environment state, and determine the first initial privacy protection budget of the target edge node based on the preset total privacy budget. Then, determine the initial privacy budget allocation strategy (i.e., the starting point of the game) based on the first and second initial privacy protection budgets.

[0056] Step 2: Based on the initial privacy budget allocation strategy, Stackelberg game theory is used to conduct multiple rounds of games to obtain multiple privacy budget allocation strategies.

[0057] Step 3: Input the current communication network environment state into the online Q network and calculate the predicted Q value corresponding to each privacy budget allocation strategy.

[0058] Step 3: Based on the greedy strategy, the target privacy budget allocation strategy corresponding to the maximum predicted Q value is selected, where the target privacy budget allocation strategy is the privacy budget allocation strategy that minimizes the communication delay and meets the preset privacy protection requirements.

[0059] Step 5: Perform corresponding privacy processing (such as noise injection, data obfuscation, etc.) according to the target privacy budget allocation strategy, and obtain corresponding rewards (such as reduced privacy leakage risk, changes in network latency, etc.) and the new state of the communication network environment;

[0060] Step 6: Take the current communication network environment state, target privacy budget allocation strategy, reward, and new communication network environment state as a sample, and then store the sample in the experience pool;

[0061] Step 7: Randomly sample multiple samples from the experience pool for online Q-network learning. The probability of each sample being sampled is related to its importance, which helps focus on the experience segments that will bring the greatest learning results. The neural network can calculate the loss function for each sample, including the mean squared error loss function and the loss function weight, to accurately reflect the contribution of the sample. The target Q value corresponding to the second objective privacy protection budget is then determined based on the calculated results.

[0062] Step 8: Based on the predicted Q-value and target Q-value corresponding to the target privacy budget allocation strategy, the network weight parameters of the online Q-network are updated through backpropagation and gradient descent to optimize the model's prediction ability.

[0063] Then, after a third preset number of computation cycles (typically less than the second preset number), the system can update the target Q network's network weight parameters based on the online Q network's network weight parameters. This helps smooth the target network's prediction curve, prevent model overfitting, and improve stability. As training progresses, the model gradually learns how to optimally allocate the first privacy-preserving budget to edge nodes and the second privacy-preserving budget to central nodes, given a given total privacy-preserving budget.

[0064] Finally, after the iterative training is completed, the obtained target Q network is used as a privacy-preserving decision model to guide the formulation of privacy-preserving strategies in real-time network environments, ensuring that network performance is maximized while maintaining data security.

[0065] Specifically, the objective function of the above Stackelberg game process can be written as:

[0066] maxεedge min εcenter [PrivacyRisk+η·CommDelay]

[0067] Where PrivacyRisk represents the privacy risk value, and i represents all nodes participating in privacy protection in the system (including central nodes and privacy nodes), CommDelay represents the communication delay, and QueueLength represents the queue delay (i.e., the waiting time for private data to be processed by the edge node). Therefore, the length of the queue delay depends on factors such as the network load, the size of the data packet, and the processing capacity of the network equipment. DataSize represents the transmission delay (i.e., the time required for private data to be transmitted from the terminal to the central node). Therefore, the transmission delay is affected by the characteristics of the transmission medium (such as the propagation speed of electromagnetic waves and the transmission speed of optical fibers), signal interference, channel conditions (such as bandwidth and signal strength), and the size of the data packet itself. In addition, η, ω are weight parameters,

[0068] The above Stackelberg game process will be briefly explained through the following two game rounds.

[0069] First, initialize the communication network environment state and the total network privacy budget ε total :Network load = 0.82 (high), attack threat = 2 (high risk), network residual privacy budget = 0.6, model convergence = 0.15, ε total =2;

[0070] In the first round of the game, the central node (i.e., the leader) first selects an action (the second initial privacy protection budget) based on the current state of the communication network environment, such as ε center =Network remaining privacy budget - 0.1 = 0.6, at this time, update the network remaining privacy budget = 2-0.6 = 1.4; then, according to the constraint ε edge ≤1.4, edge nodes (i.e. followers) choose actions (first initial privacy protection budget), i.e. ε edge = 1.2, which is used to maximize data utility; then, according to the above privacy budget

[0071] In the second round of the game, the central node (i.e., the leader) first selects an action (the second initial privacy protection budget) based on the current state of the communication network environment, such as ε center =Network remaining privacy budget + 0.1 = 0.7, at this time, update the network remaining privacy budget = 2-0.7 = 1.3; then, according to the constraint ε edge≤1.3, the edge node (i.e., follower) selects an action (the first initial privacy protection budget), i.e., ε edge = 1.0, which is used to maximize data utility; then, according to the above privacy budget

[0072] Compared with the first round of the game, the immediate reward in the second round of the game is significantly improved. Therefore, multiple rounds of games are carried out according to this process to obtain the target privacy budget allocation strategy that minimizes the communication delay and meets the preset privacy protection requirements.

[0073] Optionally, when the system calculates the predicted Q value corresponding to each privacy budget allocation strategy, its specific implementation is as follows:

[0074] Step 1: For each privacy budget allocation strategy, calculate the data utility corresponding to the privacy budget allocation strategy based on the user privacy dataset of the edge node, network operation status information, and privacy budget allocation strategy.

[0075] Data utility can quantify the impact of privacy protection measures on data availability, network performance, and user service experience. Therefore, when calculating the data utility corresponding to each privacy budget allocation strategy, the embodiments of this application can use the following utility evaluation indicators to comprehensively reflect the utility value of data in different dimensions, including:

[0076] (1) Data accuracy: By comparing the data values before and after processing, the degree of loss of data authenticity caused by the strategy is evaluated.

[0077] (2) Data integrity: Check whether the data is complete, including missing values, lost data fragments and other issues.

[0078] (3) Data consistency: Evaluate the correlation or consistency of processed data at different time points or between different data streams.

[0079] (4) Network performance indicators: such as QoS indicators, including delay, throughput, packet loss rate, etc., which reflect the impact of the strategy on the quality of network services.

[0080] (5) User behavior prediction accuracy: Evaluate the accuracy of user behavior pattern prediction by analyzing the processed data.

[0081] Time series analysis, regression models, or deep learning algorithms are then used to perform long-term data utility predictions on the simulated data. The predicted utility indicators are converted into numerical values through normalization or standardization methods. The prediction results of different utility indicators are then combined into an overall long-term data utility value through weighted averaging or comprehensive scoring. The weighting of the prediction results of different utility indicators during the weighting process should be determined based on the specific application scenario and importance, and this embodiment of the application does not impose specific limitations on this.

[0082] Step 2: Determine the predicted Q-values corresponding to various privacy budget allocation strategies based on their corresponding long-term data utility values. The predicted Q-values reflect the long-term expected benefits of adopting a privacy budget allocation strategy under the current communication network environment.

[0083] In other words, the system maps the long-term data utility corresponding to each privacy budget allocation strategy to a Q-value, ensuring that the model's learning process focuses not only on immediate rewards but also on the strategy's long-term effects. Furthermore, by comparing the long-term data utility and predicted Q-values under different privacy budget allocation strategies, the privacy-preserving decision model can identify strategies that minimize the impact on data utility and network performance while preserving privacy. This repeated process improves the model's predictive power by comparing it with actual long-term data utility. Simultaneously, the privacy budget allocation strategy is dynamically adjusted and optimized through this cycle, ensuring a continuous balance between privacy protection and data utility.

[0084] Therefore, the model training process described above establishes strategic interactions between central and edge nodes through dynamic game theory, and solves for optimal privacy parameter configurations through deep reinforcement learning, enabling real-time responses to multi-dimensional network state changes such as network load and attack threats. Compared to static federated learning frameworks, this mechanism enables more granular control of the global privacy budget and intelligently adjusts the privacy protection budget during network operation. This ensures that the system can find the optimal balance between privacy protection and network performance in a constantly changing environment, improving the security and efficiency of data processing.

[0085] Furthermore, the system can call the trained privacy protection decision model to analyze the network operation status information and user privacy data sets to obtain the target privacy budget allocation strategy that minimizes the communication delay and meets the preset privacy protection requirements.

[0086] As an optional implementation, in the technical solution provided in step S106 above, the system may perform privacy protection processing on the user privacy dataset corresponding to the target edge node in the following method, including:

[0087] Step S1061: Divide the preset time period into multiple sub-time periods according to the preset time window, and determine the spatiotemporal sensitivity of the coverage area of the target edge node in each sub-time period according to the following formula:

[0088] Δ T =α·Δ base +β·log(1+N T / N avg )

[0089] Where, Δ T represents the spatiotemporal sensitivity of the coverage area of the target edge node in the sub-time period T (where Δ T The larger the value, the more it changes according to the user distribution density and network conditions, providing a true reflection of data sensitivity under different time and space conditions). base represents the preset spatiotemporal sensitivity benchmark value (if set to 1, it is obtained based on historical data statistics), N T Indicates the number of concurrent users in the coverage area of the target edge node within the sub-time period T (which is obtained through real-time statistics and is used to reflect the user density within the sub-time period T, and N T The larger the value, the stronger the data correlation and the higher the privacy risk). avg It represents the average number of concurrent users in the coverage area of the target edge node within the preset time period (which is obtained through time window statistics and is used to reflect the average user density in the same sub-time period in the past). α and β are weight parameters (used to control the balance between static benchmark and dynamic adjustment), and α + β = 1.

[0090] Through the above-mentioned time aggregation processing, massive real-time data streams can be segmented and processed, and the aggregation results can be output in time, thereby compressing the data size and reducing storage and transmission overhead.

[0091] Step S1062 : Perform initial privacy protection processing on the user privacy dataset corresponding to the target edge node based on the first target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period. The initial privacy protection processing at least includes: noise addition processing and desensitization processing.

[0092] Step S1063 : performing secondary privacy protection processing on the user privacy dataset corresponding to the target edge node according to the second target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period, wherein the secondary privacy protection processing at least includes: noise addition processing.

[0093] Optionally, in the technical solution provided in step S1062 above, the initial privacy protection operation includes:

[0094] For each sub-time period, determine the magnitude relationship between the spatiotemporal sensitivity within the sub-time period and the spatiotemporal sensitivity benchmark value;

[0095] The spatiotemporal sensitivity corresponding to the sub-time period is not less than the spatiotemporal sensitivity baseline value. According to the first target privacy protection budget, a first amount of truncated Laplace noise is added to the spatiotemporal location data of each user in the coverage area of the target edge node, and the device identifier of each user is desensitized to obtain a device pseudo-identifier.

[0096] When the spatiotemporal sensitivity corresponding to the sub-time period is less than the spatiotemporal sensitivity baseline value, a second amount of truncated Laplace noise is added to the spatiotemporal location data of each user in the coverage area of the target edge node according to the first target privacy protection budget, and the device identifier of each user is desensitized to obtain a device pseudo-identifier.

[0097] Among them, the distribution width of the first truncated Laplace noise is smaller than the distribution width of the second truncated Laplace noise. This is because when the number of concurrent users in the coverage area of the target edge node is large (i.e., the population in the area is dense), the spatiotemporal sensitivity is large and the noise scale is small, so the added noise is relatively small to reduce data distortion; while when the number of concurrent users in the coverage area of the target edge node is small (i.e., the population in the area is sparse), the spatiotemporal sensitivity is low, the noise scale is large, and more noise is added to enhance privacy protection. This noise addition strategy based on real-time spatiotemporal sensitivity adjustment can ensure that utility loss is reduced in data-dense areas and privacy protection is enhanced in sparse areas. It overcomes the problem of excessive utility loss caused by fixed sensitivity in the existing technology, improves the availability of location coordinate information, and makes the privacy protection strategy more accurate.

[0098] Specifically, the process of adding the truncated Laplace noise is as follows: the system first determines the coordinate offset error δ of the differential privacy protection mechanism according to the requirements of the application scenario and the privacy protection requirements; then, the system calculates the scale parameter of the truncated Laplace noise corresponding to each sub-time period based on the first target privacy protection budget allocated to the target edge node and the spatiotemporal sensitivity corresponding to each sub-time period, where the scale parameter determines the distribution width of the noise; then, the amount of truncated Laplace noise is determined according to the amount of data in the user privacy data set in each sub-time period, and the obtained truncated Laplace noise is added to the original data to obtain the blurred data, thereby ensuring that each data is properly protected. For example, for the original coordinate (lat, lon), the blurred coordinate is (lat+noise lat ,lon+noise lon ), where noise lat and noise lonIt is generated based on the truncated Laplace distribution. Finally, in order to prevent the offset error of the blurred position coordinates from exceeding the coordinate offset error δ due to the addition of noise, the position coordinates after adding noise can also be constrained to adjust them to the nearest valid coordinate point to ensure that the offset error of the blurred position coordinates does not exceed the coordinate offset error δ, thereby greatly reducing geographic distortion. The blurred data volume is smaller, which can greatly reduce the cost of data transmission.

[0099] In addition, the above-mentioned desensitization process of the device identifier is as follows:

[0100] First, for each user in the coverage area of the target edge node, a preset first random number is determined, where the value of the first random number is between 0 and 1.

[0101] Next, the cumulative probability of different device pseudo identifiers is calculated according to the following formula:

[0102]

[0103] Where r represents the user's device identifier, o represents the device pseudo identifier, and ε edge represents the first target privacy protection budget, d(o,r) represents the example measurement of the device pseudo-identifier and the user's device identifier, Z represents the normalization constant, and protocolCheck(o) represents the compliance check function, which is used to perform compliance check on the device pseudo-identifier. If it complies, the output is a first valid value (such as 1), otherwise it does not comply, and the output is a second valid value (such as 0).

[0104] Finally, the device pseudo identifier corresponding to the device with a cumulative probability greater than the first random number is used as the device pseudo identifier after desensitizing the user's device identifier.

[0105] For example, if the user's device identifier is the user's IP address, such as 192.168.1.105, when desensitizing the IP address, the protocolCheck constraints are: (1) it must conform to the intranet IP format 192.168.xx; (2) the last two segments are integers from 1 to 255. Then, according to the above method, the following set of compliant candidate device pseudo-identifiers can be obtained: 192.168.45.231, 192.168.89.12, 192.168.203.78. Finally, one of the compliant candidate device pseudo-identifiers can be selected as the desensitized pseudo-IP address. This ensures that the desensitized device pseudo-identifier still complies with communication standards such as 3GPP and supports the secure processing of sensitive information such as MAC addresses and IMSIs. This not only protects data privacy, but also ensures seamless transmission and processing of data under different network protocols, which is a feature generally lacking in existing technologies.

[0106] Optionally, in the technical solution provided in step S1063 above, the secondary privacy protection operation includes:

[0107] Step 1: For each user in the coverage area of the target edge node, calculate the proportion λ of Gaussian noise added to the user's private dataset according to the following formula:

[0108]

[0109] In the formula, γ represents the weight parameter, ACC represents the control parameter, and the value of ACC changes continuously with the convergence process of the privacy-preserving decision model. The value of ACC in the early convergence stage of the privacy-preserving decision model is smaller than that in the later convergence stage. Therefore, in the early stage of model training, using Laplace noise (which has a steeper distribution and causes greater perturbation to the data) can quickly introduce sufficient randomness, prompting the federated learning model to converge faster and find a preliminary global optimal solution. As the model training deepens, adding Gaussian noise in the later convergence stage (which has a smoother distribution and causes less perturbation to the data, but can maintain high data utility) helps improve the model accuracy and reduce the degradation of model performance caused by the introduction of excessive noise.

[0110] The magnitude relationship between the proportion λ and a preset second random number is determined, wherein the value of the second random number is between 0 and 1.

[0111] When the proportion λ is not less than the second random number, Gaussian noise corresponding to each sub-time period is generated according to the second target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period, and the Gaussian noise corresponding to each sub-time period is added to the user's privacy dataset in the corresponding sub-time period.

[0112] When the proportion λ is not less than the second random number, Laplace noise corresponding to each sub-time period is generated according to the second target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period, and the Laplace noise corresponding to each sub-time period is added to the user's privacy dataset in the corresponding sub-time period.

[0113] In the above-mentioned secondary privacy protection processing process, Laplace noise has a strong defensive effect on statistical attacks (such as differential attacks), which can protect the basic distribution characteristics of data from being exploited by attackers; while Gaussian noise helps prevent model-based reverse inference attacks. Therefore, when the network load is high, increasing the proportion of Laplace noise can provide stronger privacy protection and prevent sensitive data from being intercepted by potential attackers due to network congestion; and when the network conditions are good, the proportion of Gaussian noise can be appropriately increased to reduce the degree of data fuzzification, thereby reducing the communication delay between the central node and the edge node and improving the real-time processing capability. In other words, the embodiment of the present application can adapt to various unforeseen scenario changes in the network, such as sudden data leakage risks, traffic anomalies, etc., by dynamically adjusting the ratio of Gaussian noise and Laplace noise, and enhance the system's risk resistance by adjusting the noise type and parameters in real time.

[0114] Through the above steps, it can be seen that the embodiments of this application propose a dynamic layered privacy protection model that can adaptively adjust the privacy protection strength of edge nodes and central nodes based on the real-time state of the network and the spatiotemporal characteristics of user distribution. This mechanism effectively resolves the contradiction between privacy protection and network performance optimization in existing technologies, improving the efficiency of privacy protection and data utility.

[0115] Example 2

[0116] According to an embodiment of the present application, a privacy data protection device for a communication network is provided for implementing the privacy data protection method for a communication network in Example 1. Figure 2 As shown, the privacy data protection device for communication networks includes at least: an acquisition module 22, a decision module 24 and a privacy protection module 26, wherein:

[0117] An acquisition module 22 is configured to acquire network operation status information of the communication network within a preset time period and to acquire a user privacy data set corresponding to a target edge node in the communication network;

[0118] a decision module 24 for analyzing network operation status information and user privacy datasets using a pre-trained privacy-preserving decision model to obtain a target privacy budget allocation strategy that minimizes communication latency and meets preset privacy protection requirements, wherein the privacy-preserving decision model is trained based on a dual deep Q-learning algorithm and Stackelberg game theory, and the target privacy budget allocation strategy includes at least: a first target privacy protection budget allocated to a target edge node and a second target privacy protection budget allocated to a central node;

[0119] The privacy protection module 26 is used to perform privacy protection processing on the user privacy data set according to the target privacy budget allocation strategy.

[0120] The functions of each module of the privacy data protection device for communication networks are described below in conjunction with a specific implementation process.

[0121] As an optional implementation, the acquisition module 22 can capture the network operation status information of the communication network in real time through built-in sensors or external interfaces, including but not limited to: bandwidth utilization, queue waiting time, attack threat level, historical attack frequency, sensitive field exposure risk, edge CPU utilization, center memory margin, storage I / O pressure, network privacy budget margin, data signal-to-noise ratio, spatiotemporal risk index, etc.

[0122] At the same time, the acquisition module 22 can also analyze the geographical distribution and base station configuration information of the communication network, determine the service area boundary of the target edge node, and monitor and record the privacy data set of all users in the service area of the target edge node, including but not limited to: real-time location coordinates (including longitude, latitude and real-time timestamp), device identifiers (such as media access control MAC address, international mobile equipment identity code IMEI, IP address), network signaling data (reference signal received power RSRP, signal-to-noise ratio SINR, channel quality indication CQI, etc.), these data can be directly obtained on the user's terminal device, or indirectly inferred through base stations and network signaling. Then, because in mobile communication scenarios, there is a close correlation between the user's location, time and network behavior. Therefore, this application adds spatiotemporal labels to the real-time location coordinates of each user, which can facilitate subsequent analysis algorithms to explore these correlations.

[0123] GeoHash encoding is used to discretize each user's real-time location coordinates. This is then combined with a second-level timestamp to form a {longitude, latitude, time} triple, which serves as the spatiotemporal label for the location data. This translates the longitude, latitude, and real-time timestamp into a standardized spatiotemporal feature vector, enabling subsequent processing to accurately correlate the spatial location and temporal sequence of each data point. Finally, the user privacy dataset corresponding to the target edge node is composed of the processed privacy datasets of multiple users within the target edge node's coverage area.

[0124] Specifically, the steps for implementing the above process of adding spatiotemporal tags to real-time location coordinates may include:

[0125] Step 1: Get input parameters (longitude lon, latitude lat, real-time timestamp timestamp);

[0126] Step 2: Divide the real-time timestamp into discrete time units according to the preset time granularity. If the preset time granularity is 5 minutes (i.e. 300 seconds), the discrete time unit can be expressed as: t = timestamp / 300;

[0127] Step 3: Use Geohash encoding to map (longitude lon, latitude lat) into fixed-precision 2D grid coordinates (x, y). The precision of Geohash encoding can be set based on the actual application scenario, such as 6 bits or 8 bits.

[0128] Step 3: Output the code t of the discrete time unit and the two-dimensional grid coordinates (two-dimensional integer coding based on Geohash), that is, output (x, y, t).

[0129] As an optional implementation, the privacy data protection device for a communication network in the embodiment of the present application further includes a model training module, and the model training module can train the privacy protection decision model according to the following steps, including:

[0130] First, we construct an online Q-network and a target Q-network (these two networks have the same structure) for solving the target privacy budget allocation strategy, and initialize the network weight parameters of the online Q-network and the target Q-network. The online Q-network is used to evaluate the action value in the current state, while the target Q-network is used to predict the value of the future state. Their initial weights are the same to ensure consistency of the training starting point.

[0131] Next, set up the experience pool and determine its capacity. The experience pool is used to store past state-action-reward-new state (SARS) sequence samples, and the capacity of the experience pool is pre-set to control the stored sequence samples to ensure that the model can learn from sufficient experience.

[0132] Determine a first preset number of iteration cycles (i.e., the number of iterations for multiple rounds of optimization of the strategy), and iteratively solve the privacy budget allocation strategy and network weight parameters through the following steps:

[0133] In each iteration cycle, the communication network environment state is initialized, and a second predetermined number of calculation cycles (i.e., the number of repetitions of the real-time state evaluation and action selection performed in each iteration) is determined, wherein the communication network environment state includes at least: a user privacy data set corresponding to any edge node in the communication network and network operation status information;

[0134] In each calculation cycle, the following steps are included:

[0135] Step 1: Determine the second initial privacy protection budget of the central node based on the current communication network environment state, and determine the first initial privacy protection budget of the target edge node based on the preset total privacy budget. Then, determine the initial privacy budget allocation strategy (i.e., the starting point of the game) based on the first and second initial privacy protection budgets.

[0136] Step 2: Based on the initial privacy budget allocation strategy, Stackelberg game theory is used to conduct multiple rounds of games to obtain multiple privacy budget allocation strategies.

[0137] Step 3: Input the current communication network environment state into the online Q network and calculate the predicted Q value corresponding to each privacy budget allocation strategy.

[0138] Step 3: Based on the greedy strategy, the target privacy budget allocation strategy corresponding to the maximum predicted Q value is selected, where the target privacy budget allocation strategy is the privacy budget allocation strategy that minimizes the communication delay and meets the preset privacy protection requirements.

[0139] Step 5: Perform corresponding privacy processing (such as noise injection, data obfuscation, etc.) according to the target privacy budget allocation strategy, and obtain corresponding rewards (such as reduced privacy leakage risk, changes in network latency, etc.) and the new state of the communication network environment;

[0140] Step 6: Take the current communication network environment state, target privacy budget allocation strategy, reward, and new communication network environment state as a sample, and then store the sample in the experience pool;

[0141] Step 7: Randomly sample multiple samples from the experience pool for online Q-network learning. The probability of each sample being sampled is related to its importance, which helps focus on the experience segments that will bring the greatest learning results. The neural network can calculate the loss function for each sample, including the mean squared error loss function and the loss function weight, to accurately reflect the contribution of the sample. The target Q value corresponding to the second objective privacy protection budget is then determined based on the calculated results.

[0142] Step 8: Based on the predicted Q-value and target Q-value corresponding to the target privacy budget allocation strategy, the network weight parameters of the online Q-network are updated through backpropagation and gradient descent to optimize the model's prediction ability.

[0143] Then, after a third preset number of computation cycles (typically less than the second preset number), the system can update the target Q network's network weight parameters based on the online Q network's network weight parameters. This helps smooth the target network's prediction curve, prevent model overfitting, and improve stability. As training progresses, the model gradually learns how to optimally allocate the first privacy-preserving budget to edge nodes and the second privacy-preserving budget to central nodes, given a given total privacy-preserving budget.

[0144] Finally, after the iterative training is completed, the obtained target Q network is used as a privacy-preserving decision model to guide the formulation of privacy-preserving strategies in real-time network environments, ensuring that network performance is maximized while maintaining data security.

[0145] Specifically, the objective function of the above Stackelberg game process can be written as:

[0146] max εedge min εcenter [PrivacyRisk+η·CommDelay]

[0147] Where PrivacyRisk represents the privacy risk value, and i represents all nodes participating in privacy protection in the system (including central nodes and privacy nodes), CommDelay represents the communication delay, and QueueLength represents the queue delay (i.e., the waiting time for private data to be processed by the edge node). Therefore, the length of the queue delay depends on factors such as the network load, the size of the data packet, and the processing capacity of the network equipment. DataSize represents the transmission delay (i.e., the time required for private data to be transmitted from the terminal to the central node). Therefore, the transmission delay is affected by the characteristics of the transmission medium (such as the propagation speed of electromagnetic waves and the transmission speed of optical fibers), signal interference, channel conditions (such as bandwidth and signal strength), and the size of the data packet itself. In addition, η, ω are weight parameters,

[0148] Optionally, when the model training module calculates the predicted Q value corresponding to each privacy budget allocation strategy, its specific implementation is as follows:

[0149] Step 1: For each privacy budget allocation strategy, calculate the data utility corresponding to the privacy budget allocation strategy based on the user privacy dataset of the edge node, network operation status information, and privacy budget allocation strategy.

[0150] Data utility can quantify the impact of privacy protection measures on data availability, network performance, and user service experience. Therefore, when calculating the data utility corresponding to each privacy budget allocation strategy, the model training module can use the following utility evaluation indicators to comprehensively reflect the utility value of data in different dimensions, including:

[0151] (1) Data accuracy: By comparing the data values before and after processing, the degree of loss of data authenticity caused by the strategy is evaluated.

[0152] (2) Data integrity: Check whether the data is complete, including missing values, lost data fragments and other issues.

[0153] (3) Data consistency: Evaluate the correlation or consistency of processed data at different time points or between different data streams.

[0154] (4) Network performance indicators: such as QoS indicators, including delay, throughput, packet loss rate, etc., which reflect the impact of the strategy on the quality of network services.

[0155] (5) User behavior prediction accuracy: Evaluate the accuracy of user behavior pattern prediction by analyzing the processed data.

[0156] Time series analysis, regression models, or deep learning algorithms are then used to perform long-term data utility predictions on the simulated data. The predicted utility indicators are converted into numerical values through normalization or standardization methods. The prediction results of different utility indicators are then combined into an overall long-term data utility value through weighted averaging or comprehensive scoring. The weighting of the prediction results of different utility indicators during the weighting process should be determined based on the specific application scenario and importance, and this embodiment of the application does not impose specific limitations on this.

[0157] Step 2: Determine the predicted Q-values corresponding to various privacy budget allocation strategies based on their corresponding long-term data utility values. The predicted Q-values reflect the long-term expected benefits of adopting a privacy budget allocation strategy under the current communication network environment.

[0158] Specifically, the model training module maps the long-term data utility values corresponding to each privacy budget allocation strategy to Q-values, ensuring that the model's learning process focuses not only on immediate rewards but also on the strategy's long-term effects. Furthermore, by comparing the long-term data utility and predicted Q-values under different privacy budget allocation strategies, the privacy-preserving decision model can identify strategies that minimize the impact on data utility and network performance while preserving privacy. This repeated process improves the model's predictive power through comparison with actual long-term data utility. Simultaneously, the privacy budget allocation strategy is dynamically adjusted and optimized through this cycle, ensuring a continuous balance between privacy protection and data utility.

[0159] Furthermore, the decision module 24 can call the trained privacy protection decision model to analyze the network operation status information and the user privacy data set to obtain a target privacy budget allocation strategy that minimizes the communication delay and meets the preset privacy protection requirements.

[0160] As an optional implementation, the privacy protection module 26 may perform privacy protection processing on the user privacy data set corresponding to the target edge node in the following method, including:

[0161] Step S1061: Divide the preset time period into multiple sub-time periods according to the preset time window, and determine the spatiotemporal sensitivity of the coverage area of the target edge node in each sub-time period according to the following formula:

[0162] ΔT =α·Δ base +β·log(1+N T / N avg )

[0163] Where, Δ T represents the spatiotemporal sensitivity of the coverage area of the target edge node in the sub-time period T (where Δ T The larger the value, the more it changes according to the user distribution density and network conditions, providing a true reflection of data sensitivity under different time and space conditions). base represents the preset spatiotemporal sensitivity benchmark value (if set to 1, it is obtained based on historical data statistics), N T Indicates the number of concurrent users in the coverage area of the target edge node within the sub-time period T (which is obtained through real-time statistics and is used to reflect the user density within the sub-time period T, and N T The larger the value, the stronger the data correlation and the higher the privacy risk). avg It represents the average number of concurrent users in the coverage area of the target edge node within the preset time period (which is obtained through time window statistics and is used to reflect the average user density in the same sub-time period in the past). α and β are weight parameters (used to control the balance between static benchmark and dynamic adjustment), and α + β = 1.

[0164] Through the above-mentioned time aggregation processing, massive real-time data streams can be segmented and processed, and the aggregation results can be output in time, thereby compressing the data size and reducing storage and transmission overhead.

[0165] Step S1062 : Perform initial privacy protection processing on the user privacy dataset corresponding to the target edge node based on the first target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period. The initial privacy protection processing at least includes: noise addition processing and desensitization processing.

[0166] Step S1063 : performing secondary privacy protection processing on the user privacy dataset corresponding to the target edge node according to the second target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period, wherein the secondary privacy protection processing at least includes: noise addition processing.

[0167] Optionally, in the technical solution provided in step S1062 above, the initial privacy protection operation includes:

[0168] For each sub-time period, determine the magnitude relationship between the spatiotemporal sensitivity within the sub-time period and the spatiotemporal sensitivity benchmark value;

[0169] The spatiotemporal sensitivity corresponding to the sub-time period is not less than the spatiotemporal sensitivity baseline value. According to the first target privacy protection budget, a first amount of truncated Laplace noise is added to the spatiotemporal location data of each user in the coverage area of the target edge node, and the device identifier of each user is desensitized to obtain a device pseudo-identifier.

[0170] When the spatiotemporal sensitivity corresponding to the sub-time period is less than the spatiotemporal sensitivity baseline value, a second amount of truncated Laplace noise is added to the spatiotemporal location data of each user in the coverage area of the target edge node according to the first target privacy protection budget, and the device identifier of each user is desensitized to obtain a device pseudo-identifier.

[0171] Among them, the distribution width of the first truncated Laplace noise is smaller than the distribution width of the second truncated Laplace noise. This is because when the number of concurrent users in the coverage area of the target edge node is large (i.e., the population in the area is dense), the spatiotemporal sensitivity is large and the noise scale is small, so the added noise is relatively small to reduce data distortion; while when the number of concurrent users in the coverage area of the target edge node is small (i.e., the population in the area is sparse), the spatiotemporal sensitivity is low, the noise scale is large, and more noise is added to enhance privacy protection. This noise addition strategy based on real-time spatiotemporal sensitivity adjustment can ensure that utility loss is reduced in data-dense areas and privacy protection is enhanced in sparse areas. It overcomes the problem of excessive utility loss caused by fixed sensitivity in the existing technology, improves the availability of location coordinate information, and makes the privacy protection strategy more accurate.

[0172] Specifically, the process of adding the truncated Laplace noise is as follows: the system first determines the coordinate offset error δ of the differential privacy protection mechanism according to the requirements of the application scenario and the privacy protection requirements; then, the system calculates the scale parameter of the truncated Laplace noise corresponding to each sub-time period based on the first target privacy protection budget allocated to the target edge node and the spatiotemporal sensitivity corresponding to each sub-time period, where the scale parameter determines the distribution width of the noise; then, the amount of truncated Laplace noise is determined according to the amount of data in the user privacy data set in each sub-time period, and the obtained truncated Laplace noise is added to the original data to obtain the blurred data, thereby ensuring that each data is properly protected. For example, for the original coordinate (lat, lon), the blurred coordinate is (lat+noise lat ,lon+noise lon ), where noise lat and noise lonIt is generated based on the truncated Laplace distribution. Finally, in order to prevent the offset error of the blurred position coordinates from exceeding the coordinate offset error δ due to the addition of noise, the position coordinates after adding noise can also be constrained to adjust them to the nearest valid coordinate point to ensure that the offset error of the blurred position coordinates does not exceed the coordinate offset error δ, thereby greatly reducing geographic distortion. The blurred data volume is smaller, which can greatly reduce the cost of data transmission.

[0173] In addition, the above-mentioned desensitization process of the device identifier is as follows:

[0174] First, for each user in the coverage area of the target edge node, a preset first random number is determined, where the value of the first random number is between 0 and 1.

[0175] Next, the cumulative probability of different device pseudo identifiers is calculated according to the following formula:

[0176]

[0177] Where r represents the user's device identifier, o represents the device pseudo identifier, and ε edge represents the first target privacy protection budget, d(o,r) represents the example measurement of the device pseudo-identifier and the user's device identifier, Z represents the normalization constant, and protocolCheck(o) represents the compliance check function, which is used to perform compliance check on the device pseudo-identifier. If it complies, the output is a first valid value (such as 1), otherwise it does not comply, and the output is a second valid value (such as 0).

[0178] Finally, the device pseudo identifier corresponding to the device with a cumulative probability greater than the first random number is used as the device pseudo identifier after desensitizing the user's device identifier.

[0179] Optionally, in the technical solution provided in step S1063 above, the secondary privacy protection operation includes:

[0180] Step 1: For each user in the coverage area of the target edge node, calculate the proportion λ of Gaussian noise added to the user's private dataset according to the following formula:

[0181]

[0182] In the formula, γ represents the weight parameter, ACC represents the control parameter, and the value of ACC changes continuously with the convergence process of the privacy-preserving decision model. The value of ACC in the early convergence stage of the privacy-preserving decision model is smaller than that in the later convergence stage. Therefore, in the early stage of model training, using Laplace noise (which has a steeper distribution and causes greater perturbation to the data) can quickly introduce sufficient randomness, prompting the federated learning model to converge faster and find a preliminary global optimal solution. As the model training deepens, adding Gaussian noise in the later convergence stage (which has a smoother distribution and causes less perturbation to the data, but can maintain high data utility) helps improve the model accuracy and reduce the degradation of model performance caused by the introduction of excessive noise.

[0183] The magnitude relationship between the proportion λ and a preset second random number is determined, wherein the value of the second random number is between 0 and 1.

[0184] When the proportion λ is not less than the second random number, Gaussian noise corresponding to each sub-time period is generated according to the second target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period, and the Gaussian noise corresponding to each sub-time period is added to the user's privacy dataset in the corresponding sub-time period.

[0185] When the proportion λ is not less than the second random number, Laplace noise corresponding to each sub-time period is generated according to the second target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period, and the Laplace noise corresponding to each sub-time period is added to the user's privacy dataset in the corresponding sub-time period.

[0186] In the above-mentioned secondary privacy protection processing process, Laplace noise has a strong defensive effect on statistical attacks (such as differential attacks), which can protect the basic distribution characteristics of data from being exploited by attackers; while Gaussian noise helps prevent model-based reverse inference attacks. Therefore, when the network load is high, increasing the proportion of Laplace noise can provide stronger privacy protection and prevent sensitive data from being intercepted by potential attackers due to network congestion; and when the network conditions are good, the proportion of Gaussian noise can be appropriately increased to reduce the degree of data fuzzification, thereby reducing the communication delay between the central node and the edge node and improving the real-time processing capability. In other words, the embodiment of the present application can adapt to various unforeseen scenario changes in the network, such as sudden data leakage risks, traffic anomalies, etc., by dynamically adjusting the ratio of Gaussian noise and Laplace noise, and enhance the system's risk resistance by adjusting the noise type and parameters in real time.

[0187] It should be noted that the modules in the privacy data protection device for communication networks in the embodiment of the present application correspond one-to-one to the implementation steps of the privacy data protection method for communication networks in Example 1. Since a detailed description has been given in Example 1, some details not reflected in this embodiment can be referred to Example 1 and will not be elaborated here.

[0188] Example 3

[0189] According to an embodiment of the present application, a computer program product is also provided, which includes a computer program, wherein when the computer program is executed by a processor, it implements the privacy data protection method for communication networks in Example 1.

[0190] According to an embodiment of the present application, a non-volatile storage medium is also provided, which includes a stored computer program, wherein the device where the non-volatile storage medium is located executes the privacy data protection method for communication networks in Example 1 by running the computer program.

[0191] According to an embodiment of the present application, a processor is further provided, which is used to run a computer program, wherein the computer program executes the privacy data protection method for a communication network in Example 1 when running.

[0192] According to an embodiment of the present application, an electronic device is also provided, which includes: a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to execute the privacy data protection method for a communication network in Example 1 through the computer program.

[0193] Specifically, when the computer program is running, the following steps are executed: obtaining network operation status information of the communication network within a preset time period, and obtaining a user privacy data set corresponding to a target edge node in the communication network; using a pre-trained privacy protection decision model to analyze the network operation status information and the user privacy data set, and obtaining a target privacy budget allocation strategy that minimizes the communication delay and meets the preset privacy protection requirements, wherein the privacy protection decision model is trained based on a dual deep Q learning algorithm and Stackelberg game theory, and the target privacy budget allocation strategy includes at least: a first target privacy protection budget allocated to the target edge node and a second target privacy protection budget allocated to the central node; and performing privacy protection processing on the user privacy data set according to the target privacy budget allocation strategy.

[0194] As an optional implementation, the electronic device may be in the form of a mobile terminal, a computer terminal or a similar computing device. Figure 3 The hardware structure block diagram of an electronic device for implementing a privacy data protection method for a communication network is shown. Figure 3 As shown, the electronic device 30 may include one or more (illustrated as 302a, 302b, ..., 302n in the figure) processors 302 (the processor 302 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 304 for storing data, and a transmission device 306 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It will be understood by those skilled in the art that Figure 3 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 3 More or fewer components than shown, or with Figure 3 Different configurations shown.

[0195] It should be noted that the one or more processors 302 and / or other data processing circuits described above may generally be referred to herein as "data processing circuitry". The data processing circuitry may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuitry may be a single independent processing module, or may be incorporated in whole or in part into any of the other components of the electronic device 30. As described in the embodiments of the present application, the data processing circuitry serves as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0196] The memory 304 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the privacy data protection method for communication networks in the embodiments of the present application. The processor 302 executes various functional applications and data processing by running the software programs and modules stored in the memory 304, that is, implementing the vulnerability detection method for the above-mentioned application. The memory 304 may include a high-speed random access memory and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 304 may further include a memory remotely located relative to the processor 302, and these remote memories may be connected to the electronic device 30 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0197] The transmission device 306 is used to receive or send data via a network. Specific examples of the aforementioned network may include a wireless network provided by the communications provider of the electronic device 30. In one embodiment, the transmission device 306 includes a network interface controller (NIC), which can be connected to other network devices via a base station to enable communication with the Internet. In one embodiment, the transmission device 306 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0198] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the electronic device 30 .

[0199] The serial numbers of the above embodiments are for description only and do not represent the advantages or disadvantages of the embodiments.

[0200] In the above embodiments of the present application, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0201] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only exemplary. For example, the division of units can be a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0202] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple units. Some or all of the units may be selected to achieve the purpose of the present embodiment according to actual needs.

[0203] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0204] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program code.

[0205] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A privacy data protection method for a communication network, characterized in that: include: Obtaining network operation status information of the communication network within a preset time period, and obtaining a user privacy data set corresponding to a target edge node in the communication network; The network operation status information and the user privacy dataset are analyzed using a pre-trained privacy protection decision model to obtain a target privacy budget allocation strategy that minimizes communication latency and meets preset privacy protection requirements, wherein the privacy protection decision model is trained based on a dual deep Q learning algorithm and Stackelberg game theory, and the target privacy budget allocation strategy includes at least: a first target privacy protection budget allocated to the target edge node and a second target privacy protection budget allocated to the central node; The user privacy dataset is privacy-protected according to the target privacy budget allocation strategy.

2. The method according to claim 1, characterized in that Obtaining a user privacy data set corresponding to a target edge node in the communication network includes: Determine a plurality of users in a coverage area of the target edge node; For each user, obtaining a private data set of the user, wherein the private data set includes at least: real-time location coordinates and a device identifier; adding a spatiotemporal tag to the real-time location coordinates using GeoHash encoding to obtain spatiotemporal location data; The user privacy dataset corresponding to the target edge node is composed of the privacy datasets processed by multiple users in the coverage area of the target edge node.

3. The method according to claim 1, characterized in that The training process of the privacy-preserving decision model includes: Constructing an online Q network and a target Q network for solving a target privacy budget allocation strategy, and initializing network weight parameters of the online Q network and the target Q network; Set up an experience pool and determine its capacity; Determine a first preset number of iteration cycles, and iteratively solve the privacy budget allocation strategy and network weight parameters through the following steps: In each iteration cycle, the communication network environment state is initialized, and a second preset number of calculation cycles is determined, wherein the communication network environment state includes at least: a user privacy data set corresponding to a target edge node in the communication network and network operation status information; In each calculation cycle, the second initial privacy protection budget of the central node is determined according to the current communication network environment state, and the first initial privacy protection budget of the target edge node is determined in combination with the preset total privacy budget, and the initial privacy budget allocation strategy is determined according to the first initial privacy protection budget and the second initial privacy protection budget; based on the initial privacy budget allocation strategy, multiple rounds of games are conducted using Stackelberg game theory to obtain multiple privacy budget allocation strategies; the current communication network environment state is input into the online Q network, the predicted Q value corresponding to each privacy budget allocation strategy is calculated, and the target privacy budget allocation strategy corresponding to the maximum predicted Q value is selected based on the greedy strategy, wherein the target privacy budget allocation strategy is to minimize the communication delay and meet the requirements. a privacy budget allocation strategy that meets preset privacy protection requirements; performing corresponding privacy processing according to the target privacy budget allocation strategy, obtaining corresponding rewards and new states of the communication network environment, and taking the current state of the communication network environment, the target privacy budget allocation strategy, the rewards and the new state of the communication network environment as a sample, and then storing the sample in the experience pool; inputting a plurality of randomly sampled samples from the experience pool into the neural network, calculating the probability of each sample being sampled, the mean square error loss function and the loss function weight, and determining the target Q value corresponding to the second target privacy protection budget based on the obtained calculation results; updating the network weight parameters of the online Q network by the gradient descent method based on the predicted Q value and the target Q value corresponding to the target privacy budget allocation strategy; After a third preset number of calculation cycles, updating the network weight parameters of the target Q network according to the network weight parameters of the online Q network, wherein the third preset number is less than the second preset number; After the iteration is completed, the obtained target Q network is used as the privacy protection decision model.

4. The method according to claim 3, characterized in that Inputting the current communication network environment state into the online Q network, and calculating the predicted Q values corresponding to various privacy budget allocation strategies, including: For each of the privacy budget allocation strategies, calculating the data utility corresponding to the privacy budget allocation strategy based on the user privacy dataset of the edge node, the network operation status information, and the privacy budget allocation strategy; The predicted Q values corresponding to the various privacy budget allocation strategies are determined according to the long-term data utility values corresponding to the various privacy budget allocation strategies.

5. The method according to claim 1, wherein Performing privacy protection processing on the user privacy dataset according to the target privacy budget allocation strategy, including: The preset time period is divided according to a preset time window to obtain multiple sub-time periods, and the spatiotemporal sensitivity of the coverage area of the target edge node corresponding to each of the sub-time periods is determined according to the following formula: D T =a·D base +β·log(1+N T / N avg ) Where, Δ T represents the spatiotemporal sensitivity of the coverage area of the target edge node in the sub-time period T, Δ base Indicates the preset temporal and spatial sensitivity reference value, N T N represents the number of concurrent users in the coverage area of the target edge node within the sub-time period T, avg represents the average number of concurrent users in the coverage area of the target edge node within a preset time period, α and β represent weight parameters respectively, and α+β=1; Performing initial privacy protection processing on the user privacy dataset corresponding to the target edge node according to the first target privacy protection budget and the spatiotemporal sensitivity corresponding to each of the sub-time periods, wherein the initial privacy protection processing at least includes: noise addition processing and desensitization processing; A secondary privacy protection process is performed on the user privacy dataset corresponding to the target edge node according to the second target privacy protection budget and the spatiotemporal sensitivity corresponding to each of the sub-time periods, wherein the secondary privacy protection process at least includes: noise addition process.

6. The method according to claim 5, characterized in that Performing initial privacy protection processing on the user privacy dataset corresponding to the target edge node according to the first target privacy protection budget and the spatiotemporal sensitivity corresponding to each of the sub-time periods, including: For each of the sub-time periods, determining a magnitude relationship between the spatiotemporal sensitivity within the sub-time period and the spatiotemporal sensitivity reference value; When the spatiotemporal sensitivity corresponding to the sub-time period is not less than the spatiotemporal sensitivity baseline value, adding a first truncated Laplace noise to the spatiotemporal location data of each user in the coverage area of the target edge node according to the first target privacy protection budget, and desensitizing the device identifier of each user to obtain a device pseudo identifier; When the spatiotemporal sensitivity corresponding to the sub-time period is less than the spatiotemporal sensitivity baseline value, adding a second truncated Laplace noise to the spatiotemporal location data of each user in the coverage area of the target edge node according to the first target privacy protection budget, and desensitizing the device identifier of each user to obtain a device pseudo identifier; The distribution width of the first truncated Laplace noise is smaller than the distribution width of the second truncated Laplace noise.

7. The method according to claim 6, characterized in that Desensitizing the device identifier of each user to obtain a pseudo device identifier, including: For each user within the coverage area of the target edge node, determining a preset first random number, wherein the value of the first random number is between 0 and 1; The cumulative probability of different device pseudo identifiers is calculated according to the following formula: Where r represents the user's device identifier, o represents the device pseudo identifier, ε edge represents the first target privacy protection budget, d(o, r) represents an example measurement of the device pseudo-identifier and the user's device identifier, Z represents a normalization constant, and protocolCheck(o) represents a compliance check function for performing compliance check on the device pseudo-identifier; The device pseudo identifier corresponding to the cumulative probability being greater than the first random number is used as the device pseudo identifier after desensitizing the device identifier of the user.

8. The method according to claim 5, characterized in that Performing secondary privacy protection processing on the user privacy dataset corresponding to the target edge node according to the second target privacy protection budget and the spatiotemporal sensitivity corresponding to each of the sub-time periods, including: For each user in the coverage area of the target edge node, the proportion λ of Gaussian noise added to the user's private dataset is calculated according to the following formula: Wherein, γ represents a weight parameter, ACC represents a control parameter, and the value of ACC changes continuously with the convergence process of the privacy-preserving decision model. The value of ACC in the early convergence stage of the privacy-preserving decision model is smaller than the value in the late convergence stage of the privacy-preserving decision model. Determine the magnitude relationship between the proportion λ and a preset second random number, wherein the value of the second random number is between 0 and 1; If the proportion λ is not less than the second random number, generating Gaussian noise corresponding to each sub-time period according to the second target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period, and adding the Gaussian noise corresponding to each sub-time period to the user's private dataset in the corresponding sub-time period; When the proportion λ is not less than the second random number, Laplace noise corresponding to each sub-time period is generated according to the second target privacy protection budget and the spatiotemporal sensitivity corresponding to each sub-time period, and the Laplace noise corresponding to each sub-time period is added to the privacy dataset of the user in the corresponding sub-time period.

9. A privacy data protection device for a communication network, characterized in that: include: An acquisition module, configured to acquire network operation status information of a communication network within a preset time period, and to acquire a user privacy data set corresponding to a target edge node in the communication network; a decision module, configured to analyze the network operation status information and the user privacy dataset using a pre-trained privacy-preserving decision model to obtain a target privacy budget allocation strategy that minimizes communication latency and meets preset privacy protection requirements, wherein the privacy-preserving decision model is trained based on a dual deep Q-learning algorithm and Stackelberg game theory, and the target privacy budget allocation strategy includes at least: a first target privacy protection budget allocated to the target edge node and a second target privacy protection budget allocated to the central node; The privacy protection module is used to perform privacy protection processing on the user privacy data set according to the target privacy budget allocation strategy.

10. An electronic device, characterized in that: include: A memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the privacy data protection method for a communication network according to any one of claims 1 to 8 through the computer program.

Images