Reputation-based privacy protection truth value discovery method in block chain

Abstract

The invention discloses a privacy protection truth value discovery method based on reputation in a block chain. The method provided by the invention aims at the problems of perception data quality reduction, participant privacy information leakage and the like caused by lack of a trusted participant screening mechanism and insufficient privacy protection in the crowd intelligence network. Therefore, participants with high creditworthiness are screened on the premise of privacy protection to complete a data perception task, so that high-quality data are obtained. According to the method, firstly, a double-chain collaborative architecture is constructed based on a block chain, and privacy protection is performed on creditworthiness of participants through exclusive or homomorphic encryption; in the subsequent task allocation stage, the high-reputation participants are preferentially selected to execute the data acquisition task in combination with multi-round encryption communication and a dynamic screening mechanism; and finally, on the basis of the data of the high-credibility participants, calculating the task scores of the participants with unknown reputation, dynamically updating the reputation of the participants, and ensuring the authenticity and accuracy of the data.

Description

Technical Field

[0001] The present invention belongs to the field of high-quality data collection in crowd intelligence networks, and particularly relates to a method for collecting accurate and authentic data by effectively evaluating the credibility of participants in crowd intelligence networks. Background Art

[0002] Crowd-sensing networks are an innovative model for information collection and processing. Leveraging the ubiquity of mobile devices, they treat ordinary users as sensing nodes. Using mobile devices like phones, users can collect data on various aspects of the environment, transportation, and daily life. This data is then transmitted over the network to a central platform, where it is analyzed and processed to transform valuable knowledge and decision-making insights. Crowd-sensing networks offer advantages such as low cost, wide coverage, and strong real-time performance. However, while these advantages are crucial, data quality and privacy protection remain crucial issues that cannot be ignored.

[0003] In crowdsensing networks, various services require extensive data support. Data quality plays a crucial role in these networks, directly determining the quality of services. Poor data quality can significantly compromise service quality and potentially lead to serious losses, including loss of property, equipment, and even human life. For example, in environmental monitoring services, if the collected data on air and water quality exhibits significant discrepancies, the system may issue erroneous environmental warnings, misleading government officials into making inappropriate decisions. This not only increases challenges in environmental governance but also can lead to misallocation of resources and delays in timely response to pollution. More seriously, if erroneous environmental warnings prevent governments from promptly identifying and addressing pollution incidents, they can have long-term impacts on ecosystems and public health. In this context, poor data quality not only undermines the scientific nature of environmental governance but can also directly endanger ecological security and public health, highlighting the importance of ensuring data quality for crowdsensing networks and the services they provide.

[0004] Secondly, privacy protection is a key issue in crowdsensing networks. To encourage workers to actively participate in sensing and data collection, crowdsensing networks must not only provide workers with financial rewards to compensate for their resources and time, but also properly address the critical issue of privacy protection. Much of the data perceived by workers often contains sensitive personal information. For example, in health monitoring scenarios, workers' vital signs data carries strong privacy implications. Leakage can lead to serious consequences such as insurance denial and employment discrimination. When workers perceive that their privacy is not effectively protected, their willingness to participate in data sensing and providing data decreases significantly. This not only reduces the amount of perceived data but also compromises data quality, ultimately impacting the effectiveness of the entire crowdsensing network. If a large number of workers refuse to sense data due to privacy concerns, those who remain willing to participate may only be those with low privacy sensitivity. This will significantly reduce the diversity and comprehensiveness of the data, severely hindering the sustainable development of the crowdsensing network. Therefore, establishing a comprehensive privacy protection mechanism is key to incentivizing workers to continue participating in data sensing and ensuring the healthy development of the crowdsensing network.

[0005] Therefore, in a crowd-sensing network, data quality is related to service quality and security, and privacy protection affects worker participation and data quality and diversity, both of which are crucial to network performance and sustainable development. The method of the present invention proposes an effective solution to the above problems. Summary of the Invention

[0006] This invention discloses a reputation-based privacy-preserving truth-value discovery method in a blockchain. This method addresses the difficulties in identifying the quality of participant-submitted data in crowd-intelligence networks due to the lack of standards for comparing data submitted by participants. This can lead to increasingly poor data quality due to the increasing number of dishonest participants, and the lack of an assessment of participants' performance in submitting data over past time periods. The method proposes a method for obtaining high-quality data by effectively evaluating and obtaining participant reputation. Typically, in mobile crowd-intelligence systems, platforms lack standards for comparing participant-submitted data, leading to dishonest participants intentionally reporting low-quality data. Generally speaking, higher data quality requires more time and resources, resulting in higher perceived costs. Dishonest participants tend to submit low-quality data to reduce costs and maximize revenue, but this can harm the crowd-intelligence network. Therefore, the method of the present invention proposes a low-cost, effective method for achieving this goal by effectively evaluating and obtaining participant reputation and then selecting trustworthy participants to perceive data. The method proposed in the present invention is based on the premise that the initial platform starts with a small number of highly credible participants. Thereafter, when selecting participants, multiple participants are selected for a unified data collection task under privacy protection. Participants with higher trust will be selected from the participants who apply to participate in the task for data collection. On the other hand, the trust of the participants is evaluated by comparing the data of participants with unknown trust with the data of highly credible participants, thereby effectively identifying whether the data collected by the participants is true, and thereby evaluating the credibility of the participants. It should be noted that the method proposed in the present invention adopts a dual-chain collaborative architecture to achieve dynamic reputation management. When selecting participants to complete the task, the screening of credible participants is carried out by comparing the reputation scores of the participants (after XOR homomorphic encryption). At the same time, the corresponding relationship between the participants and the reputation scores is not known to others during the screening process, thus ensuring the screening of credible workers under privacy protection.

[0007] The technical solutions of the invention are as follows:

[0008] 1. A reputation-based privacy-preserving truth value discovery method in a blockchain, characterized by comprising the following steps:

[0009] Prerequisites:

[0010] (1) Elliptic Curve Cryptography. Elliptic Curve Cryptography is a public key cryptography system. Given an elliptic curve EC of prime order l, with G as its base point, SP p,i is a randomly selected number and must satisfy 0 <SP p,i <l,SP p,i As the private key, the public key is PK p,i =SP p,i G.

[0011] (2) Schnorr signature. Schnorr signature allows the signer to prove that he owns the private key SP p,i , because the signature allows the recipient to use the corresponding public key PK p,i To verify.

[0012] Initialization: DR publishes a data collection task in the data processing chain, which contains information such as the number of participants required is n, the number of participants required to estimate the true value of the data collection task is k, where k < n. The public key of the participant is PK p,i , the private key is SP p,i PK p,i Publicly known. Given a base point G of the elliptic curve EC of prime order l. The public key of DR is J and the private key is j. It is assumed that there are a number of high-trust participants at the beginning. These high-trust participants are obtained through external methods. The data collected by these participants are completely authentic and trustworthy, and their trustworthiness is always considered to be 1. It is assumed that except for the high-trust participants whose credibility is 1 and known, the credibility status of other participants is unknown, and their credibility is set to 0.5. The credibility value range is [0,1]. The condition that needs to be met for a trusted participant is greater than the trust threshold For example

[0013] The components of the system and their corresponding functions are as follows:

[0014] Data processing chain: a blockchain network used to store tasks issued by DR and data collected by participants;

[0015] Reputation Processing Chain: A blockchain network used to manage the identities of participants, and store, update, and maintain their reputations and task scores.

[0016] Data requesters: can publish data collection tasks on the data processing chain, which will score the data of participants and synchronize the encrypted task scores on the reputation processing chain. Hereinafter referred to as DR;

[0017] Participants: can view the task list on the data processing chain. After collecting data, it will be encrypted and uploaded to the data processing chain. When a participant completes the specified number of data collection tasks, the participant needs to update his or her reputation on the reputation processing chain;

[0018] Phase 1: Each participant selects a unique secret identifier

[0019] (1) After the participants in the network learn the task information, they voluntarily choose whether to send an application to the DR. When sending an application, the participants randomly select a secret identifier. Its range is t i ∈[1,n]. Convert to binary number B i , after encryption, it forms P i Send to DR

[0020]

[0021] where τ i,n (f) means only participant U i The one-time password of the f-th message of length n that you know. The initial value of f is 1. It is an XOR calculation;

[0022] (2) DR receives P from n participants i Then calculate P i The XOR of P0 is sent to these n participants

[0023]

[0024] Afterwards, DR compares count(P0) with n, where P0 is a binary number, count(P0) is the number of 1s in the binary number P0, and count(·) is a function that counts the number of 1s in the binary number, and the same applies hereafter. If count(P0) = n, it can be seen from the properties of the XOR operation that each participant has generated a unique secret identifier, and step (4) is executed. Otherwise, step (3) is executed. Since τ i,n (f) Known only to the corresponding participants, DR does not know the actual Correspondence with participants. Since participants do not know who sent the application to DR, when the number of applicants exceeds the predetermined number of participants n, DR will randomly select n participants' applications from the applicants, and the remaining participants' applications will be discarded;

[0025] (3) When participant U i Receive P0, check P0 Position If it is 0, it means that the secret identifiers screened by the participants have collided and they need to reselect t i Until it appears again like Keep No change. After that, f increases by 1, and participant U i Update P i

[0026]

[0027] Pi Send to DR. Then execute step (2). Participants whose applications are discarded will not receive P0. After waiting for a while, they will think that their applications have been discarded by DR.

[0028] (4) Participant U i calculate

[0029]

[0030] in P0 from 0th to A substring of bits. As a participant i The identifier is used to facilitate the subsequent screening of participants for the true value estimation. At this point, the work of generating unique identifiers for participants is completed;

[0031] Phase 2: Platform privacy protection selects k participants with the highest credibility from n participants;

[0032] (5) DR accesses the encrypted reputation X of the corresponding participant on the reputation processing chain i . DR then initializes the variable Y cnt =0,res=0,v=1,E i =1,q=1;

[0033] Where q is used to indicate that the participant with the qth highest credibility is selected. Every time step (6) is executed l times, q will be incremented by 1 and the variable Y will be reset. cnt ,res,v,E i is the initial state, until q = k, the participants with the highest credibility are selected. r,i Is an encrypted binary number with a length of l bits. cnt It is used to record the number of participants whose reputation is greater than the qth largest reputation. res is used to store the encrypted qth largest reputation, v is used to indicate that the vth bit of the reputation is compared in step (6), E i Each participant U i state variables;

[0034] (6) When q≤k, repeat the following operation until the qth maximum credibility of the encryption is determined:

[0035] a. Compare the reputation of each participant one by one: Participant U i Send an encrypted binary message M′ of length n to DR i

[0036]

[0037] Among them, Mi is a binary number message of length n, M' i is the encrypted binary number message. s r,i [v] is the value of the v-th bit of s r,i For the message M i :

[0038]

[0039] b. DR Aggregation and Statistics: After DR receives all M' i it will calculate

[0040]

[0041] T = count(M0) (8)

[0042] Once T + Y cnt < n - q + 1, then

[0043] Y cnt = T + Y cnt (9)

[0044] res = res||1 (10)

[0045] Otherwise

[0046] res = res||0 (11)

[0047] c. Participant Status Update: Broadcast the last bit of res to the participants. After participant U i receives the last bit of res, check whether it is equal to s r,i [v], if not equal, then set E i = 0;

[0048] d. Round Advancement: If v = l, it means that the reputation of the q-th largest participant has been found, q is incremented by 1, and at the same time, reset the variables Y cnt , res, v, E i to the initial state, continue to execute step (6) until q > k, then execute step (7). Otherwise, increment v by 1, and then continue to execute step (6). || is the binary string concatenation operator;

[0049] (7) After finding the top k encrypted reputations among the participants participating in the data collection task, DR will publish k tokens on the reputation processing chain. Each token contains the encrypted reputation of the corresponding participant, where the one-time address K of the token O is generated using the public key PK of the participant p,i The generation steps are described as follows:

[0050] (a) Select a random number r∈[0,l-1] and generate the reputation token public key R

[0051] R=rG (12)

[0052] (b) Calculate K O

[0053] K O =H n (r,PK p,i )G (13)

[0054] Among them H n is a hash function that maps to integers from 0 to l-1;

[0055] (8) Participants determine whether they are selected: Participants need to verify whether these reputations belong to themselves on the reputation processing chain, using their own private key SP p,i Check whether it has the newly issued token. Participants calculate the one-time private key k O

[0056] k O =H n (SP p,i ,R) (14)

[0057] Using k O Calculate K′=k O G. If K′=K O , you can use the one-time private key k O Unlock the token;

[0058] Phase 3: Submit perception data and mark task scores;

[0059] (9) After completing the task, the participant needs to call the data submission contract to submit the data. There is a transaction public key in the task data transaction. DR public key J, transaction key The sensing data DATA collected by the participant, the participant's public key PK p,i , the block height h used to indicate the location of the participant's reputation token on the reputation processing chain r wait;

[0060] in is a randomly selected integer in the range is a shared secret The hash of . DR is calculated as follows

[0061]

[0062] When DR receives the sensing data transaction, it uses its private key j and the transaction public key Φ to reconstruct the symmetric key H n (jΦ) to decrypt task-related data;

[0063] (10) DR verifies the identity of the participant who completed the task: DR verifies the identity of the participant through h r After finding the corresponding participant’s reputation token, verify its Schnorr signature θ v , the specific steps are as follows:

[0064] DR calculation challenge c′ and Γ′ calculated based on c′

[0065] c′=H n (ξ′,αG) (16)

[0066] Γ′=αG+c′PK p,i (17)

[0067] If Γ = Γ′, then the signature verification is successful, and the response γ is defined as

[0068] Γ=γG (18)

[0069] γ=α+cSP p,i (max l) (19)

[0070] Where c is the challenge calculated by the participant, and the participant selects a secret random number α∈[0,l-1]

[0071] c=H n (ξ,αG)

[0072] If Schnorr signature θ v If it is verified to be valid, then DR can believe that the perception data is indeed submitted by the participant;

[0073] (11) DR updates the task score. The steps for updating the task score are as follows:

[0074] Check the comparison results of the data collected by the collector and the data collected by the high-credibility participants. If they are within the error range, then the number of collections C of the i-th participant at time t is i,t Increase by 1. If there is no high-credibility participant in the data collection task, and there is a credible participant, then check the comparison results of the data collected by the collector and the data collected by the credible participant. If they are within the error range, then C i,t Increase by 1. Calculate the comparative accuracy σ of the participant at time t r

[0075]

[0076] Among them Ci,tot is the total number of comparisons for the participant;

[0077] Update the task score a of the i-th participant at time t i,t

[0078]

[0079] where a i,t-1 represents the task score of the i-th participant at time t-1. The task score update step λ satisfies λ<1 and is used to control the speed of each task score update. The threshold σ s Used to determine whether the task score update should increase or decrease;

[0080] (12) DR marks the final task score and calls the task score contract to create an encrypted task token on the reputation processing chain. Task score a i Encrypted to get s t :

[0081]

[0082] Will s t Published in a transaction. DR submits a task token transaction to call the task score contract on the reputation processing chain. The task token transaction is included in a block, which is eventually written into the reputation processing chain to generate a new task token with a public block height of h. t Used to indicate its location;

[0083] (13) When a new task token is generated, the participant will use the token’s one-time address K O , token public key R and your own private key SP p,i To check whether the task token is assigned to yourself, the checking steps are the same as step (8). Calculation Decryption task score;

[0084] The fourth stage: updating the credibility;

[0085] (14) If a predefined number z of task tokens have been received, the participant will collect his own tokens and generate a new reputation token. The participant uses the z task scores {a1,...,a z Credibility can be calculated

[0086]

[0087] Calculated credibility X i Used to generate the encrypted score sr,i

[0088]

[0089] Among them rPK p,i is a shared key.

[0090] Beneficial effects

[0091] This invention discloses a reputation-based privacy-preserving truth value discovery method in a blockchain. The basic concept of the invented method is that the crowd-sensing network platform initially has a small number of highly trusted participants. Subsequently, when selecting participants for a unified data collection task, multiple participants are selected while protecting privacy. Among the participants who apply for the task, participants with higher trustworthiness are prioritized for data collection. This ensures that the platform collects authentic data and achieves good results. This reduces or even prevents damage to the system caused by malicious participants. Furthermore, the trustworthiness of participants is evaluated by comparing the data of participants with unknown trustworthiness with that of participants with higher trustworthiness. This effectively identifies the authenticity of the data collected by the participants and uses this to evaluate the participants' reputation. As this process continues, the platform will be able to identify more and more participants, and the identification speed will increase. Once the platform has identified a sufficient number of trustworthy participants, it can select participants from these participants to complete the data collection task to meet the requirements. This reduces the platform's participant identification process and thus reduces costs. Furthermore, because a sufficient number of trustworthy participants have been identified to provide data, the authenticity and trustworthiness of the data are guaranteed. This achieves the design goal. BRIEF DESCRIPTION OF THE DRAWINGS

[0092] Figure 1 This is a system network configuration diagram of the method of the present invention.

[0093] Figure 2 The conditions of the participants required to complete the data collection task;

[0094] Figure 3 Methods are to obtain the accuracy of data; DETAILED DESCRIPTION

[0095] To facilitate understanding of the present invention, the present invention will be described in more comprehensive and detailed form below in conjunction with the accompanying drawings and preferred embodiments. However, the scope of protection of the present invention is not limited to the following specific embodiments.

[0096] Unless otherwise defined, all technical terms used hereinafter have the same meanings as those generally understood by those skilled in the art. The technical 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.

[0097] 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.

[0098] Example:

[0099] In the forest fire early warning system, it is necessary to monitor the temperature, humidity, smoke concentration and wind direction data of the forest area in real time, and upload the perception data to the emergency management platform for analysis, so as to timely predict the fire risk level and initiate emergency response. In this scenario, various sensors deployed in the forest area (such as IoT temperature and humidity nodes, drone inspection equipment) and volunteer handheld devices participate in data collection as perception nodes. However, some sensors may report abnormal data due to equipment failure or malicious tampering, and volunteers may also misreport or fabricate fire information, causing the platform to misjudge the risk and delay disaster relief. Therefore, there is an urgent need for a truth value discovery method that can dynamically evaluate the credibility of the data source and protect privacy to ensure the accuracy and reliability of the early warning system.

[0100] The experimental results of the inventive method are given below.

[0101] 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 data processing chain for storing the tasks issued by the DR and the data collected by the participants; a reputation processing chain for managing the identities of the participants, storing, updating and maintaining the reputation and task scores of the participants; the data requester is the entity that needs the data and will score the task scores of the participants; and a large number of participants can perform perception tasks and update their own reputations. The specific data collection, task score evaluation and reputation update process can be found in the description of the claims of the method of the present invention.

[0102] Figure 2 The number of participants required to complete the data collection task by the method of the present invention and the conventional method is given. From the experimental results, it can be seen that the method of the present invention requires far fewer participants than the conventional method.

[0103] Figure 3 The error between the data obtained using the method of the present invention and conventional methods for completing data acquisition tasks and the true value is presented. Experiments show that the accuracy of the data obtained by the method of the present invention increases as the method is implemented. Furthermore, after stabilization, the accuracy of the data obtained by the method of the present invention is far higher than that of conventional methods.

Claims

1. A reputation-based privacy-preserving truth value discovery method in blockchain, characterized by The following steps are involved: Prerequisites: (1) Elliptic Curve Cryptography. Elliptic Curve Cryptography is a public key cryptography system. Given an elliptic curve EC of prime order l, with G as its base point, SP p,i is a randomly selected number and must satisfy 0 <SP p,i <l,SP p,i As the private key, the public key is PK p,i =SP p,i G. (2) Schnorr signature. Schnorr signature allows the signer to prove that he owns the private key SP p,i , because the signature allows the recipient to use the corresponding public key PK p,i To verify. Initialization: DR publishes a data collection task in the data processing chain, which contains information such as the number of participants required is n, the number of participants required to estimate the true value of the data collection task is k, where k < n. The public key of the participant is PK p,i , the private key is SP p,i PK p,i Publicly available. Given an elliptic curve EC of prime order l, a base point is G. The public key of DR is J and the private key is j. It is assumed that there are a number of high-trust participants at the beginning. These high-trust participants are obtained through external methods. The data collected by these participants is completely authentic and trustworthy, and their trustworthiness is always considered to be 1. It is assumed that except for the high-trust participants whose credibility is 1 and known, the credibility status of other participants is unknown and their credibility is set to 0.

5. The credibility value range is [0,1]. The condition that needs to be met for a trusted participant is that the credibility threshold is greater than For example The components of the system and their corresponding functions are as follows: Data processing chain: a blockchain network used to store tasks issued by DR and data collected by participants; Reputation Processing Chain: A blockchain network used to manage the identities of participants, and store, update, and maintain their reputations and task scores. Data requesters: can publish data collection tasks on the data processing chain, which will score the data of participants and synchronize the encrypted task scores on the reputation processing chain. Hereinafter referred to as DR; Participants: can view the task list on the data processing chain. After collecting data, it will be encrypted and uploaded to the data processing chain. When a participant completes the specified number of data collection tasks, the participant needs to update his or her reputation on the reputation processing chain; Phase 1: Each participant selects a unique secret identifier (1) After the participants in the network learn the task information, they voluntarily choose whether to send an application to the DR. When sending an application, the participants randomly select a secret identifier. Its range is t i ∈[1,n]. Convert to binary number B i , after encryption, it forms P i Send to DR where τ i,n (f) means only participant U i The one-time password of the f-th message of length n that you know. The initial value of f is 1. It is an XOR calculation; (2) DR receives P from n participants i Then calculate P i The XOR of P0 is sent to these n participants Afterwards, DR compares count(P0) with n, where P0 is a binary number, count(P0) is the number of 1s in the binary number P0, and count(·) is a function that counts the number of 1s in the binary number, and the same applies hereafter. If count(P0) = n, it can be seen from the properties of the XOR operation that each participant has generated a unique secret identifier, and step (4) is executed. Otherwise, step (3) is executed. Since τ i,n (f) Known only to the corresponding participants, DR does not know the actual Correspondence with participants. Since participants do not know who sent the application to DR, when the number of applicants exceeds the predetermined number of participants n, DR will randomly select n participants' applications from the applicants, and the remaining participants' applications will be discarded; (3) When participant U i Receive P0, check P0 Position If it is 0, it means that the secret identifiers screened by the participants have collided and they need to reselect t i Until it appears again like Keep No change. After that, f increases by 1, and participant U i Update P i P i Send to DR. Then execute step (2). Participants whose applications are discarded will not receive P0. After waiting for a while, they will think that their applications have been discarded by DR. (4) Participant U i calculate in P0 from 0th to A substring of bits. As a participant i The identifier is used to facilitate the subsequent screening of participants for the true value estimation. At this point, the work of generating unique identifiers for participants is completed; Phase 2: Platform privacy protection selects k participants with the highest credibility from n participants; (5) DR accesses the encrypted reputation X of the corresponding participant on the reputation processing chain i . DR then initializes the variable Y cnt =0,res=0,v=1,E i =1,q=1; Where q is used to indicate that the participant with the qth highest credibility is selected. Every time step (6) is executed l times, q will be incremented by 1 and the variable Y will be reset. cnt ,res,v,E i is the initial state, until q = k, the participants with the highest credibility are selected. r,i Is an encrypted binary number with a length of l bits. cnt It is used to record the number of participants whose reputation is greater than the qth largest reputation. res is used to store the encrypted qth largest reputation, v is used to indicate that the vth bit of the reputation is compared in step (6), E i Each participant U i state variables; (6) When q≤k, repeat the following operation until the qth maximum credibility of the encryption is determined: a. Compare the reputation of each participant one by one: Participant U i Send an encrypted binary message M′ of length n to DR i Among them, M i is a binary message of length n, M′ i is the encrypted binary message. r,i [v] is s r,i The value of the vth bit. For message M i : b.DR aggregation and statistics: DR receives all M′ i Will calculate later T=count(M0) (8) Once T + Y cnt <n - q + 1, then AND cnt =T+Y cnt (9) res=res||1 (10) otherwise res=res||0 (11) c. Participant status update: broadcast the last bit of res to the participants. Participant U i After receiving the last bit of res, check it with s r,i [v] are equal, if not equal, let E i =0; d. Round advancement: If v = 1, it indicates that the qth largest participant’s credibility has been found, q is incremented by 1, and the variable Y is reset. cnt ,res,v,E i The initialization state continues with step (6) until q>k, then proceed to step (7). Otherwise, increment v by 1, then proceed to step (6). || is the binary string concatenation operator; (7) After finding the top k encrypted reputations of the participants in the data collection task, DR will issue k tokens on the reputation processing chain. Each token contains the encrypted reputation of the corresponding participant, where the one-time address of the token is K O Use the participant's public key PK p,i Generate. The generation steps are described as follows: (a) Select a random number r∈[0,l-1] and generate the reputation token public key R R=rG (12) (b) Calculate K O K O =H n (r,PK p,i )G (13) Among them H n is a hash function that maps to integers from 0 to l-1; (8) Participants determine whether they are selected: Participants need to verify whether these reputations belong to themselves on the reputation processing chain, using their own private key SP p,i Check whether it has the newly issued token. Participants calculate the one-time private key k O k O =H n (SP p,i ,R) (14) Using k O Calculate K′=k O G. If K′=K O , you can use the one-time private key k O Unlock the token; Phase 3: Submit perception data and mark task scores; (9) After completing the task, the participant needs to call the data submission contract to submit the data. There is a transaction public key in the task data transaction. DR public key J, transaction key The sensing data DATA collected by the participant, the participant's public key PK p,i , the block height h used to indicate the location of the participant's reputation token on the reputation processing chain r wait; in is a randomly selected integer in the range is a shared secret The hash of . DR is calculated as follows When DR receives the sensing data transaction, it uses its private key j and the transaction public key Φ to reconstruct the symmetric key H n (jφ) to decrypt task-related data; (10) DR verifies the identity of the participant who completed the task: DR verifies the identity of the participant through h r After finding the corresponding participant’s reputation token, verify its Schnorr signature θ v , the specific steps are as follows: DR calculation challenge c′ and Γ′ calculated based on c′ c′=H n (ξ′,αG) (16) Γ′=αG+c′PK p,i (17) If Γ = Γ′, then the signature verification is successful, and the response γ is defined as Γ=γG (18) γ=α+cSP p,i (mod l) (19) Where c is the challenge calculated by the participant, and the participant selects a secret random number α∈[0,l-1] c=H n (ξ,αG) If Schnorr signature θ v If it is verified to be valid, then DR can believe that the perception data is indeed submitted by the participant; (11) DR updates the task score. The steps for updating the task score are as follows: Check the comparison results of the data collected by the collector and the data collected by the high-credibility participants. If they are within the error range, then the number of collections C of the i-th participant at time t is i,t Increase by 1. If there is no high-credibility participant in the data collection task, and there is a credible participant, then check the comparison results of the data collected by the collector and the data collected by the credible participant. If they are within the error range, then C i,t Increase by 1. Calculate the comparative accuracy σ of the participant at time t r Among them C i,tot is the total number of comparisons for the participant; Update the task score a of the i-th participant at time t i,t where a i,t-1 represents the task score of the i-th participant at time t-1. The task score update step λ satisfies λ<1 and is used to control the speed of each task score update. The threshold σ s Used to determine whether the task score update should increase or decrease; (12) DR marks the final task score and calls the task score contract to create an encrypted task token on the reputation processing chain. Task score a i Encrypted to get s t : Will s t Published in a transaction. DR submits a task token transaction to call the task score contract on the reputation processing chain. The task token transaction is included in a block, which is eventually written into the reputation processing chain to generate a new task token with a public block height of h. t Used to indicate its location; (13) When a new task token is generated, the participant will use the token’s one-time address K O , token public key R and your own private key SP p,i To check whether the task token is assigned to yourself, the checking steps are the same as step (8). Calculation Decryption task score; The fourth stage: updating the credibility; (14) If a predefined number z of task tokens have been received, the participant will collect his own tokens and generate a new reputation token. The participant uses the z task scores {a1,...,a z Credibility can be calculated Calculated credibility X i Used to generate the encrypted score s r,i Among them rPK p,i is a shared key.

Images