Mobile crowd sensing task allocation method based on confusion privacy protection

Abstract

The invention discloses a mobile crowd sensing task allocation method based on confusion privacy protection. An MCS platform publicly releases a sensing task set; a user locally maintains a position application frequency statistical table and a total application frequency counter, the utility of tasks sent by a platform is determined, random sampling is carried out on a task set by adopting a differential privacy index mechanism based on the utility, the sampling probability of each task is obtained, and an initial application task is determined; the user adopts a Laplace mechanism to disturb the real distance of the initial application task to obtain a disturbed distance, and submits pre-application information to the platform; the platform summarizes pre-application information of all users, and counts and discloses the number of applicants of each task; and the user determines the expected income of the applied task, the task with the maximum expected income is screened out, the Laplace mechanism which is the same as the pre-application stage is adopted to carry out distance disturbance, the disturbed distance is obtained, and the MCS platform completes task distribution and determines the payment amount according to the disturbance information submitted by the user.

Description

Technical Field

[0001] This invention belongs to the technical field of data processing systems or methods specifically applicable to administrative, commercial, financial, management, supervision or forecasting purposes, and specifically relates to a mobile crowd-sensing task allocation method based on obfuscation-based privacy protection. Background Technology

[0002] With the widespread adoption of 5G / 6G wireless communication technologies and mobile smart terminals such as smartphones and wearable devices, Mobile Crowdsourcing (MCS) has evolved into a core data processing paradigm specifically suited for administrative, commercial, financial, management, supervisory, and forecasting purposes. By aggregating the sensing capabilities of massive numbers of user terminals, it provides low-cost, wide-coverage data support for multiple sectors. In the administrative sector, it can be used for urban governance, data collection for public services in remote areas, and tracking of policy implementation effectiveness; in the commercial sector, it can serve market research in lower-tier cities, targeted marketing, and monitoring of consumer trends; in the financial sector, it can support the assessment of credit demand in remote areas, cross-regional risk monitoring, and evaluation of financial service coverage; in the management sector, it can assist in optimizing resource allocation and managing cross-regional service efficiency; in the supervisory sector, it is suitable for environmental compliance verification and cross-regional regulatory data collection; and in the forecasting sector, it can achieve market demand forecasting and analysis of public service supply and demand trends based on data from across the entire region, becoming a key foundation for data-driven decision-making in various fields.

[0003] In an MCS system, task allocation is a core element in ensuring service quality and data value transformation. The MCS platform publishes geographically specific task awareness features, and users apply for tasks by submitting their distance information. The platform then allocates tasks to users based on the nearest proximity principle. While this mechanism improves task allocation efficiency, it exposes the following key issues when adapting to the application needs of administrative, commercial, and financial sectors:

[0004] First, location privacy leaks pose risks across multiple scenarios. Users must submit location-related distance information to the platform to apply for tasks. The platform, along with businesses and government departments, analyzes users' historical task application data, combined with auxiliary information such as community maps, transportation routes, and consumption records, to accurately infer users' active areas and frequent locations (e.g., residence, workplace, and frequently visited places). This privacy leak not only exposes users' sensitive personal information but can also be used for commercial harassment, financial credit discrimination, and excessive administrative intervention, seriously infringing on users' legitimate rights and hindering the widespread application of MCS technology in highly privacy-sensitive fields. Although differential privacy technology has been widely used for MCS privacy protection, most existing solutions only design indistinguishability constraints based on the distance dimension, failing to consider that users' preference for near-distance tasks leads to spatial concentration in historical application data. Platforms can still breach privacy protection boundaries through statistical analysis, failing to meet the stringent privacy protection requirements of administrative, commercial, and financial sectors. Second, task completion failures in sparse areas affect data service coverage. Users, driven by a rational choice to reduce execution costs, have a natural preference for tasks located near their homes. This results in extremely low application rates for tasks in sparsely populated areas such as suburbs and remote monitoring stations, often leading to situations where no one applies. This issue is particularly prominent in scenarios such as administrative supervision (e.g., environmental monitoring and rural public service data collection in remote areas), consumer behavior research in lower-tier markets, and credit demand assessment in remote areas. It directly leads to incomplete data collection and insufficient decision-making basis in these fields, severely impacting the service coverage and availability of the MCS system and limiting its deeper application in a wider range of scenarios.

[0005] Therefore, existing MCS task allocation schemes have significant limitations when adapting to specific data processing scenarios such as administration, commerce, finance, and management. Regarding privacy protection, they only focus on distance factors and fail to dynamically adjust protection strength based on the number of task applications, making them ineffective against inference attacks based on historical application data. Furthermore, the lack of a task guidance mechanism for sparse areas leads to a high failure rate for tasks in sparse regions. Some schemes introduce excessive noise to enhance privacy, resulting in distorted allocation decisions, while others sacrifice privacy protection strength to improve efficiency, leading to a difficulty in balancing privacy protection and allocation efficiency. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a mobile crowd-sensing task allocation method based on obfuscation privacy protection that significantly improves the accuracy of privacy protection and the efficiency of task allocation.

[0007] The technical solution adopted to solve the above-mentioned technical problems is: a mobile crowd-sensing task allocation method based on obfuscation-based privacy protection, comprising the following steps:

[0008] Step 1. The MCS platform publicly releases a set of perception tasks to be executed. It also provides the geographical location information for each task and builds a global location request count statistics table. ;

[0009] Step 2. The user maintains a location application count statistics table L and a total application count counter C on the local terminal to record the historical application count of the user in different location cells. Each time a new task is applied for, the application count for the corresponding location and the total application count are updated.

[0010] Step 3. The user determines the task issued by the MCS platform according to the following formula. The utility ,

[0011]

[0012] In the formula, For users and tasks The actual distance For the task The percentage of applications received based on location;

[0013] Step 4. User based on utility A differential privacy index mechanism is used to randomly sample the task set T to obtain the results for each task. sampling probability Determine the initial application task ;

[0014] Step 5. The user uses the Laplace mechanism to process the initial application task. The actual distance Perform a perturbation and obtain the distance after the perturbation. To prevent users from submitting their actual distances and revealing their location information;

[0015] Step 6. The user submits pre-application information to the platform. , A unique identifier for the user. To adjust the preferences, , This indicates acceptance of a secondary assignment if no initial task is received. This indicates that you do not accept it;

[0016] Step 7. The MCS platform aggregates all users' pre-application information, compiles statistics, and publishes the information for each task. Number of applicants ;

[0017] Step 8. The number of applicants based on the number published on the MCS platform. From oneself to the task The actual distance The application task is determined according to the following formula. Expected returns ,

[0018]

[0019] In the formula, As a measure of returns, ;

[0020] Step 9. The user iterates through all applications to ensure that the number of applicants meets the requirements. Task , For users to initially apply for tasks Based on the number of applicants, tasks with the highest expected returns are selected. and the task The distance is perturbed using the same Laplace mechanism as in the pre-application phase to obtain the perturbed distance. Submit the revised application information to the MCS platform. ;

[0021] Step 10. The MCS platform completes task allocation and determines payment amount based on the disturbance information submitted by the user.

[0022] Step 10.1. The MCS platform constructs a user set S to record the user allocation status, with the initial state of all users being unassigned; for each task... Collect users who have applied for the task, forming a user set B. Select the user with the smallest perturbation distance from user set B as the winner, and update user set S.

[0023] Step 10.2. The MCS platform counts the remaining set of tasks that have not been applied for. Statistics table based on global location application frequency The recorded user application history is sorted by application frequency from lowest to highest for the remaining tasks;

[0024] Step 10.3. For each remaining task Users who were not assigned a role and accepted reassignment were selected, resulting in a user set. The user set is determined according to the following formula. Comprehensive user metrics ,

[0025]

[0026] In the formula, These are the weighting coefficients. For the set of location cells covered by the MCS platform, For the remaining tasks The cell in question;

[0027] Step 10.4. Select the user with the lowest overall score as the winner and assign tasks accordingly. And update the user set S;

[0028] Step 10.5. For each winning user, the MCS platform determines the payment amount based on the perturbation distance they submitted.

[0029] As a preferred technical solution, in step 3, the task Percentage of applications based on location for:

[0030] .

[0031] As a preferred technical solution, in step 4, each task sampling probability for:

[0032] In the formula, Budget for privacy protection.

[0033] As a preferred technical solution, in step 5, the distance after disturbance for:

[0034]

[0035] In the formula, This is Laplace distributed noise.

[0036] As a preferred technical solution, in step 10.5, for each winning user, the MCS platform determines the payment amount based on the perturbation distance they submitted. , .

[0037] The beneficial effects of this invention are as follows:

[0038] This invention proposes a task request obfuscation-based privacy approach. By dynamically constraining the indistinguishability of probabilities when users choose different tasks, it prevents platforms from making privacy inferences solely based on distance. It not only considers users' preferences for nearby tasks but also incorporates historical information about task request frequency, making privacy protection more precise. This effectively prevents platforms from inferring users' permanent location from historical request data, significantly improving privacy protection and overcoming the shortcomings of traditional differential privacy schemes based solely on distance.

[0039] This invention allows users to maintain task application records locally, preventing the leakage of privacy information during network transmission. It employs a utility function that combines distance and the proportion of application frequency, satisfying users' preference for near-field tasks and providing a basis for guiding users to apply for tasks in sparse areas. Through differential privacy index sampling, tasks are randomly selected to ensure that the task selection probability satisfies obfuscation privacy. Furthermore, the real distance is perturbed by Laplace distance perturbation to further protect user location privacy.

[0040] This invention employs a user-subject adjustment mechanism based on expected returns to encourage users to actively apply for tasks in sparse regions, significantly improving the application and completion rates of such tasks and optimizing task allocation efficiency. A mechanism based on perturbation distance to determine payment amounts achieves full task allocation and non-negative returns for users, ensuring the long-term sustainability of the system. Simultaneously, global optimization reduces system operating costs. Attached Figure Description

[0041] Figure 1 This is a flowchart illustrating the mobile crowd-sensing task allocation method based on obfuscation-based privacy protection according to the present invention. Detailed Implementation

[0042] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments, but the present invention is not limited to the following embodiments.

[0043] exist Figure 1 In this embodiment, a mobile crowd sensing task allocation method based on obfuscation-based privacy protection includes the following steps:

[0044] Step 1. The MCS platform publicly releases the perception tasks to be executed.

[0045] The MCS platform publicly releases a set of perception tasks to be executed. It also provides the geographical location information for each task and builds a global location request count statistics table. ;

[0046] Step 2. The user maintains a location application count statistics table L and a total application count counter C on the local terminal to record the historical application count of the user in different location cells. Each time a new task is applied for, the application count for the corresponding location and the total application count are updated.

[0047] Step 3. The user determines the task issued by the MCS platform according to the following formula. The utility ,

[0048]

[0049] In the formula, For users and tasks The actual distance For the task The percentage of applications based on location

[0050] Among them, the task Percentage of applications based on location for:

[0051] .

[0052] Step 4. User based on utility A differential privacy index mechanism is used to randomly sample the task set T to obtain the results for each task. sampling probability Determine the initial application task ;

[0053] Sample probability for:

[0054]

[0055] In the formula, Budget for privacy protection;

[0056] Step 5. The user uses the Laplace mechanism to process the initial application task. The actual distance Perform a perturbation and obtain the distance after the perturbation. , , To use Laplace-distributed noise, we prevent the real distance submitted by the user from revealing location information.

[0057] Step 6. The user submits pre-application information to the platform. , A unique identifier for the user. To adjust the preferences, , This indicates acceptance of a secondary assignment if no initial task is received. This indicates that you do not accept it;

[0058] Step 7. The MCS platform aggregates all users' pre-application information, compiles statistics, and publishes the information for each task. Number of applicants ;

[0059] Step 8. The number of applicants based on the number published on the MCS platform. From oneself to the task The actual distance The application task is determined according to the following formula. Expected returns ,

[0060]

[0061] In the formula, As a measure of returns, ;

[0062] Step 9. The user iterates through all applications to ensure that the number of applicants meets the requirements. Task , For users to initially apply for tasks Based on the number of applicants, tasks with the highest expected returns are selected. and the task The distance is perturbed using the same Laplace mechanism as in the pre-application phase to obtain the perturbed distance. Submit the revised application information to the MCS platform. ;

[0063] Step 10. The MCS platform completes task allocation and determines payment amount based on the disturbance information submitted by the user.

[0064] Step 10.1. The MCS platform constructs a user set S to record the user allocation status, with the initial state of all users being unassigned; for each task... Collect users who have applied for the task, forming a user set B. Select the user with the smallest perturbation distance from user set B as the winner, and update user set S.

[0065] Step 10.2. The MCS platform counts the remaining set of tasks that have not been applied for. Statistics table based on global location application frequency The recorded user application history is sorted by application frequency from lowest to highest for the remaining tasks;

[0066] Step 10.3. For each remaining task Users who were not assigned a role and accepted reassignment were selected, resulting in a user set. The user set is determined according to the following formula. Comprehensive user metrics ,

[0067]

[0068] In the formula, These are the weighting coefficients. For the set of location cells covered by the MCS platform, For the remaining tasks The cell in question;

[0069] Step 10.4. Select the user with the lowest overall score as the winner and assign tasks accordingly. And update the user set S;

[0070] Step 10.5. For each winning user, the MCS platform determines the payment amount based on the perturbation distance they submitted. , .

Claims

1. A method for assigning mobile crowd-sensing tasks based on obfuscation-based privacy protection, characterized in that, Includes the following steps: Step 1. The MCS platform publicly releases a set of perception tasks to be executed. It also provides the geographical location information for each task and builds a global location request count statistics table. ; Step 2. The user maintains a location application count statistics table L and a total application count counter C on the local terminal to record the historical application count of the user in different location cells. Each time a new task is applied for, the application count for the corresponding location and the total application count are updated. Step 3. The user determines the task issued by the MCS platform according to the following formula. The utility , In the formula, For users and tasks The actual distance For the task The percentage of applications received based on location; Step 4. User based on utility A differential privacy index mechanism is used to randomly sample the task set T to obtain the results for each task. sampling probability Determine the initial application task ; Step 5. The user uses the Laplace mechanism to process the initial application task. The actual distance Perform a perturbation and obtain the distance after the perturbation. To prevent users from submitting their actual distances and revealing their location information; Step 6. The user submits pre-application information to the platform. , A unique identifier for the user. To adjust the preferences, , This indicates acceptance of a secondary assignment if no initial task is received. This indicates that you do not accept it; Step 7. The MCS platform aggregates all users' pre-application information, compiles statistics, and publishes the information for each task. Number of applicants ; Step 8. The number of applicants based on the number published on the MCS platform. From oneself to the task The actual distance The application task is determined according to the following formula. Expected returns , In the formula, As a measure of returns, ; Step 9. The user iterates through all applications to ensure that the number of applicants meets the requirements. Task , For users to initially apply for tasks Based on the number of applicants, tasks with the highest expected returns are selected. and the task The distance is perturbed using the same Laplace mechanism as in the pre-application phase to obtain the perturbed distance. Submit the revised application information to the MCS platform. ; Step 10. The MCS platform completes task allocation and determines payment amount based on the disturbance information submitted by the user. Step 10.

1. The MCS platform constructs a user set S to record the user allocation status, with the initial state of all users being unassigned; for each task... Collect users who have applied for the task, forming a user set B. Select the user with the smallest perturbation distance from user set B as the winner, and update user set S. Step 10.

2. The MCS platform counts the remaining set of tasks that have not been applied for. Statistics table based on global location application frequency The recorded user application history is sorted by application frequency from lowest to highest for the remaining tasks; Step 10.

3. For each remaining task Users who were not assigned a role and accepted reassignment were selected, resulting in a user set. The user set is determined according to the following formula. Comprehensive user metrics , In the formula, These are the weighting coefficients. For the set of location cells covered by the MCS platform, For the remaining tasks The cell in question; Step 10.

4. Select the user with the lowest overall score as the winner and assign tasks accordingly. And update the user set S; Step 10.

5. For each winning user, the MCS platform determines the payment amount based on the perturbation distance they submitted.

2. The mobile crowd-sensing task allocation method based on obfuscation-based privacy protection according to claim 1, characterized in that, In step 3, the task Percentage of applications based on location for: 。 3. The mobile crowd-sensing task allocation method based on obfuscation-based privacy protection according to claim 1, characterized in that, In step 4, each task sampling probability for: In the formula, Budget for privacy protection.

4. The mobile crowd-sensing task allocation method based on obfuscation-based privacy protection according to claim 1, characterized in that, In step 5, the distance after disturbance for: In the formula, This is Laplace distributed noise.

5. The mobile crowd-sensing task allocation method based on obfuscation-based privacy protection according to claim 1, characterized in that, In step 10.5, for each winning user, the MCS platform determines the payment amount based on the perturbation distance they submitted. , .

Images