Methods, systems, devices and media for publishing traffic flow data with privacy protection
By establishing intersection models and constraints in traffic flow data, and combining Laplace's mechanism and Lagrange least squares optimization, the problems of computational complexity and insufficient data validity in existing privacy protection methods are solved, and efficient privacy-preserving data release is achieved.
Patent Information
- Application Number
- CN202311303007.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-10
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2043-10-10
AI Technical Summary
Existing methods for protecting traffic flow data privacy struggle to maintain data validity and computational efficiency while ensuring privacy. K-anonymity techniques are vulnerable to attacks, while differential privacy techniques are computationally complex and resource-intensive.
By establishing intersection models and constraints, differential privacy noise addition is performed using the Laplace mechanism, and post-processing optimization is achieved using the Lagrange least squares method. This ensures that the noise-added data meets both differential privacy conditions and intersection constraints, thereby improving data usability.
This approach achieves adjustable differential privacy budgets while improving data availability and computational efficiency, and reducing computational resource consumption.
Smart Images

Figure CN117392838B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of traffic information security technology, and in particular to a method, system, computer device, and storage medium for publishing traffic flow data with privacy protection in mind. Background Technology
[0002] With the development and popularization of intelligent connected vehicles, a large amount of trajectory data will be generated. This data can be used for road planning, site selection, advertising services and other fields. However, directly publishing the trajectory may leak personal privacy such as home address and workplace. But blurring the data in order to protect privacy will reduce the effectiveness of the data. Therefore, how to improve the effectiveness of trajectory data while ensuring personal privacy has become a research topic of great interest to many scholars.
[0003] Existing methods for protecting traffic flow data privacy mainly include K-anonymity and differential privacy techniques. K-anonymity uses algorithms to ensure that each record in a data table has at least the same quasi-identifier attribute value as at least k-1 other records in the table, thereby reducing privacy leaks caused by link attacks. However, this method is based on the assumption that attackers have limited capabilities, resulting in imperfect privacy protection for trajectory data, making it susceptible to cracking and difficult to quantify the degree of privacy protection. Differential privacy techniques inject appropriate noise into the data so that the operation of adding or deleting a data record has an indistinguishable impact on the database query output, thus ensuring the privacy of individuals in the database. Differential privacy techniques applied to trajectory data protection mainly include spatial histogram, prefix tree, and clustering techniques. However, the noise level and grid granularity selected in spatial histogram techniques directly affect data availability and cannot meet the accuracy requirements of data applications. Furthermore, prefix tree and clustering techniques require generating tree-structured data based on the trajectory and performing clustering operations, making the algorithms too complex and requiring significant computational resources for applications with large areas, locations, and a large number of trajectories. Summary of the Invention
[0004] The purpose of this invention is to provide a privacy-preserving traffic flow data publishing method. It determines the intersection constraints of trajectory data through an intersection model, and then performs post-processing on differential privacy-enhanced noisy data based on the intersection constraints. This addresses the application defects of existing traffic flow data privacy protection schemes. It can not only adjust the degree of privacy protection according to the size of the differential privacy budget, but also ensure the availability of data while meeting the differential privacy conditions, and improve data processing efficiency and save computing resources.
[0005] To achieve the above objectives, it is necessary to provide a privacy-preserving method, system, computer device, and storage medium for publishing traffic flow data, addressing the aforementioned technical issues.
[0006] In a first aspect, embodiments of the present invention provide a privacy-preserving method for publishing traffic flow data, applied to a traffic flow data center, the method comprising the following steps:
[0007] Obtain the intersection models within the target traffic area, and establish corresponding intersection constraints based on each intersection model;
[0008] Receive vehicle route information uploaded by each vehicle user within the target traffic area, and statistically analyze the road traffic flow data of each road within the target traffic area in real time based on the road route information of each vehicle.
[0009] Differential privacy noise enhancement is performed on each road traffic flow data to obtain the corresponding noisy road traffic flow data.
[0010] Based on the intersection constraints, the noisy road traffic flow data is post-processed and optimized to obtain corrected road traffic flow data.
[0011] The corrected road traffic flow data will be published to the public through the corresponding traffic flow database.
[0012] Furthermore, the intersection constraints are expressed as follows:
[0013]
[0014] Where, N c X represents the total number of roads corresponding to intersection c; c,i (t) represents the total number of vehicles entering intersection c at time t; Y c,i (t′) represents the total number of vehicles that leave intersection c at time t′.
[0015] Furthermore, the step of performing differential privacy noise processing on the number of vehicles passing through each road to obtain the corresponding noisy road traffic flow data includes:
[0016] Differential privacy noise enhancement is applied to the number of vehicles passing through each road using the Laplace mechanism to obtain the corresponding noisy road traffic flow data; the noisy road traffic flow data is represented as follows:
[0017]
[0018] Where, x i (t) and Let represent the number of vehicles passing through road i at time t and the corresponding noisy road traffic flow data obtained by adding noise, respectively; ΔQ represents the sensitivity of trajectory data query Q; ε represents the size of the differential privacy budget; This represents random noise that satisfies an ε-difference privacy Laplace distribution.
[0019] Furthermore, the step of post-processing and optimizing the noisy road traffic flow data according to the intersection constraints to obtain corrected road traffic flow data includes:
[0020] Under the intersection constraints, a corresponding correction optimization model is established with the optimization objective of minimizing the distance between the corrected road traffic flow data and the noisy road traffic flow data;
[0021] The Lagrange least squares method is used to solve the correction optimization model to obtain the corrected road traffic flow data.
[0022] Furthermore, the correction optimization model is expressed as:
[0023]
[0024]
[0025] in, and These represent the noisy road traffic flow data and the corrected road traffic flow data for road i, respectively. This indicates the intersection constraints.
[0026] Furthermore, the step of solving the correction optimization model using the Lagrange least squares method to obtain the corrected road traffic flow data includes:
[0027] Based on the objective function and intersection constraints of the correction optimization model, Lagrange multipliers are introduced to obtain the corresponding Lagrange function, and based on the Lagrange function, the corresponding partial derivative matrix equation is obtained.
[0028] Based on the partial derivative law of the intersection constraint conditions with respect to the corrected road traffic flow data, the partial derivative matrix equation is simplified to obtain the corrected road traffic flow data function.
[0029] Based on the corrected road traffic flow data function, intersection constraints, and the intersection entry and exit relationships between adjacent intersections, the Lagrange multipliers corresponding to each intersection constraint are obtained.
[0030] The corrected road traffic flow data is obtained based on each Lagrange multiplier and the corrected road traffic flow data function.
[0031] Furthermore, the corrected road traffic flow data function is expressed as follows:
[0032]
[0033] Where, λ idx_in(i) and λ idx_out(i)These represent the intersection numbers at which road i enters and exits; and These represent the noisy road traffic flow data and the corrected road traffic flow data for road i, respectively.
[0034] Secondly, embodiments of the present invention provide a traffic flow data publishing system for privacy protection, the system comprising:
[0035] The constraint creation module is used to obtain the intersection models of each intersection within the target traffic area and establish corresponding intersection constraint conditions based on each intersection model.
[0036] The data statistics module is used to receive road information uploaded by each vehicle user in the target traffic area, and to statistically analyze the road traffic flow data of each road in the target traffic area in real time based on the road information of each vehicle.
[0037] The data noise-adding module is used to perform differential privacy noise-adding processing on traffic flow data of various roads to obtain the corresponding noisy road traffic flow data.
[0038] The noise correction module is used to perform post-processing optimization on the noisy road traffic flow data according to the intersection constraints to obtain corrected road traffic flow data.
[0039] The data publishing module is used to publish the corrected road traffic flow data to the outside world through the corresponding traffic flow database.
[0040] Thirdly, embodiments of the present invention also provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.
[0041] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.
[0042] This application provides a privacy-preserving traffic flow data publishing method, system, computer device, and storage medium. The method enables a traffic flow data center to establish corresponding intersection constraints based on acquired intersection models within a target traffic area. It then performs real-time statistical analysis of road traffic flow data for each road within the target traffic area based on received road information uploaded by various vehicle users. Differential privacy noise addition processing is applied to each road traffic flow data to obtain noisy road traffic flow data. Post-processing optimization is then performed on the noisy road traffic flow data according to the intersection constraints to obtain corrected road traffic flow data. Finally, the corrected road traffic flow data is published externally through a corresponding traffic flow database. Compared with existing technologies, this privacy-preserving traffic flow data publishing method not only adjusts the degree of privacy protection according to the differential privacy budget, but also ensures data availability while meeting differential privacy conditions, improves data processing efficiency, and saves computing resources. Attached Figure Description
[0043] Figure 1 This is a schematic diagram of a crossroads model in an embodiment of the present invention;
[0044] Figure 2 This is a flowchart illustrating the privacy-preserving traffic flow data publishing method in an embodiment of the present invention.
[0045] Figure 3 This is a schematic diagram illustrating the algorithm implementation of the Laplace mechanism used in this embodiment of the invention;
[0046] Figure 4 This is a schematic diagram of the structure of a traffic flow data publishing system for privacy protection in an embodiment of the present invention;
[0047] Figure 5 This is an internal structural diagram of the computer device in an embodiment of the present invention. Detailed Implementation
[0048] To make the objectives, technical solutions, and beneficial effects of this application clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Obviously, the embodiments described below are only part of the embodiments of the present invention and are used to illustrate the present invention, but are not intended to limit the scope of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.
[0049] like Figure 1 As shown, assume that the number of vehicles entering the intersection in a certain time period t is X(t). 1~4 The number of vehicles leaving the intersection during time interval t′ (t′ is after t) is Y(t′).1~4 Based on the common sense that cars at intersections don't appear or disappear out of thin air, that is... Figure 1 The intersection model shown has intersection constraints: If X1 = 100, X2 = X3 = X4 = 0, Y 1~4 If the value is arbitrary, then without privacy protection, if an attacker has background knowledge that a car leaves the intersection from y1 (lowercase letters in variables represent the corresponding roads), the attacker can obtain the car's trajectory x1→y1 from the database, because it can only originate from x1. Conversely, if only one car enters the intersection from x1 within a relatively short timeframe (i.e., X1=1, X2=X3=X4=0), then without privacy protection, the attacker doesn't need background knowledge; they only need to compare Y... 1~4 The value of the vehicle whose trajectory increases by 1 within a given time period can be used to determine the vehicle's trajectory. Clearly, introducing noise into the trajectory data to prevent attackers from inferring an individual's trajectory using background knowledge and the database is an effective way to address the aforementioned privacy issues. However, adding random noise means that the data from different roads no longer meets this constraint, making it difficult to guarantee data availability.
[0050] The privacy-preserving traffic flow data publishing method provided by this invention can be understood as a traffic flow data privacy protection method proposed based on the aforementioned traffic flow data privacy protection problem. It is applied to a traffic flow data center in a centralized differential privacy protection scenario, using a post-processing mechanism based on intersection constraints to ensure that the noisy data satisfies both differential privacy conditions and intersection constraints, thus being "closer" to the original data. This method supports high availability for querying and publishing noisy traffic flow data and protects the privacy of vehicle user trajectory data. Centralized differential privacy refers to users directly uploading their original data to a data center, which then processes and publishes the data. This type of method requires a reliable data center to ensure privacy is not leaked. The following embodiments will provide a detailed description of the privacy-preserving traffic flow data publishing method of this invention.
[0051] In one embodiment, such as Figure 2 As shown, a privacy-preserving traffic flow data publishing method is provided, applied to a traffic flow data center, including the following steps:
[0052] S11. Obtain the intersection models within the target traffic area and establish corresponding intersection constraints based on each intersection model. The target traffic area can be understood as a traffic area associated with a database that publishes traffic flow data (supporting external queries). An intersection model can be understood as an intersection network structure, such as a crossroads or a T-junction. The corresponding intersection constraints can be understood as constraints established based on the principle that cars on the road do not appear or disappear out of thin air, combined with the actual intersection model, to describe the inherent relationship between traffic flow data of each road at the intersection. It should be noted that different intersection models will result in different total numbers of roads in the corresponding intersection constraints, but the overall constraint relationship remains consistent. Specifically, the intersection constraints are expressed as follows:
[0053]
[0054] Where, N c X represents the total number of roads corresponding to intersection c; c,i (t) represents the total number of vehicles entering intersection c at time t; Y c,i (t′) represents the total number of vehicles that leave intersection c at time t′.
[0055] S12. Receive vehicle route information uploaded by each vehicle user within the target traffic area, and based on this information, calculate real-time traffic flow data for each road within the target traffic area. The vehicle route information can be understood as whether a vehicle passed through any road in the target traffic area, and the time taken to pass through each road. Assuming the target traffic area contains m roads and n intersections, when each car passes through road i, it will upload its road-passing information to the traffic flow data center, which will then count the number of vehicles passing through each road, denoted as x. i (t) represents the number of vehicles passing through road i at time t, i.e., x i (t) represents the original road traffic flow data that needs to be protected. It should be noted that the specific method for counting the number of vehicles corresponding to each road in the target traffic area at each time point can be implemented with reference to existing technologies, and will not be detailed here.
[0056] S13. Perform differential privacy noise addition processing on the traffic flow data of each road to obtain the corresponding noisy road traffic flow data; wherein, in principle, other noise addition methods can be used for differential privacy noise addition processing, but in order to ensure the convenience of supporting any type of digital query processing, this embodiment preferably uses differential privacy noise addition processing based on the Laplace mechanism; specifically, the step of performing differential privacy noise addition processing on the number of vehicles passing through each road to obtain the corresponding noisy road traffic flow data includes:
[0057] Differential privacy noise enhancement is applied to the number of vehicles passing through each road using the Laplace mechanism to obtain the corresponding noisy road traffic flow data. The Laplace mechanism is an algorithm that satisfies differential privacy and is primarily used in database counting queries. Specifically, it can be understood as follows: assuming Q is a trajectory data query from database I, then the Laplace mechanism... To satisfy ε-differential privacy, we have:
[0058]
[0059] in, Indicates that it comes from the scale parameter is The random noise is distributed by the Laplace distribution; ΔQ represents the sensitivity of trajectory data query Q, which actually measures how much a small change in the traffic flow database will cause a change in the query result. If the sensitivity is high, the difference between samples / records in the database will cause a large change in the query result, and attackers can more easily infer the information of the record by issuing queries to a pair of adjacent databases that are different from the record; in the traffic flow data privacy protection scenario of this embodiment, ΔQ represents the length L of the longest trajectory in the database, that is, ΔQ = L, because if the length of the longest trajectory is L, then a change in a trajectory will affect the data of at most L intersections, which can be specifically expressed as:
[0060]
[0061] Where I represents the traffic flow database; I′ represents the neighboring database of database I, i.e., the database that differs from database I by at most one record, I′∈nbrs(I), |(II′)∪(I′-I)|=1; Q(I) and Q(I′) represent the query results of trajectory data query Q corresponding to databases I and I′, respectively; ε represents the differential privacy budget size, and the corresponding ε-differential privacy can be understood as: for a random algorithm A, if for all databases I, the neighboring databases I′∈nbrs(I) and the output There exists Pr[A(I)∈S]≤e ε If Pr[A(I')∈S], then algorithm A satisfies ε-differential privacy;
[0062] The algorithm code corresponding to the Laplace mechanism mentioned above is as follows: Figure 3 As shown, it can be directly applied to trajectory data to meet differential privacy conditions, and the degree of privacy protection can be adjusted according to the size of the privacy budget. It overcomes the vulnerability of the K-anonymity algorithm, thereby improving the reliability of user privacy. The resulting noisy road traffic flow data with quantifiable privacy protection is specifically represented as follows:
[0063]
[0064] Where, xi (t) and Let represent the number of vehicles passing through road i at time t and the corresponding noise-added road traffic flow data, respectively.
[0065] However, considering that using only the Laplace mechanism would introduce excessive noise, which greatly reduces the accuracy of the query due to its high sensitivity, in order to ensure the validity and usability of the query data as much as possible, this embodiment preferably introduces a post-processing mechanism to correct noise under intersection constraints, so that the noisy data also meets the intersection constraints, thereby improving the accuracy and usability of the query data without compromising its privacy protection.
[0066] S14. Based on the intersection constraints, perform post-processing optimization on the noisy road traffic flow data to obtain corrected road traffic flow data; wherein, the corrected road traffic flow data can be understood as the closest value of the noisy road traffic flow data obtained based on the intersection constraints; specifically, the step of performing post-processing optimization on the noisy road traffic flow data based on the intersection constraints to obtain corrected road traffic flow data includes:
[0067] Under the intersection constraints, a corresponding correction optimization model is established with the optimization objective of minimizing the distance between the corrected road traffic flow data and the noisy road traffic flow data. Preferably, the distance between the corrected road traffic flow data and the noisy road traffic flow data is measured using L2 Euclidean distance. The resulting correction optimization model is expressed as follows:
[0068]
[0069]
[0070] in, and These represent the noisy road traffic flow data and the corrected road traffic flow data for road i, respectively. Indicates the intersection constraints;
[0071] The Lagrange least squares method is used to solve the correction optimization model to obtain the corrected road traffic flow data. Specifically, the step of using the Lagrange least squares method to solve the correction optimization model to obtain the corrected road traffic flow data includes:
[0072] Based on the objective function and intersection constraints of the correction optimization model, Lagrange multipliers are introduced to obtain the corresponding Lagrange function, and the corresponding partial derivative matrix equation is obtained based on the Lagrange function; wherein, the Lagrange function is expressed as:
[0073]
[0074] Where, λ j Let represent the Lagrange multiplier corresponding to the constraint condition at the j-th intersection;
[0075] The constrained corrected optimization model is then transformed into an unconstrained optimization model using the least squares method, as shown below:
[0076]
[0077] make have:
[0078]
[0079]
[0080] Writing the above system of equations in matrix form, we obtain the partial derivative matrix equations, expressed as:
[0081]
[0082] In the formula,
[0083]
[0084] Where A and B represent the system of equations and The coefficient matrix; m and n represent the total number of roads and intersections within the target traffic area, respectively;
[0085] Considering that directly solving the above partial derivative matrix equation is very complex, this embodiment preferably simplifies matrix A based on the partial derivative law of the corrected road traffic flow data according to the intersection constraint conditions before solving it.
[0086] Based on the partial derivative law of the intersection constraints with respect to the corrected road traffic flow data, the partial derivative matrix equation is simplified to obtain the corrected road traffic flow data function; the specific process of obtaining the corrected road traffic flow data function is as follows:
[0087] Note the intersection constraint function Regarding the variable x for each road i partial derivatives There is a certain pattern: since a road connects at most two intersections, that is, one road variable x i It is related to at most two intersection constraints, therefore, for With the same variable x iThe relevant partial derivatives, of which at most two are non-zero; assuming road i connects intersection j and intersection k, road i is the direction of entry into intersection j and the direction of exit from intersection k. Since the traffic flow entering and exiting the same intersection is equal in the intersection constraint conditions, then we have:
[0088]
[0089] Then, the system of equations The coefficients are all 1, and the values of the non-zero partial derivatives are only ±1. The direction of the road determines whether the partial derivative is positive or negative. For example, when the road direction is the direction of entering the intersection, the non-zero partial derivative related to the road variable is +1, and when the road direction is the direction of leaving the intersection, the non-zero partial derivative related to the road variable is -1.
[0090] Based on the above conclusions, matrix A shown in equation (11) can be simplified using A. i Let the vector be the i-th row vector of matrix A, i.e.:
[0091]
[0092] It can be simplified to:
[0093]
[0094] For ease of result representation, this embodiment defines the function idx_in(i) to represent that the input value is road number i and the return value is the road i entering the intersection number, and defines the function idx_out(i) to represent that the input value is road number i and the return value is the road i leaving the intersection number;
[0095] Based on equations (10) and (12), the corrected road traffic flow data function can be expressed as:
[0096]
[0097] Where, λ idx_in(i) and λ idx_out(i) These represent the intersection numbers at which road i enters and exits; and These represent the noisy road traffic flow data and the corrected road traffic flow data for road i, respectively.
[0098] Based on the corrected road traffic flow data function, intersection constraints, and the entry and exit relationships between adjacent intersections, the Lagrange multipliers corresponding to each intersection constraint are obtained, specifically including:
[0099] Let R_in p R_out represents the set of all roads leading to intersection p. pLet P represent the set of all roads leading out of intersection P, and Adj(p) represent all intersections adjacent to intersection P. Then, for intersection P, we have:
[0100]
[0101] Substituting equation (13) into equation (14), we get:
[0102]
[0103] Based on the entry and exit relationships between adjacent intersections, we can obtain:
[0104]
[0105] Based on the above equation, the vector λ = (λ1, ..., λ) can be obtained. n The solution to )
[0106] The corrected road traffic flow data is obtained based on each Lagrange multiplier and the corrected road traffic flow data function, specifically including:
[0107] Substituting the solution of the vector λ obtained in the above steps back into equation (13), we can obtain the optimized result of the algorithm, that is, the corrected road traffic flow data.
[0108] Based on the characteristic that intersection constraints represent the inherent relationship of road traffic flow data, this embodiment proposes a post-processing algorithm to correct noise under intersection constraints. Without the need to construct a complex data structure design, it reduces the added noise, so that the noisy data can also meet the intersection constraints, becoming "closer" to the original data, while also satisfying differential privacy conditions. This improves data processing and publishing efficiency, saves computing resources, and effectively meets the privacy protection publishing requirements of traffic flow data.
[0109] S15. The corrected road traffic flow data is published externally through the corresponding traffic flow database; wherein, the traffic flow database can be understood as a database used to store road route data information of all user vehicles passing through the target traffic area, and supports external querying. The specific storage form and content are not specifically limited here; it should be noted that the specific method of publishing the corrected road traffic flow data externally through the corresponding traffic flow database can be implemented using existing technology, and will not be elaborated here.
[0110] This application embodiment establishes corresponding intersection constraints based on the obtained intersection models within the target traffic area through a traffic flow data center. It then performs real-time statistical analysis of road traffic flow data for each road within the target traffic area based on the road information uploaded by various vehicle users within the target traffic area. Differential privacy noise enhancement is then applied to the traffic flow data to obtain noisy road traffic flow data. Following the intersection constraints, the noisy road traffic flow data undergoes post-processing optimization to obtain corrected road traffic flow data. Finally, the corrected road traffic flow data is published externally through the corresponding traffic flow database. This technical solution effectively addresses the application deficiencies of existing traffic flow data privacy protection schemes. It not only adjusts the degree of privacy protection according to the differential privacy budget but also ensures data availability while meeting differential privacy conditions, improves data processing efficiency, and saves computing resources.
[0111] It should be noted that although the steps in the flowchart above are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order requirement for the execution of these steps, and they can be executed in other orders.
[0112] In one embodiment, such as Figure 4 As shown, a privacy-preserving traffic flow data publishing system is provided, applied to a traffic flow data center. The system includes:
[0113] Constraint creation module 1 is used to obtain the intersection models of each intersection within the target traffic area and establish corresponding intersection constraint conditions based on each intersection model.
[0114] The data statistics module 2 is used to receive the road information of each vehicle user in the target traffic area, and to count the road traffic flow data of each road in the target traffic area in real time based on the road information of each vehicle.
[0115] Data noise addition module 3 is used to perform differential privacy noise addition processing on each road traffic flow data to obtain the corresponding noisy road traffic flow data;
[0116] The noise correction module 4 is used to perform post-processing optimization on the noisy road traffic flow data according to the intersection constraints to obtain corrected road traffic flow data.
[0117] The data publishing module 5 is used to publish the corrected road traffic flow data to the outside world through the corresponding traffic flow database.
[0118] Specific limitations regarding the privacy-preserving traffic flow data publishing system can be found in the limitations of the privacy-preserving traffic flow data publishing method described above; the corresponding technical effects are equivalent and will not be repeated here. Each module in the aforementioned privacy-preserving traffic flow data publishing system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.
[0119] Figure 5 An internal structural diagram of a computer device is shown in one embodiment. This computer device may specifically be a terminal or a server. Figure 5 As shown, the computer device includes a processor, memory, network interface, display, camera, and input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a privacy-preserving traffic flow data publishing method. The display screen can be an LCD screen or an e-ink display screen. The input device can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.
[0120] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computing devices may include more or fewer components than those shown in the figure, or combine certain components, or have the same component arrangement.
[0121] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method described above.
[0122] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.
[0123] In summary, the present invention provides a privacy-preserving traffic flow data publishing method and system. The method involves a traffic flow data center establishing corresponding intersection constraints based on the acquired intersection models within a target traffic area. It then performs real-time statistical analysis of road traffic flow data for each road within the target traffic area based on the road information uploaded by various vehicle users within the target traffic area. This is followed by differential privacy noise addition processing to obtain noisy road traffic flow data. Furthermore, post-processing optimization is applied to the noisy road traffic flow data based on the intersection constraints to obtain corrected road traffic flow data. Finally, the corrected road traffic flow data is published externally through the corresponding traffic flow database. This method not only adjusts the degree of privacy protection according to the differential privacy budget but also ensures data availability while meeting differential privacy conditions. It also improves data processing efficiency and saves computational resources.
[0124] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts of the embodiments, refer to each other. Each embodiment focuses on its differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.
[0125] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the technical principles of this invention, and these improvements and substitutions should also be considered within the scope of protection of this application. Therefore, the scope of protection of this patent application should be determined by the scope of the claims.
Claims
1. A privacy-preserving traffic flow data publishing method, characterized in that, Applied to traffic flow data centers, the method includes the following steps: Obtain the intersection models within the target traffic area, and establish corresponding intersection constraints based on each intersection model; the intersection constraints are expressed as follows: in, Indicates intersection c The corresponding total number of roads; express t Enter the intersection at any time c The total number of vehicles; express Always leave the intersection c The total number of vehicles; Receive information on the roads traveled by vehicles uploaded by each vehicle user within the target traffic area, and statistically analyze the road traffic flow data of each road within the target traffic area in real time based on the information on the roads traveled by each vehicle. Differential privacy noise enhancement is performed on each road traffic flow data to obtain the corresponding noisy road traffic flow data. Based on the intersection constraints, the noisy road traffic flow data is post-processed and optimized to obtain corrected road traffic flow data, including: Under the intersection constraints, a corresponding correction optimization model is established with the optimization objective of minimizing the distance between the corrected road traffic flow data and the noisy road traffic flow data; The Lagrange least squares method is used to solve the correction optimization model to obtain the corrected road traffic flow data; The corrected road traffic flow data will be published to the public through the corresponding traffic flow database.
2. The method for publishing privacy-preserving traffic flow data as described in claim 1, characterized in that, The step of performing differential privacy noise addition processing on each road traffic flow data to obtain the corresponding noisy road traffic flow data includes: Differential privacy noise enhancement is applied to the number of vehicles passing through each road using the Laplace mechanism to obtain the corresponding noisy road traffic flow data; the noisy road traffic flow data is represented as follows: in, and They represent the roads at time t, respectively. i The number of vehicles passing through the road and the corresponding noise level are used to obtain noisy road traffic flow data; Indicates trajectory data query Sensitivity; Indicates the size of the differential privacy budget; Indicates satisfaction - Random noise with a Laplace distribution for differential privacy.
3. The privacy-preserving traffic flow data publishing method as described in claim 1, characterized in that, The correction optimization model is expressed as follows: min St. in, and Representing roads i Noisy road traffic flow data and corrected road traffic flow data; This indicates the intersection constraints.
4. The traffic flow data publishing method for privacy protection as described in claim 3, characterized in that, The step of solving the correction optimization model using the Lagrange least squares method to obtain the corrected road traffic flow data includes: Based on the objective function and intersection constraints of the correction optimization model, Lagrange multipliers are introduced to obtain the corresponding Lagrange function, and based on the Lagrange function, the corresponding partial derivative matrix equation is obtained. Based on the partial derivative law of the intersection constraint conditions with respect to the corrected road traffic flow data, the partial derivative matrix equation is simplified to obtain the corrected road traffic flow data function. Based on the corrected road traffic flow data function, intersection constraints, and the intersection entry and exit relationships between adjacent intersections, the Lagrange multipliers corresponding to each intersection constraint are obtained. The corrected road traffic flow data is obtained based on each Lagrange multiplier and the corrected road traffic flow data function.
5. The privacy-preserving traffic flow data publishing method as described in claim 4, characterized in that, The corrected road traffic flow data function is expressed as follows: in, and Representing roads i The entrance and exit intersection numbers; and Representing roads i Noisy road traffic flow data and corrected road traffic flow data.
6. A traffic flow data publishing system with privacy protection as its core feature, characterized in that, The privacy-preserving traffic flow data publishing method as described in claim 1 is applied to a traffic flow data center, and the system includes: The constraint creation module is used to obtain the intersection models of each intersection within the target traffic area and establish corresponding intersection constraint conditions based on each intersection model. The data statistics module is used to receive the road information of each vehicle user in the target traffic area, and to count the road traffic flow data of each road in the target traffic area in real time based on the road information of each vehicle. The data noise-adding module is used to perform differential privacy noise-adding processing on traffic flow data of various roads to obtain the corresponding noisy road traffic flow data. The noise correction module is used to perform post-processing optimization on the noisy road traffic flow data according to the intersection constraints to obtain corrected road traffic flow data. The data publishing module is used to publish the corrected road traffic flow data to the outside world through the corresponding traffic flow database.
7. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.
8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.
Citation Information
Patent Citations
Time-space type trajectory big data differential privacy protection method
CN108763947A
Differential privacy-based vehicle trajectory data protection method and system, computer equipment and storage medium
CN115221557A
