Trusted data acquisition method for bilateral location privacy protection

Abstract

The invention discloses a trusted data acquisition method for bilateral position privacy protection. According to the method, an effective solution is provided for solving the problems that a platform is difficult to recruit appropriate workers for distance and position related tasks under worker and task position privacy protection, and low-credibility even malicious workers in a network are identified. The method comprises the following steps: firstly, protecting position privacy of a worker and a task through matrix disturbance encryption, and calculating a distance closeness degree of the worker and the task through encryption; and then the truth is calculated through the data of the known completely-trusted workers and the highly-trusted workers, the data of the unknown workers are verified in combination with the truth, the trust degree of the unknown workers is calculated, and the highly-trusted workers and the malicious workers are identified. Therefore, after the workers are identified, the workers with high credibility and proper proximity degree can be recruited for each task, so as to realize credible data acquisition of bilateral position privacy protection.

Description

Technical Field

[0001] The present invention relates to the field of data collection for crowd-sensing, and in particular to a method for achieving high-quality and low-cost data collection under the premise of bilateral privacy protection in a crowd-sensing system. Background Art

[0002] Mobile crowd sensing is an emerging data collection method that collects and analyzes environmental data by integrating a large number of widely distributed mobile devices (such as smartphones) into a sensing network. In crowd sensing, data requesters need to build services and publish tasks to collect data through the platform, while workers carrying mobile sensing devices are paid by participating in tasks. Due to the large number of data participants, the crowd sensing network can obtain data for a long time, over a large area, and at a low cost. Typical application scenarios include air quality monitoring, noise detection, and street view mapping.

[0003] Since the data submitted by data participants directly affects the quality of platform construction applications, selecting credible workers to collect high-quality data is the key to the crowd-sensing network. Since workers' perception of data consumes resources such as computing, storage, communication resources, and time, and the higher the perception quality, the more resources and time they pay. However, there are some low-credibility or even malicious participants in the crowd-sensing network. These participants want to pay as little as possible and earn the maximum reward. In order to reduce costs, these workers often directly fabricate false data to the platform, so that they can get a large reward with the least cost. Malicious participants not only want to get a large reward, but also use the reported malicious data to attack the platform, causing the system to suffer greater losses than false data. The previous method believed that the platform can obtain the trust of the workers after receiving the workers' data, and it only needs to directly select high-credibility workers. But in fact, after the workers submit the data, the platform often finds it difficult to distinguish the quality of the workers' data. Therefore, it is necessary to execute the truth discovery algorithm through the workers' data to obtain the truth before evaluating the authenticity of the data and obtaining the trust of the workers.

[0004] In addition to worker trust evaluation, bilateral location privacy protection has gradually become an important issue that cannot be ignored in crowd sensing systems. Specifically, for task requesters, when publishing tasks, they do not want the location of the task to be disclosed to the platform, which may bring a series of negative effects, such as exposing environmental problems in a specific area, thereby affecting local economic activities. For workers participating in the task, they do not want their location information to be disclosed because this may expose the worker's personal privacy information, such as their place of residence, place of work, etc. However, for the platform, selecting workers to perform tasks may also require the worker and the task location to calculate the distance between them, especially for location-sensitive tasks (such as sound and image acquisition), the data quality is usually closely related to the distance between the worker and the task. Therefore, the present invention proposes a trusted data collection method with bilateral location privacy protection, which allows the platform to achieve high-quality worker and task matching recruitment while protecting the bilateral location privacy of workers and tasks. Summary of the invention

[0005] The present invention discloses a trusted data collection method with bilateral location privacy protection. Aiming at the situation that low-quality workers and malicious workers in the crowd intelligence perception network reduce the quality of network services and the bilateral privacy of workers and tasks needs to be protected, the inventive method proposes an effective worker trust evaluation method and a distance proximity calculation method under bilateral location privacy protection, which effectively improves the credibility and quality of collected data while protecting the privacy of workers and tasks. Because in the crowd intelligence perception network, there are some malicious workers who may fabricate false data to reduce the quality of building applications, but it is difficult for the existing platform to directly verify the authenticity and credibility of the data, so that workers with high trust cannot be identified. At the same time, because the privacy of tasks and workers' locations is unwilling to be disclosed, for some distance-related tasks, it is difficult for the platform to calculate the proximity of tasks and workers and cannot select suitable workers. Therefore, the present invention proposes a trusted data collection method with bilateral location privacy protection, firstly, a matrix perturbation method is used to protect location privacy, and the proximity of workers and tasks is effectively calculated, and then, through the data aggregated by some partially known fully trusted workers and the data discovered by highly trusted workers, the data of unknown workers is effectively verified, and trusted workers are identified, so as to finally realize high-quality and highly trusted data collection under the protection of unknown privacy of workers and tasks.

[0006] The technical solution of the inventive method is as follows:

[0007] 1. A trusted data collection method with bilateral location privacy protection, characterized by comprising the following steps:

[0008] (1) The crowd-sensing system mainly includes task requesters, participating workers, an intermediary platform, and a trusted third-party platform; the task set of the requester is defined as R = {r1, r2, …, r M}, each task will last for several rounds, and the set of workers is defined as W = {w1,w2,…,w N}, the locations of tasks and workers fall into a k×k grid; the platform initially knows the trustworthiness of some workers, where Represents worker w i Trust in round t; The platform puts the workers into a completely trusted gold worker collection. i ∈S Gold ;for and The worker platform puts it into a basically trusted silver worker collection, w i ∈S Silver ; When the task requester needs to publish a task, the following method is used to achieve trusted data collection with bilateral location privacy protection:

[0009] (2) Task requesters and workers protect location privacy through the following methods:

[0010] First, the trusted third party generates a set of k×k position perturbation matrices for each worker and task. and Distribute to workers w i and task r j The requester will also randomly generate an α(α>k 2 ) is sent to all requesters and workers, where E represents the identity matrix;

[0011] In order to prevent the platform from knowing the location of the task, the task requester will represent the location of the task in the grid (p, q) as a k×k location matrix And the value at position (x,y) in the matrix is ​​assigned to α 2k-2-|x-p|-|y-q| , and encrypt it with the perturbation matrix Then send it to the platform; workers will also generate a position matrix like the requester And perturb encryption Send to the platform;

[0012] (3) The platform calculates the proximity between workers and tasks and selects workers using the following method:

[0013] In each round, for each encrypted task and worker location received and After that, the platform first calculates the relative position matrix R as follows

[0014]

[0015] Then calculate the proximity between the worker and the task ξ i,j=Trace(R), Trace(·) means finding the trace of the matrix; after obtaining the relative distance, the platform selects workers by calculating the service quality of the workers. The service quality is calculated as follows:

[0016]

[0017] After calculating the service quality, for each task, the platform will select the top K workers to participate in the task and collect data based on the service quality;

[0018] (4) The platform calculates the truth data according to the following method:

[0019] In round t, for task r j , each worker w participating in the task i The collected data are expressed as According to the above platform's classification of workers, if for task r j If there are gold workers participating in the task, the platform will directly aggregate the truth based on the gold workers' data.

[0020]

[0021] If there is no gold worker, the truth discovery algorithm will be executed through the silver worker data. The platform will continue to calculate the following two formulas until the truth error of the two calculations is less than a value to obtain the final truth:

[0022]

[0023] (5) The platform’s trust in workers is updated as follows:

[0024] Task j The final truth obtained in round t For the workers w involved in this task i Collected data Then the worker's trust is updated as follows:

[0025] if Then it will increase the trust of workers β represents the maximum allowed error, and ρ represents the update rate step size.

[0026] if Then it will increase the trust of workers β represents the maximum allowed error, and ρ represents the update rate step size.

[0027] Subsequently, each updated worker is reclassified into gold workers, silver workers, and untrusted workers based on the threshold.

[0028] Beneficial Effects

[0029] The present invention discloses a trusted data collection method with bilateral location privacy protection. In a crowd-sensing network, there are many tasks that are related to distance and location. The data collection quality of these tasks is correlated to the distance between the worker and the task. At the same time, workers and requesters have important requirements for location privacy protection, and it is difficult for the platform to recruit suitable workers for these location and distance-related tasks. Our proposed method proposes an effective location privacy protection method to address this problem. At the same time, under privacy protection, the platform can also effectively balance the trust of workers and the proximity to the task to match the appropriate task for high-quality data collection. , BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 : Total data utility comparison;

[0031] Figure 2 : Estimate the law of relative distance change; DETAILED DESCRIPTION

[0032] In order to facilitate the understanding of the present invention, the present invention will be described in more comprehensive and detailed manner below in conjunction with the accompanying drawings and preferred embodiments of the specification, but the protection scope of the present invention is not limited to the following specific embodiments.

[0033] Unless otherwise defined, all professional terms used below have the same meanings as those generally understood by those skilled in the art. The professional terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the scope of protection of the present invention.

[0034] Unless otherwise specified, various raw materials, reagents, instruments and equipment used in the present invention can be purchased from the market or prepared by existing methods.

[0035] Example:

[0036] Real-time monitoring of urban environment, such as air quality, noise level, traffic flow, etc., is one of the main application scenarios of crowd sensing systems. To this end, the platform usually attracts widely distributed mobile device holders in the city to participate in data perception by issuing tasks and giving certain rewards. However, since the perception quality of workers in the crowd sensing system is unknown in advance, it is necessary to evaluate the perception quality of workers. In addition, it is also necessary to consider the cost of workers performing tasks and the distance of workers moving. However, while obtaining the distance factor, the location privacy of workers and tasks may be leaked. Therefore, the present invention discloses a trusted data collection method with bilateral location privacy protection to achieve high-quality and low-cost data collection under privacy protection conditions. The experimental method is as follows: Assume that there are 100 workers in the city, who undertake 30 perception tasks, the number of workers required for each task is 10, and the budget for each task ranges from 6000 to 12000. The α parameter of the position matrix of workers and tasks is set to 15, and the real matrix coordinates of each worker and task are randomly generated. In addition, the workers' perception abilities are evenly distributed in four intervals: [0, 0.25], [0.25, 0.5], [0.5, 0.75], [0.75, 1]. The perception ability of each worker determines the accuracy of the data reported. The higher the perception ability, the closer the reported data is to the true value. The cost of each worker is randomly generated within (0, 1). When the platform releases the task information, all workers respond to the perception task in the same proportion. According to the method of the present invention, the platform selects workers according to the quality-cost index, and then evaluates the data trust of the participating workers and updates the workers' estimated perception quality.

[0037] The experimental results of the inventive method are given below.

[0038] Figure 1 The total data utility of the method of the present invention and other baseline comparison methods under different task budgets are given. Specifically, the task budget is gradually increased from 6000 to 12000, with an increase of 2000 each time. The results show that the total data utility of the method of the present invention is always the highest except for the theoretical optimal method, and always has the lowest regret, reflecting the excellent performance of the method of the present invention.

[0039] Figure 2 The accuracy of the method of the present invention in estimating the relative distance between the worker and the task is given, where the symbol d represents the actual distance between the worker and the task. The results show that the method of the present invention can accurately estimate the actual distance within a small and certain error range, proving the superiority of the method of the present invention.

Claims

1. A trusted data collection method with bilateral location privacy protection, characterized in that The following steps are involved: (1) The crowd-sensing system mainly includes task requesters, participating workers, an intermediary platform, and a trusted third-party platform; the task set of the requester is defined as R = {r1, r2, …, r M }, each task will last for several rounds, and the set of workers is defined as W = {w1,w2,…,w N }, the locations of tasks and workers fall into a k×k grid; the platform initially knows the trustworthiness of some workers, where Represents worker w i Trust in round t; The platform puts the workers into a completely trusted gold worker collection. i ∈S Gold ;for and The worker platform puts it into a basically trusted silver worker collection, w i ∈S Silver ; When the task requester needs to publish a task, the following method is used to achieve trusted data collection with bilateral location privacy protection: (2) Task requesters and workers protect location privacy through the following methods: First, the trusted third party generates a set of k×k position perturbation matrices for each worker and task and Distribute to workers w i and task r j The requester will also randomly generate an α(α>k 2 ) is sent to all requesters and workers, where E represents the identity matrix; In order to prevent the platform from knowing the location of the task, the task requester will represent the location of the task in the grid (p, q) as a k×k location matrix And the value at position (x,y) in the matrix is ​​assigned to α 2k-2-|x-p|-|y-q| , and encrypt it with the perturbation matrix Then send it to the platform; the worker will also generate a position matrix like the requester And perturb encryption Send to the platform; (3) The platform calculates the proximity between workers and tasks and selects workers using the following method: In each round, for each encrypted task and worker location received and After that, the platform first calculates the relative position matrix R as follows Then calculate the proximity between the worker and the task ξ i,j =Trace(R), Trace(·) means finding the trace of the matrix; after obtaining the relative distance, the platform selects workers by calculating the service quality of the workers. The service quality is calculated as follows: After calculating the service quality, for each task, the platform will select the top K workers to participate in the task and collect data based on the service quality; (4) The platform calculates the truth data according to the following method: In round t, for task r j , each worker w participating in the task i The collected data are expressed as According to the above platform's classification of workers, if for task r j If there are gold workers participating in the task, the platform will directly aggregate the truth based on the gold workers' data. If there is no gold worker, the truth discovery algorithm will be executed through the silver worker data. The platform will continue to calculate the following two formulas until the truth error of the two calculations is less than a value to obtain the final truth: (5) The platform’s trust in workers is updated as follows: Task j The final truth obtained in round t For the workers w involved in this task i Collected data Then the worker's trust is updated as follows: if Then it will increase the trust of workers β represents the maximum allowed error, and ρ represents the update rate step size. if Then it will increase the trust of workers β represents the maximum allowed error, and ρ represents the update rate step size. Subsequently, each updated worker is reclassified into gold workers, silver workers, and untrusted workers based on the threshold.

Images