A secure truth value discovery method for crowdsensing based on homomorphic encryption

Abstract

The present invention discloses a method for discovering a secure truth value of crowd intelligence perception based on homomorphic encryption, comprising the following steps: CS A Generate a public-private key pair (pk, sk) and send the public key pk to CS B ; Each participant sends perturbation data to CS A , CS A Invert and encrypt the perturbation data, and then send the ciphertext to CS B ; Data upload: Each participant sends the perturbed perception data to CS B , CS B Use Paillier plus homomorphic encryption algorithm to obtain the ciphertext of the perception data; weight update: in CS A With the assistance of CS B Use Paillier plus homomorphic encryption algorithm to obtain the participant weight ciphertext; truth value update: in CS B With the assistance of CS A The Paillier-based homomorphic encryption algorithm is used to obtain the true value of the observed object. Weight and true value updates are iteratively performed until the true value meets convergence criteria. Participants do not participate in the encryption and decryption of the data, and the truth value discovery operation is performed between two non-colluding cloud servers, reducing the computational effort and computational overhead of the participants.

Description

Technical Field

[0001] The present invention belongs to the field of crowd sensing technology, and more specifically, the present invention relates to a crowd sensing secure truth value discovery method based on homomorphic encryption. Background Art

[0002] In recent years, the widespread adoption of portable intelligent mobile devices (such as smartphones and smartwatches) equipped with a variety of sensors (GPS, accelerometers, voice sensors, etc.) and the widespread expansion of intelligent cloud computing have greatly promoted the development of mobile crowdsensing. Unlike traditional sensor network-based sensing, mobile crowdsensing relies on a large number of users, using their perception data as a data source. It leverages the widespread distribution of the public and the flexibility of mobility to collect data. Mobile crowdsensing has applications in many important areas, such as transportation planning, urban management, environmental monitoring, and social governance.

[0003] Mobile crowdsensing can collect perception information from numerous users. However, due to factors such as irregular user operation, the impact of the data collection environment, and varying hardware and software specifications, the perceived data of some users may differ to varying degrees from the actual data. To mitigate the impact of these factors on the collection of real information and improve data quality, crowdsensing-based truth discovery has been proposed and widely applied in various fields. Truth discovery is an effective method for efficiently obtaining real data from multiple data sources. Current truth discovery methods include TruthFinder and CRH. However, these algorithms fail to consider the potential for privacy leakage of participants during execution. They operate without encryption or perturbation, making it highly likely that sensitive information of participants could be intercepted during algorithm execution. Privacy issues in crowdsensing-based truth discovery still exist, such as insufficient privacy protection. Currently, researchers in this field have mostly used homomorphic encryption to protect data privacy. For example, Miao et al. proposed a privacy-preserving scheme for truth discovery called PPTD, which uses a threshold Paillier cryptographic system to protect the privacy of perception data collected by participants. However, each participant in this scheme needs to participate in encryption calculations in real time, which leads to the problem that encryption calculations are large and time-consuming, and there is also huge computing overhead on the part of the participants. Summary of the Invention

[0004] The present invention provides a truth value discovery method for crowd-sensing based on homomorphic encryption, aiming to improve the above problems.

[0005] The present invention is implemented as follows: a method for discovering secure truth values ​​of crowd-sensing based on homomorphic encryption, the method specifically comprising the following steps:

[0006] S1. Data requester sends data to cloud server CS A Issue data request, cloud server CS A Generate a public-private key pair (pk, sk) and send the public key pk to the cloud server CS B ; Each participant sends the randomly generated perturbation data to the cloud server CS through a secure channel A , Cloud Server CS A Invert and encrypt the perturbation data, and then send the perturbation ciphertext to the cloud server CS B ;

[0007] S2. Each participant’s perception of the observed data and Perform disturbance and make the perturbation perception data Send to cloud server CS B , Cloud Server CS B Perception data after disturbance After encryption, the disturbance is eliminated based on the disturbance ciphertext to obtain the perception data ciphertext

[0008] S3, in cloud server CS A With the assistance of Cloud Server CS B Based on the observation object O calculated in the previous round m The true value is used to calculate the distance ciphertext of each observation object in the current round t, and the weight ciphertext of each participant in the current round iteration t is calculated based on the distance ciphertext of all observation objects of each participant;

[0009] S4, in the cloud server CS B With the assistance of Cloud Server CS A According to the participants i The weighted ciphertext and observer u in the current iteration t i Ciphertext of the perception data of the measured object Get the true value of the observed object in the current iteration t;

[0010] S5, Cloud Server CS A Determine the truth value Whether the convergence conditions are met, if not, the cloud server CS A The truth value Send to cloud server CS B , proceed as the next round of iteration, return to step S3, until the calculated true value meets the convergence condition, and return the final true value to the data requester.

[0011] Furthermore, step S1 specifically includes the following steps:

[0012] S11, Cloud Server CS A Generate a private key pair (pk, sk) using the Paillier encryption algorithm;

[0013] S12, Cloud Server CS A Send the public key pk to the cloud server CS B ;

[0014] S13. Each participant u i The randomly generated perturbation data Sent to the cloud server CS through a secure channel A ,in, Represents participant u i For the mth observation object O m Randomly generated α value, Represents participant u i For the mth observation object O m Randomly generated beta values;

[0015] S14, Cloud Server CS A For perturbation data Invert and encrypt to get the disturbed ciphertext and Cloud Server CS A Send the perturbed ciphertext to the cloud server CS B .

[0016] Furthermore, step S2 includes the following steps:

[0017] S21, Participant u i Using perturbed data and For the observed object o m Perception data and Add disturbance and send the perturbation data to and Send to cloud server CS B ,in,

[0018] S22, Cloud Server CS B Perception data after disturbance and Encryption to obtain the disturbed perception data ciphertext and And through the Paillier addition homomorphism property and Eliminate disturbances and obtain the ciphertext of the perception data and

[0019] Furthermore, step S3 includes the following steps:

[0020] S31, Cloud Server CS B For the observation object O in round t-1 m The true value of Square to encrypt and form the true value ciphertext Then calculate the distance ciphertext of all observation objects of each participant Cloud Server CS B Aggregate the distance ciphertexts of all observed objects of each participant to obtain the distance ciphertext V of the participant i , and randomly generate a disturbance data r1 i , the distance ciphertext V of the participants i Add perturbation to obtain the distance ciphertext of the participants after perturbation

[0021] S32, Cloud Server CS B The distance ciphertext V of all participants i Perform aggregation to obtain the total distance ciphertext V all , then generate a random perturbation data r2 to the total distance ciphertext V all Add perturbation to get the total distance ciphertext after perturbation Cloud Server CS B The disturbed ciphertext and Send to cloud server CS A ;

[0022] S33, Cloud Server CS A First, the perturbed participant distance ciphertext and total distance ciphertext Decrypt and get the distance between participants after disturbance and Cloud Server CS A Calculate the CRH weight for each participant u in the current round iteration i Weight after perturbation Cloud Server CS A right Encrypt the ciphertext Send to cloud server CS B ;

[0023] S34, Cloud Server CS B According to the known disturbance data r1 i and r2, the perturbed weighted ciphertext is encrypted by the Paillier additive homomorphic encryption algorithm Eliminate the disturbance and get the participant u in the current round of iterationi The weighted ciphertext E pk (w i ).

[0024] Furthermore, the distance ciphertext of the observed object in the current round t The calculation formula is as follows:

[0025]

[0026] Furthermore, the weight The calculation formula is as follows:

[0027]

[0028] Furthermore, the weighted ciphertext E pk (w i ) is calculated as follows:

[0029]

[0030] Furthermore, step S4 specifically includes the following steps:

[0031] S41, Cloud Server CS B Randomly generate perturbation data And encrypt the participants' perception data Adding perturbation yields Then according to the disturbance data r1 i The weighted ciphertext E of each participant pk (w i ) Add the perturbation to get E pk (r1 i ×w i ), Finally, the disturbed perception data ciphertext and weighted ciphertext E pk (r1 i ×w i ) sent to the cloud server CS A ;

[0032] S42, Cloud Server CS A Perturbed ciphertext and E pk (r1 i ×w i ) to decrypt and get and r1 i ×w i , Cloud Server CS A Perform multiplication on the decrypted data and encrypt it to obtain the disturbed weighted data ciphertext And send it to the cloud server CS B ;

[0033] S43, Cloud Server CS B According to the known disturbance data and r1 i , through the Paillier addition homomorphic encryption algorithm Cancel the disturbance and obtain the weighted perception data ciphertext Cloud Server CS B Aggregate the weighted ciphertexts and weighted data ciphertexts of all participants to obtain the aggregated weighted ciphertexts of all participants and all participants on the observed object O m Aggregate weighted data ciphertext Cloud Server CS B Aggregate weight ciphertext and weighted data ciphertext Send to cloud server CS A ;

[0034] S44, Cloud Server CS A Aggregate weight ciphertext and aggregated weighted data ciphertext Decrypt to get the aggregate weight W all and aggregate weighted data VW m ; Cloud Server CS A Calculate the true value of each observation object in the current round iteration t

[0035] Furthermore, the truth value The calculation formula is as follows:

[0036]

[0037] The lightweight privacy-preserving truth value discovery method based on mobile crowd intelligence perception in this invention has the following features:

[0038] Beneficial effects:

[0039] 1) Participants only add perturbations and do not participate in the encryption and decryption of data. The truth discovery operation is performed between two non-colluding cloud servers, which reduces the amount of computation for participants while lowering the computational overhead on the participant side, thereby stimulating participants to actively participate in various perception tasks; 2) Participants can go offline after submitting data. During the truth discovery process performed by the cloud server, participants do not have to stay online all the time and do not need to interact with the cloud server. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A schematic diagram of the structure of a crowd-sensing secure truth value discovery system based on homomorphic encryption provided by an embodiment of the present invention;

[0041] Figure 2 Flowchart of a method for discovering secure truth values ​​of crowd-sensing based on homomorphic encryption provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The specific implementation methods of the present invention will be further explained in detail below by describing the embodiments with reference to the accompanying drawings, so as to help those skilled in the art to have a more complete, accurate and in-depth understanding of the inventive concept and technical solution of the present invention.

[0043] Figure 1 This is a schematic diagram of the structure of a crowd-sensing secure truth value discovery system based on homomorphic encryption provided by an embodiment of the present invention. For ease of explanation, only the parts related to the embodiment of the present invention are shown. The system includes:

[0044] Data requester, non-colluded cloud server CS A and cloud server CS B And several participants, the data requester sends the data to the cloud server CS A Send a data request, which carries the observation object corresponding to the requested data. The participant uploads the perception data generated by the corresponding observation object to the cloud server CS A , in the cloud server CS B With the help of , the true value is updated.

[0045] Figure 2 A flowchart of a method for discovering secure truth values ​​of crowd-sensing based on homomorphic encryption provided by an embodiment of the present invention, the method specifically includes the following steps:

[0046] (1) Initialization: The data requester sends a request to the cloud server CS A Issue data request, cloud server CS A Generate a public-private key pair (pk, sk) based on the data request and send the public key pk to the cloud server CS B ; Each participant sends the randomly generated perturbation data to the cloud server CS through a secure channel A , Cloud Server CS A Invert and encrypt the perturbation data, and then send the perturbation ciphertext to the cloud server CS B .

[0047] In this embodiment of the present invention, the initialization process specifically includes the following steps:

[0048] S11, Cloud Server CS A Generate a private key pair (pk, sk) using the Paillier encryption algorithm, where sk is the private key;

[0049] S12, Cloud Server CSA Send the public key pk to the cloud server CS B ;

[0050] S13. Each participant u i The randomly generated perturbation data Sent to the cloud server CS through a secure channel A ,in, Represents participant u i For the mth observation object O m Randomly generated α value, Represents participant u i For the mth observation object O m The β value is randomly generated, and the value of m ranges from 1 to M;

[0051] S14, Cloud Server CS A For perturbation data Invert and encrypt to get the disturbed ciphertext and Cloud Server CS A Send the perturbed ciphertext to the cloud server CS B .

[0052] (2) Data upload: Perception data observed by each participant and Perform disturbance and make the perturbation perception data Send to cloud server CS B , Cloud Server CS B Perception data after disturbance After encryption, the disturbance is eliminated based on the disturbance ciphertext to obtain the perception data ciphertext

[0053] In an embodiment of the present invention, the data uploading process specifically includes the following steps:

[0054] S21, Participant u i Using perturbed data and For the observed object o m Perception data and Add disturbance and send the perturbation data to and Send to cloud server CS B ,in,

[0055] S22, Cloud Server CS B Perception data after disturbance and Encryption to obtain the disturbed perception data ciphertext and And through the Paillier addition homomorphism property and Eliminate disturbances and obtain the ciphertext of the perception data and

[0056] (3) Weight update: In the cloud server CS A With the assistance of Cloud Server CS B Based on the observation object O calculated in the previous round m The true value is used to calculate the distance ciphertext of each observation object, and the weight ciphertext of each participant in the current round iteration t is calculated based on the distance ciphertext of all observation objects of each participant.

[0057] In an embodiment of the present invention, the weight updating process specifically includes the following steps:

[0058] S31, Cloud Server CS B For the observation object O in round t-1 m The true value of Square to encrypt and form the true value ciphertext Then calculate the distance ciphertext of all observation objects of each participant The calculation formula is as follows:

[0059]

[0060] Then, the cloud server CS B Aggregate the distance ciphertexts of all observed objects of each participant to obtain the distance ciphertext V of the participant i , And randomly generate a perturbation data r1 i , the distance ciphertext V of the participants i Add perturbation to obtain the distance ciphertext of the participants after perturbation

[0061] S32, Cloud Server CS B The distance ciphertext V of all participants i Perform aggregation to obtain the total distance ciphertext V all , Then generate a random perturbation data r2 to the total distance ciphertext V all Add perturbation to get the total distance ciphertext after perturbation Finally, Cloud Server CS B The disturbed ciphertext and Send to cloud server CS A;

[0062] S33, Cloud Server CS A First, the perturbed participant distance ciphertext and total distance ciphertext Decrypt and get the distance between participants after disturbance and in, Then the cloud server CS A According to the CRH weight calculation formula, each participant u in the current round of iteration is obtained i Weight after perturbation Last Cloud Server CS A right Encrypt the line and write the ciphertext Send to cloud server CS B ;

[0063] S34, Cloud Server CS B According to the known disturbance data r1 i and r2, the perturbed weighted ciphertext is encrypted by the Paillier additive homomorphic encryption algorithm Eliminate the disturbance and get the participant u in this round of iteration i The weighted ciphertext E pk (w i ),because therefore,

[0064] (4) True value update: In the cloud server CS B With the assistance of Cloud Server CS A According to the participants i The weighted ciphertext and observer u in the current iteration t i Ciphertext of the perception data of the measured object Get the true value of the observed object in the current iteration t.

[0065] In the embodiment of the present invention, the process of obtaining the true value of the observed object specifically includes the following steps:

[0066] S41, Cloud Server CS B Randomly generate perturbation data And encrypt the participants' perception data Adding perturbation yields Then according to the disturbance data r1 i The weighted ciphertext E of each participant pk (w i ) Add the perturbation to get E pk (r1 i ×w i), Finally, the disturbed perception data ciphertext and weighted ciphertext E pk (r1 i ×w i ) sent to the cloud server CS A ;

[0067] S42, Cloud Server CS A Perturbed ciphertext and E pk (r1 i ×w i ) to decrypt and get and r1 i ×w i , Cloud Server CS A Perform multiplication on the decrypted data and encrypt it to obtain the disturbed weighted data ciphertext And send it to the cloud server CS B ;

[0068] S43, Cloud Server CS B According to the known disturbance data and r1 i , through the Paillier addition homomorphic encryption algorithm Cancel the disturbance and obtain the weighted perception data ciphertext Cloud Server CS B Aggregate the weighted ciphertexts and weighted data ciphertexts of all participants to obtain the aggregated weighted ciphertexts of all participants and all participants on the observed object O m Aggregate weighted data ciphertext in, Last Cloud Server CS B Aggregate weight ciphertext and weighted data ciphertext Send to cloud server CS A ;

[0069] S44, Cloud Server CS A Aggregate weight ciphertext and aggregated weighted data ciphertext Decrypt to get the aggregate weight W all and aggregate weighted data VW m ,in, Cloud Server CS A According to the CRH truth value calculation formula, the true value of each observation object in the current round iteration t is obtained

[0070] (5) Cloud Server CS A Determine the truth value Whether the convergence conditions are met, if not, the cloud server CS A The truth value Send to cloud server CS B , and proceed as the next round of iteration, returning to step S3 until the calculated true value meets the convergence condition, and the final true value is returned to the data requester through a secure channel.

[0071] The present invention has been described exemplarily. Obviously, the specific implementation of the present invention is not limited to the above-mentioned method. As long as various non-substantial improvements are made using the method concept and technical solution of the present invention, or the concept and technical solution of the present invention are directly applied to other occasions without improvement, they are all within the scope of protection of the present invention.

Claims

1. A crowd-sensing secure truth value discovery method based on homomorphic encryption, characterized in that: The method specifically comprises the following steps: S1. Data requester sends data to cloud server CS A Issue data request, cloud server CS A Generate a public-private key pair (pk, sk) and send the public key pk to the cloud server CS B ; Each participant sends the randomly generated perturbation data to the cloud server CS through a secure channel A , Cloud Server CS A Invert and encrypt the perturbation data, and then send the perturbation ciphertext to the cloud server CS B ; S2. Each participant’s perception of the observed data and Perform disturbance and make the perturbation perception data Send to cloud server CS B , Cloud Server CS B Perception data after disturbance After encryption, the disturbance is eliminated based on the disturbance ciphertext to obtain the perception data ciphertext S3, in cloud server CS A With the assistance of Cloud Server CS B Based on the observation object O calculated in the previous round m The true value is used to calculate the distance ciphertext of each observation object in the current round t, and the weight ciphertext of each participant in the current round iteration t is calculated based on the distance ciphertext of all observation objects of each participant; S4, in the cloud server CS B With the assistance of Cloud Server CS A According to the participants i The true value of the observed object in the current round iteration t is obtained by using the weight ciphertext and the perception data ciphertext of the observed object in the current round iteration t; S5, Cloud Server CS A Determine the truth value Whether the convergence conditions are met, if not, the cloud server CS A The truth value Send to cloud server CS B , proceed as the next round of iteration, return to step S3, until the calculated true value meets the convergence condition, and return the final true value to the data requester; Step S3 includes the following steps: S31, Cloud Server CS B For the observation object O in round t-1 m The true value of Square to encrypt and form the true value ciphertext Then calculate the distance ciphertext of all observation objects of each participant Cloud Server CS B Aggregate the distance ciphertexts of all observed objects of each participant to obtain the distance ciphertext V of the participant i , and randomly generate a perturbation data The distance ciphertext V of the participants i Add perturbation to obtain the distance ciphertext of the participants after perturbation S32, Cloud Server CS B The distance ciphertext V of all participants i Perform aggregation to obtain the total distance ciphertext V all , then generate a random perturbation data r2 to the total distance ciphertext V all Add perturbation to get the total distance ciphertext after perturbation Cloud Server CS B The disturbed ciphertext and Send to cloud server CS A ; S33, Cloud Server CS A First, the perturbed participant distance ciphertext and total distance ciphertext Decrypt and get the distance between participants after disturbance and Cloud Server CS A Calculate the CRH weight for each participant u in the current round iteration i Weight after perturbation Cloud Server CS A right Encrypt the line and write the ciphertext Send to cloud server CS B ; S34, Cloud Server CS B According to the known disturbance data and r2, the perturbed weighted ciphertext is encrypted by the Paillier additive homomorphic encryption algorithm Eliminate the disturbance and get the participant u in the current round of iteration i The weighted ciphertext E pk (w i ); Step S4 specifically includes the following steps: S41, Cloud Server CS B Randomly generate perturbation data And encrypt the participants' perception data Adding perturbation yields Then according to the disturbance data r1 i The weighted ciphertext E of each participant pk (w i ) Add the disturbance to get Finally, the disturbed perception data ciphertext and weighted ciphertext Send to cloud server CS A ; S42, Cloud Server CS A Perturbed ciphertext and Decrypt and get and r1 i ×w i , Cloud Server CS A Perform multiplication on the decrypted data and encrypt it to obtain the disturbed weighted data ciphertext And send it to the cloud server CS B ; S43, Cloud Server CS B According to the known disturbance data and r1 i , through the Paillier addition homomorphic encryption algorithm Cancel the disturbance and obtain the weighted perception data ciphertext Cloud Server CS B Aggregate the weighted ciphertexts and weighted data ciphertexts of all participants to obtain the aggregated weighted ciphertexts of all participants and all participants on the observed object O m Aggregate weighted data ciphertext Cloud Server CS B Aggregate weight ciphertext and weighted data ciphertext Send to cloud server CS A ; S44, Cloud Server CS A Aggregate weight ciphertext and aggregated weighted data ciphertext Decrypt to get the aggregate weight W all and aggregate weighted data VW m ; Cloud Server CS A Calculate the true value of each observation object in the current round iteration t 2. The method for discovering secure truth values ​​based on crowd intelligence perception based on homomorphic encryption as claimed in claim 1, characterized in that: Step S1 specifically includes the following steps: S11, Cloud Server CS A Generate a private key pair (pk, sk) using the Paillier encryption algorithm; S12, Cloud Server CS A Send the public key pk to the cloud server CS B ; S13. Each participant u i The randomly generated perturbation data Sent to the cloud server CS through a secure channel A ,in, Represents participant u i For the mth observation object O m Randomly generated α value, Represents participant u i For the mth observation object O m Randomly generated beta values; S14, Cloud Server CS A For perturbation data Invert and encrypt to get the disturbed ciphertext and Cloud Server CS A Send the perturbed ciphertext to the cloud server CS B .

3. The method for discovering secure truth values ​​based on crowd intelligence perception based on homomorphic encryption as claimed in claim 1, characterized in that: Step S2 includes the following steps: S21, Participant u i Using perturbed data and For the observed object o m Perception data and Add disturbance and send the perturbation data to and Send to cloud server CS B ,in, S22, Cloud Server CS B Perception data after disturbance and Encryption to obtain the disturbed perception data ciphertext and And through the Paillier addition homomorphism property and Eliminate disturbances and obtain the ciphertext of the perception data and 4. The method for discovering secure truth values ​​based on crowd intelligence perception based on homomorphic encryption as claimed in claim 1, characterized in that: The distance ciphertext of the observed object in the current round t The calculation formula is as follows:

5. The method for discovering secure truth values ​​based on crowd intelligence perception based on homomorphic encryption as claimed in claim 1, characterized in that: Weight The calculation formula is as follows:

6. The method for discovering secure truth values ​​of crowd-sensing based on homomorphic encryption as claimed in claim 1, characterized in that: Weighted ciphertext E pk (w i ) is calculated as follows:

7. The method for discovering secure truth values ​​of crowd-sensing based on homomorphic encryption according to claim 1, characterized in that: truth value The calculation formula is as follows:

Images