Trusted worker recruitment method based on identity privacy protection

Abstract

The invention discloses a trusted worker recruitment method based on identity privacy protection, and relates to the field of data collection of trusted computing and quality assurance of a crowd-sourcing network. The network structure is composed of a data requester, a server, a reputation authentication center TA for short and a worker. A worker obtains a pseudonym by interacting with the TA in advance, and the pseudonym is used in the data interaction process to protect privacy. The reputation information of the worker is stored in the TA, and the encrypted reputation is carried when the task is applied; then, the server adopts a two-party reputation verification process based on zero-knowledge proof, so that the server can verify whether the reputation submitted by the worker is true or not to the TA under the condition of not knowing the reputation value of the worker; and the server selects N workers to collect data, after data collection, the server calculates a truth value by adopting a truth value calculation method to obtain a truth value, sends the truth value to the data request, carries out comparison calculation on the creditworthiness of the workers, and updates the creditworthiness of the workers through the TA.

Description

Technical Field

[0001] The present invention relates to the field of trusted computing and quality assurance data collection in a crowd intelligence network, and more particularly to a trusted worker recruitment method based on identity privacy protection. Background Art

[0002] As an emerging data collection technology, Mobile Crowdsensing (MCS) utilizes mobile devices equipped with sensors and a large number of user groups to complete sensing tasks, and has shown great application potential in many fields such as military, industry, life health, and environmental monitoring. However, the wide application of crowdsensing technology faces problems such as privacy protection and low data quality.

[0003] In a mobile crowdsensing network, the data sensing process is as follows: The data requester, abbreviated as DR, publishes the sensing data task to be done to the platform; Workers view the published tasks through the platform; If a worker is willing to execute the task, the worker applies to the platform to do the task; The platform selects workers and notifies the selected workers to execute the task; After the workers execute the task, they submit the sensed data to the platform; The platform generally uses the CRH method to calculate the true value and sends the result of the true value calculation to the DR, thus completing the entire data collection process.

[0004] However, the above general task collection method for a mobile crowdsensing network has the following deficiencies: First, there is no privacy protection, and the information of workers is directly exposed. Workers are reluctant to participate in data collection due to concerns about their privacy, which will thus hinder the development of the crowdsensing network. The second deficiency is that: In the currently commonly used CRH method in the data collection method, N workers are recruited to execute a task to obtain N data. Then, these N data are weighted to obtain the true value. The basic idea of true value calculation is that the data submitted by workers that is closer to the true value has a greater weight, and vice versa, the weight is smaller. The idea of the CRH method is that most workers in the network are trusted, so the data they submit is real, and thus is closer to the finally calculated true value. Therefore, the weight of trusted workers is large, while the data reported by dishonest or malicious workers is far from the true value, and its weight is small. In this way, the result after weighted calculation is close to the actual true value. However, the data quality obtained by this method is not necessarily good. In this method, as long as there is false data among the N data, the data quality will be reduced, and the more false data there is, the lower its quality. When multiple malicious workers collude and conspire to attack, the attacker can obtain the result they want, so the CRH method completely fails.

[0005] To change this situation, some researchers have proposed a worker recruitment method based on reputation selection. The idea of this method is to evaluate the reputation of workers; then, select workers with high reputation to obtain high-quality data. However, the challenge of this method is that the reputation of workers is private information of workers, and workers are not willing to let the platform know their reputation information. Therefore, it is a great challenge to collect data while protecting trust privacy, and there is no good solution currently.

[0006] The existing technologies have deficiencies in simultaneously meeting high data quality and worker reputation privacy protection. Therefore, there is an urgent need for an inventive method that can comprehensively solve the above problems and can well solve the above problems. Summary of the Invention

[0007] In view of this, the present invention provides a trusted worker recruitment method based on identity privacy protection to solve the problems in the background technology.

[0008] To achieve the above object, the present invention adopts the following technical solutions:

[0009] A trusted worker recruitment method based on identity privacy protection, characterized by including the following steps:

[0010] The data requester submits a task request to the cloud server;

[0011] The cloud server publishes the task to the workers;

[0012] Workers who hope to participate in the task register their identities with the trust certification center. When registering, the workers generate an elliptic curve encrypted key pair locally; after retaining the private key, they send the public key to the certification center, and the certification center will generate a pseudonym for them according to the public key of the workers;

[0013] The certification center sends the pseudonym to the corresponding workers;

[0014] The workers apply for sensing tasks to the cloud server using their own pseudonyms and reputation values. The reputation values of the workers are obtained by querying the certification center based on their identities;

[0015] The cloud server interacts with the certification center to verify whether the reputation values of the workers have been tampered with. Only the workers who pass the verification will be recruited;

[0016] The cloud server selects N workers from the workers who pass the verification to perform the sensing task and notifies the selected workers;

[0017] The selected workers sense the data and submit the data to the cloud server;

[0018] The cloud server performs true value calculation and reputation evaluation on the sensing data submitted by the workers;

[0019] The cloud server sends the true value to the data requester and sends the reputation value evaluation result to the certification center;

[0020] The certification center updates the reputation value of the worker.

[0021] Optionally, a pseudonym refers to a technology for protecting identity privacy that replaces personal identity information with a pseudo-identifier, ensuring that the data subject cannot be identified through the data without using other additional information; the worker generates a key pair encrypted by an elliptic curve locally, where the public key is Pk i , and the private key is Sk i ; when registering, the worker sends the public key Pk i to the certification center; use the SHA1 algorithm as the hash function, denoted as Hash(·). Before processing by the hash algorithm, first add a unique salt value salt i for each worker ω i , and the salt value is a randomly generated string salt of any length l i ←Rand(l). After concatenating the salt value as an attribute to the worker's public key Pk i , then perform hash processing to calculate the pseudonym identifier p i of the worker = Hash(Pk i ||salt i ), where "||" represents the string concatenation operation; finally, the worker uses the pseudonym p i to apply to the cloud server to participate in the sensing task; since the pseudonym of each worker is generated by its unique salt value, and the cloud server cannot access the original Pk i and salt i , thus preventing the cloud server or the data requester from linking to the specific user identity through the sensing data.

[0022] Optionally, a two-party reputation verification process based on zero-knowledge proof is adopted, enabling the cloud server to verify with the certification center whether the reputation submitted by the worker is true without obtaining the true reputation value of the worker;

[0023] Use zero-knowledge proof to verify whether the reputation value of the worker is equal to the reputation value stored in the certification center, that is, whether the reputation value of the worker has been tampered with;

[0024] The prover proves to the verifier that the reputation value t′ owned by the prover is the same as the reputation value t owned by the verifier; at the same time, the prover will not disclose t′ to the verifier during this process; according to the reputation value, the verifier measures whether the reputation value of each prover passes the verification through the formula υ = ν(t, t′).

[0025] Optionally, the trusted center acts as a verifier, while the worker is a prover. The worker proves to the server that the reputation value it holds is valid. The commitment is used to prevent the cloud server from knowing the reputation value, thus protecting the privacy of the worker's reputation value. The specific process is as follows:

[0026] G is a cyclic group of prime order p, and g1 is an element of, and g2, h1, and h2 are elements generated by g1. Based on the discrete logarithm of h1, the discrete logarithm of g1, the discrete logarithm of h2, and the discrete logarithm of g2, the worker cannot know any of them; assume that t is the worker's reputation value stored by the trusted center, and t′ is the reputation value submitted by the worker; l, s1, and s2 are three security parameters;

[0027] A worker willing to participate in the sensing task first selects three random integers ω, λ1, λ2, and the three satisfy ω ∈ [1, 2 l ·p + 1], and then calculates Through and to obtain Then, the worker calculates D = ω + H·t′, D1 = λ1 + H·r1, and D2 = λ2 + H·r2; Hash(·) is a hash function; finally, the worker sends H, D, D1, D2, C m ′ to the cloud server;

[0028] After receiving the data from the worker side, the cloud server obtains the commitment from the trusted center where t is the worker's reputation value stored by the trusted center and

[0029] The cloud server verifies whether the reputation value has been tampered with by checking whether H is equal to ; if they are equal, the worker's reputation value has not been tampered with and the worker passes the verification; otherwise, the verification fails and the worker will not be recruited by the server to participate in the subsequent steps; for the convenience of the following description, if u k passes the trust verification algorithm, then υ k = ν(t, t′) = 1; otherwise, υ k = ν(t, t′) = 0; only the workers who pass the trust verification will be recruited to participate in the subsequent truth value calculation.

[0030] Optionally, a truth discovery and reputation evaluation method based on CRH is adopted to improve data quality by iteratively optimizing the data source weights and task truth values;

[0031] The true value calculation process is a process of iteratively calculating the estimated true value of the data. At the beginning of the iteration, the average value of the data collected for the current task is first calculated. Where Q represents the total number of workers who submitted this task data; then the standard deviation of the current task data is calculated based on this average value Calculate the data submitted by a single worker and the estimated true value x of the data through standard deviation * The standard error between At the beginning of the iteration, set To obtain the standard error of the task data submitted by each worker; then, based on the data submitted by the worker and the estimated true value x * The standard error between them is used to calculate the data weight of the data submitted by the workers. Bring the worker's data weight into x * The calculation formula Thus, a new round of estimated true value is obtained; this iterative process is repeated until the estimated true value x * The iteration ends after the value of converges, and then observes the data weights of the workers at the time of convergence, selects the k workers with higher data weights, marks their performance as credible in this round, adopts their data and pays them; the cloud server sends the data weights of the workers to the authentication center; the authentication center divides the workers into three categories according to their reputation values: malicious, unknown and credible. If the reputation value of the worker is greater than the high threshold T h , then it belongs to the trusted set; if the reputation value of the worker is in T h With T l If the reputation value is less than the lower threshold T l , it belongs to the malicious set; the authentication center selects a trusted worker from among them, assuming that there are K trusted workers in total, and takes the average of their data weights Obtaining task scores for unknown workers If the worker's task score ρ i Greater than or equal to the task score threshold ρ △ , that is, i ≥ρ △ , then through t new =min{(1-μ)ρ+ηt,1} to calculate the new reputation value after the increase; otherwise, if ρ i < △ , then through t new =max{-(1-μ)(1-ρ i )+ηt,0} to calculate the new reduced reputation value, where η is called the rate coefficient, which is used to control the influence of the historical reputation value on the new reputation value; after the reputation value of a worker is updated, the certification center classifies it into the corresponding category based on the updated reputation value.

[0032] As can be seen from the above technical solutions, compared with the prior art, the present invention provides a trusted worker recruitment method based on identity privacy protection, which performs excellently in terms of data quality and obtains the lowest mean absolute error and the lowest root mean square error in experiments. The reduction amplitudes are 21.75% and 80.17% respectively. It also has good performance in worker reputation identification and has the function of protecting worker reputation privacy. In summary, the present invention realizes privacy protection, improves data quality, provides an efficient, reliable and innovative solution for the wide application of the crowd wisdom network, and has important practical application value and promotion prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0034] Figure 1 System network composition diagram of the invention method;

[0035] Figure 2 Situation of obtaining the mean absolute error and root mean square error of data;

[0036] Figure 3 Change situation of the credibility of different workers. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present invention.

[0038] Embodiment 1

[0039] The embodiment of the present invention discloses a trusted worker recruitment method based on identity privacy protection, including the following steps:

[0040] Step 1: The data requester, namely the Data Requester (DR for short), submits a task request to the cloud server, namely the Cloud Server (CS for short).

[0041] Step 2: The CS publishes the task to the workers.

[0042] Step 3: Workers who wish to participate in the task register their identities with the Trust Authority (TA). When registering, the worker generates a key pair encrypted by an elliptic curve locally. After retaining the private key, the worker sends the public key to the TA, and the TA generates a pseudonym for the worker based on the public key. The specific process is described in Article 2;

[0043] Step 4: The TA sends the pseudonym to the corresponding worker;

[0044] Step 5: The worker applies to the CS for a sensing task using their own pseudonym and reputation value. The worker's reputation value can be obtained by querying the TA with their own identity;

[0045] Step 6: The CS interacts with the TA to verify whether the worker's reputation value has been tampered with. Only workers who pass the verification will be recruited. The specific process of the credibility verification is described in Article 3;

[0046] Step 7: The TA returns the result of the reputation value verification to the CS;

[0047] Step 8: The CS selects N workers from the workers who pass the verification for the sensing task and notifies the selected workers;

[0048] Step 9: The selected workers sense the data and submit the data to the CS;

[0049] Step 10: The CS performs true value calculation and reputation evaluation on the sensing data submitted by the workers. The specific process of the true value discovery and reputation evaluation method is described in Article 4;

[0050] Step 11: The CS sends the true value to the DR and sends the reputation value evaluation result to the TA. There are only two types of changes in the reputation value: increase and decrease.

[0051] Step 12: The TA updates the reputation value of the worker.

[0052] Furthermore, this embodiment uses a salted hash algorithm to implement pseudonymization to enhance the unpredictability of traditional hash pseudonyms, prevent rainbow table and duplicate password attacks, thereby enhancing the privacy protection level of the worker's identity;

[0053] Pseudonymization refers to a technique for protecting identity privacy that replaces personal identity information with a pseudo-identifier to ensure that the data subject cannot be identified through the data without using other additional information. The worker generates a key pair encrypted by an elliptic curve locally, where the public key is Pk i , and the private key is Sk i . When registering, the worker sends the public key Pk i to the TA. The present invention uses the SHA1 algorithm as the hash function (denoted as Hash(·)). Before processing by the hash algorithm, first for each worker ω iAdd a unique salt value salt i , where the salt value is a randomly generated string salt of any length l i ←Rand(l), and use this salt value as an attribute to concatenate to the worker's public key Pk i After that, perform a hash operation to calculate the worker's pseudonym identifier p i =Hash(Pk i ||salt i )

[0054] , where "||" represents the string concatenation operation. Finally, the worker uses the pseudonym p i to apply to the cloud server to participate in the sensing task. Since each worker's pseudonym is generated from its unique salt value, and the CS cannot access the original Pk i and salt i , thus preventing the CS or DR from linking to the specific user identity through the sensing data.

[0055] Furthermore, this embodiment adopts a two-party reputation verification process based on zero-knowledge proof, enabling the CS to verify with the TA whether the reputation submitted by the worker is true without obtaining the true value of the worker's reputation;

[0056] Zero-knowledge proof is a cryptographic verification mechanism that allows a prover to prove to a verifier that the prover knows a message m using an encryption commitment scheme without revealing any other information about m. In the present invention, we use zero-knowledge proof to verify whether the reputation value of the worker is equal to the reputation value stored in the TA, that is, whether the reputation value of the worker has been tampered with.

[0057] The prover proves to the verifier that the reputation value t' possessed by the prover is the same as the reputation value t possessed by the verifier. At the same time, the prover does not reveal t' to the verifier during this process. According to the reputation value, the verifier measures whether the reputation value of each prover passes the verification through the formula υ = ν(t, t').

[0058] In the present invention, the trusted center acts as the verifier, while the worker is the prover. The worker needs to prove to the server that the possessed reputation value is valid. The Pedersen commitment can prevent the cloud server from knowing the reputation value, thus protecting the privacy of the worker's reputation value. The specific process is as follows.

[0059] Step 1. G is a cyclic group with a prime order p, g1 is an element of , g2, h1, and h2 are elements generated by g1. Based on the discrete logarithm of h1, the discrete logarithm of g1, the discrete logarithm of h2, the discrete logarithm of g2, the worker cannot know any of them. Assume that t is the worker's reputation value stored in the trusted center, and t' is the reputation value submitted by the worker (which may be tampered with). l, s1, and s2 are three security parameters.

[0060] Step 2. Workers willing to participate in the sensing task first select three random integers ω, λ1, λ2, where the three satisfy ω ∈ [1, 2 l ·p + 1], and then calculate through and to obtain Then, the worker calculates '

[0061] D = ω + H·t, D1 = λ1 + H·r1 and D2 = λ2 + H·r2. Hash(·) is a hash function. Finally, the worker sends H, D, D1, D2, C m ′ to the cloud server.

[0062] Step 3. After receiving the data from the worker side, the cloud server obtains the commitment from the trusted center where t is the reputation value of the worker stored in the trusted center and

[0063] Step 4. The cloud server verifies whether the reputation value has been tampered with by checking whether H is equal to If they are equal, the reputation value of the worker has not been tampered with and the worker passes the verification. Otherwise, the verification fails and the worker will not be recruited by the server to participate in the subsequent steps. For the convenience of later description, if u k passes the trust verification algorithm, then υ k = ν(t, t′) = 1; otherwise, υ k = ν(t, t′) = 0. Only workers who pass the trust verification will be recruited to participate in the subsequent true value calculation.

[0064] Furthermore, this embodiment adopts a truth discovery and reputation evaluation method based on CRH, which improves data quality by iteratively optimizing the data source weights and task truth values;

[0065] The truth value calculation process of the present invention is a process of continuously iteratively calculating the estimated truth value of the data. At the beginning of the iteration, first calculate the average value of the data collected for the current task where Q represents the total number of workers submitting the data for this task; then calculate the standard deviation of the current task data based on this average value Through the standard deviation, we can calculate the standard error between the data submitted by a single worker and the estimated truth value x of the data * ​In the first round at the beginning of the iteration, we set to obtain the standard error of the task data submitted by each worker; subsequently, based on the standard error between the data submitted by the worker and the estimated true value x of the data * , we can calculate the data weight of the data submitted by the worker Then we substitute the data weight of the worker into the calculation formula of x * to obtain a new estimated true value. We repeat this iterative process until the value of the estimated true value x converges and the iteration ends. Subsequently, we observe the data weights of the workers at the time of convergence, select the top k workers among them, mark their performance in this round as credible, adopt their data and pay them. CS will send the data weights of these workers to TA. TA classifies the workers into malicious, unknown, and credible categories according to their reputation values. If the reputation value of a worker is greater than the high threshold * , then it belongs to the credible set; if the reputation value of the worker is between and , then it belongs to the unknown set; if the reputation value is less than the low threshold , then it belongs to the malicious set; TA selects the credible workers among them (assuming there are a total of K), calculates the average value of their data weights For unknown workers, calculate the task score If the task score ρ of the worker is greater than or equal to the task score threshold ρ i , that is, ρ △ ≥ρ i ≥ρ △ , then calculate the new increased reputation value through t new =min{(1 - μ)ρ + ηt, 1}; conversely, if ρ i <ρ △ , then calculate the new decreased reputation value through t new =max{-(1 - μ)(1 - ρ i ) + ηt, 0}. η in the two formulas is called the rate coefficient, which is used to control the influence degree of the historical reputation value on the new reputation value. After the reputation value of the worker is updated, TA classifies the worker into the corresponding category according to the updated reputation value.

[0066] Example 2

[0067] The difference between this example and Example 1 is only that:

[0068] In a crowd-sensing network, when it is necessary to obtain real data of a certain place and requires relatively high data quality, such as data on temperature and humidity, air quality, traffic conditions, etc. for sensing application scenarios, and the sensed data is uploaded to a platform for processing. In the tasks issued by the platform, workers apply to participate in the tasks under false names and provide different data. By applying the method of the present invention to the above data collection, the accuracy of the data can be improved without revealing the credibility of the workers.

[0069] The experimental results of the inventive method are given below.

[0070] Figure 1 The network structure diagram of the method of the present invention is given. The method of the present invention has the following components: a credibility certification center, which is trustworthy; a server, the server is curious and cautious and will strictly execute according to the data collection process, but may be curious to obtain the credibility of the workers; a data requester, which is an entity that needs data; and numerous workers who can execute sensing tasks. The specific data collection process can be seen in the description of the claims of the method of the present invention.

[0071] Figure 2 The experimental results of the method of the present invention and other methods in terms of data accuracy are given. There are two methods compared with the method of the present invention: one is the CRH method, that is, N workers are recruited to execute a sensing task at the same time, and then the weighted average of the N obtained data is taken as the final result; the other is the CRH method with privacy protection, and this method also does not select workers based on the credibility of the workers; finally, it is the method of the present invention. From the experimental results Figure 2 , and the corresponding experimental data tables 1 and 2, it can be seen that since the method of the present invention verifies the credibility of the workers and calculates the true value based on the credibility of the workers, the obtained results are more accurate. Compared with the other two methods, after a relatively long-term data collection, the average absolute error of the method of the present invention has decreased by 21.75%. The root mean square error has decreased by 80.17%.

[0072] Table 1: Comparison of different methods in terms of average absolute error

[0073]

[0074]

[0075] Table 2: Comparison of different methods in terms of root mean square error

[0076]

[0077] Figure 3The change in the credibility of workers in the method of the present invention is given. It can be seen from the experimental result graph that the credibility of trustworthy workers is continuously rising, while the credibility of dishonest workers is continuously falling. This shows that the method of the present invention can effectively identify the credibility of workers, so as to guide the selection of trustworthy workers for data collection and improve data quality.

[0078] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same and similar parts between the various embodiments, reference can be made to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and reference can be made to the description of the method part for the relevant parts.

[0079] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A trusted worker recruitment method based on identity privacy protection, characterized in that, It includes the following steps: The data requester submits a task request to the cloud server; The cloud server publishes the task to the workers; The workers who wish to participate in the task register their identities with the trust certification center. When registering, the workers generate a key pair encrypted by elliptic curve cryptography locally; After retaining the private key, the public key is sent to the certification center, and the certification center will generate a pseudonym for the worker according to the worker's public key; The certification center sends the pseudonym to the corresponding worker; The worker applies for a sensing task from the cloud server using its own pseudonym and reputation value. The worker's reputation value is obtained by querying the certification center based on its identity; The cloud server interacts with the certification center to verify whether the worker's reputation value has been tampered with. Only the workers who pass the verification will be recruited; The cloud server selects N workers from the workers who pass the verification to perform the sensing task and notifies the selected workers; The selected workers sense the data and submit the data to the cloud server; The cloud server performs true value calculation and reputation evaluation on the sensing data submitted by the workers; The cloud server sends the true value to the data requester and sends the reputation value evaluation result to the certification center; The certification center updates the reputation value of the worker.

2. The trusted worker recruitment method based on identity privacy protection according to claim 1, characterized in that, Pseudonymization refers to an identity privacy protection technology that replaces personal identity information with pseudo-identifiers to ensure that the data subject cannot be identified from the data without using other additional information; the worker generates a key pair encrypted by an elliptic curve locally, where the public key is Pk i , and the private key is Sk i ; when the worker registers, the public key Pk i is sent to the authentication center; the SHA1 algorithm is used as the hash function, denoted as Hash(·). Before the hash algorithm processing, first add a unique salt value salt i for each worker ω i . The salt value is a randomly generated string salt of any length l i ←Rand(l). After concatenating the salt value as an attribute to the worker's public key Pk i , then perform the hash processing to calculate the worker's pseudonym identifier p i = Hash(Pk i ||salt i ), where "||" represents the string concatenation operation; finally, the worker uses the pseudonym p i to apply to the cloud server to participate in the sensing task; since each worker's pseudonym is generated by its unique salt value, and the cloud server cannot access the original Pk i and salt i , it prevents the cloud server or data requester from linking to the specific user identity through the sensing data.

3. The trusted worker recruitment method based on identity privacy protection according to claim 1, wherein A two-party reputation verification process based on zero-knowledge proof is adopted, enabling the cloud server to verify with the certification center whether the reputation submitted by the worker is true without obtaining the true reputation value of the worker; Zero-knowledge proof is used to verify whether the reputation value of the worker is equal to the reputation value stored in the certification center, that is, whether the reputation value of the worker has been tampered with; The prover proves to the verifier that the reputation value t' owned by the prover is the same as the reputation value t owned by the verifier; at the same time, the prover will not disclose t' to the verifier during this process. According to the reputation value, the verifier measures whether the reputation value of each prover passes the verification through the formula υ = ν(t, t'); 4. A trusted worker recruitment method based on identity privacy protection according to claim 1, characterized in that The trusted center acts as the verifier, and the worker is the prover. The worker proves to the server that the reputation value it owns is valid. The commitment is used to prevent the cloud server from knowing the reputation value, thereby protecting the privacy of the worker's reputation value. The specific process is as follows: G is a cyclic group of prime order p, and g1 is an element. g2, h1, and h2 are elements generated by g1. The discrete logarithm based on h1, the discrete logarithm based on g1, the discrete logarithm based on h2, and the discrete logarithm based on g2 are all unknown to the workers. Assume that t is the worker reputation value stored by the trusted center, and t′ is the reputation value submitted by the worker. l, s1, and s2 are three security parameters. Workers willing to participate in the sensing task first select three random integers ω, λ1, λ2, which satisfy ω ∈ {1, 2 l ·p + 1}, Then calculate Through And Get Then, the worker calculates D = ω + H · t′, D1 = λ1 + H · r1 and Hash(·) is a hash function; finally, the worker sends H, D, D1, D2, C m ′ to the cloud server; After receiving data from the worker side, the cloud server obtains a commitment from the trusted center where t is the worker reputation value stored in the trusted center and The cloud server verifies whether the reputation value has been tampered with by checking whether H is equal to ; if they are equal, the reputation value of the worker has not been tampered with and the worker passes the verification; otherwise, the verification fails and the worker will not be recruited by the server to participate in the subsequent steps; for the convenience of the following description, if u k passes the trust verification algorithm, then υ k = ν(t, t′) = 1; otherwise, υ k = ν(t, t′) = 0; only the workers who pass the trust verification will be recruited to participate in the subsequent true value calculation.

5. A trusted worker recruitment method based on identity privacy protection according to claim 1, characterized in that A true value discovery and reputation evaluation method based on CRH is adopted to improve the data quality by iteratively optimizing the data source weight and task true value; The true value calculation process is a process of continuously iteratively calculating the estimated true value of the data; At the beginning of the iteration, first calculate the average value of the data collected for the current task where Q represents the total number of workers who submitted the data for this task; then calculate the standard deviation of the current task data based on this average value Through the standard deviation, calculate the standard error between the data submitted by a single worker and the estimated true value x of the data * In the first round at the beginning of the iteration, set to obtain the standard error of the task data submitted by each worker; subsequently, based on the standard error between the data submitted by the worker and the estimated true value x of the data * calculate the data weight of the data submitted by the worker Substitute the data weight of the worker into the calculation formula of x * to obtain a new round of estimated true value;​​ Repeat this iterative process until the estimated true value x * converges, then end the iteration. Subsequently, observe the data weights of the workers when convergence occurs, select the top k workers with higher weights, mark their performance in this round as credible, adopt their data, and pay them. The cloud server will send the data weights of the workers to the certification center. The certification center classifies the workers into malicious, unknown, and credible categories according to their reputation values. If a worker's reputation value is greater than the high threshold , then it belongs to the credible set. If a worker's reputation value is between and , then it belongs to the unknown set. If a worker's reputation value is less than the low threshold , then it belongs to the malicious set. The certification center selects the credible workers among them. Assume there are a total of K credible workers, and calculate the average value of their data weights For unknown workers, calculate the task score If a worker's task score ρ i is greater than or equal to the task score threshold ρ △ , that is, ρ i ≥ρ △ , then calculate the new increased reputation value through t new =min{(1 - μ)ρ + ηt, 1}. Conversely, if ρ i <ρ △ , then calculate the new decreased reputation value through t new =max{-(1 - μ)(1 - ρ i ) + ηt, 0}, where η is called the rate coefficient, which is used to control the influence degree of the historical reputation value on the new reputation value. After the reputation value of the worker is updated, the certification center classifies it into the corresponding category according to the updated reputation value.

Images