A privacy protection task allocation method and system for mobile group cognitive perception

Abstract

This invention discloses a privacy-preserving task allocation method and system for mobile crowdsourcing sensing, comprising: a trusted institution generating system parameters, key pairs, and a Beaver triple pool, and pre-distributing them to two fog servers; the task requester and worker respectively using a planar Laplace mechanism to differentially privacy-perturb the location, and sharing the required skill vector and the skill vector bitwise XORed to secretly form two shares, which are encrypted with the public keys of the two fog servers and uploaded to a cloud platform; the cloud platform filters candidate workers based on the perturbation distance, and distributes the candidate worker information and coverage state shares to the two fog servers; the two servers collaboratively perform secure bitwise AND and secure bitwise OR calculations using the triple shares, iteratively updating the coverage state shares, and returning them to the cloud platform after random scrambling; the cloud platform restores the scrambled coverage state and selects collaborating workers based on its Hamming weight. This invention can achieve efficient task allocation while protecting location and skill privacy.

Description

Technical Field

[0001] This invention relates to the field of crowd sensing task allocation technology, and in particular to a task allocation method and system for mobile crowd sensing with multi-skill collaborative privacy protection. Background Technology

[0002] With the development of mobile internet, smart terminals, and the Internet of Things (IoT), Mobile Crowd Sensing (MCS) has become an important means of acquiring large-scale environmental information and crowd behavior data. By mobilizing participants distributed across cities, transportation networks, or specific areas, MCS can collect various information such as environmental perception, traffic flow, air quality, and public facility usage. Participants typically collect data through smartphones, in-vehicle devices, or wearable terminals and upload it to a cloud platform or central server via mobile networks.

[0003] In MCS (Multi-Channel System), task allocation is one of the core issues. Task allocation involves rationally assigning tasks published by the publisher to suitable participants to meet the task requirements. Traditional task allocation methods typically consider factors such as task location, participant location, participant preferences, and budget costs, and are generally suitable for simple tasks completed by a single worker. However, as perception tasks have evolved from single-worker tasks to more complex forms involving multi-dimensional capabilities and cross-domain collaboration, more and more MCS applications are requiring multiple participants to collaborate on tasks at the same task location or in the same task scenario. Tasks often involve multiple skill requirements, and a single worker may only possess some of these skills. The platform needs to select a collaborative set from candidate workers whose skill union covers the task requirements and complete the allocation and scheduling while satisfying spatiotemporal constraints.

[0004] With increasing demands for data privacy protection, sensitive information such as location and skills involved in task allocation may be at risk of leakage. For example, publicly disclosing the geographical location of a task may expose participants' daily whereabouts; publicly disclosing skill requirements or participants' abilities may reveal personal occupational or interest information. Traditional task allocation schemes lack effective protection measures when handling this privacy information, easily leading to security and privacy issues and limiting the widespread adoption of MCS systems in practical applications. Existing research on privacy-preserving multi-skill collaborative task allocation methods is still limited, and a systematic solution has not yet been formed. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a privacy-preserving task allocation method and system for mobile crowd intelligence sensing, thereby solving the technical problem of lack of privacy protection in the task allocation process in existing technologies.

[0006] This invention provides a privacy-preserving task allocation method for mobile crowd sensing, comprising the following steps:

[0007] Step 1: The trusted institution generates a master key and asymmetric key pairs for the two fog servers, and generates a Beaver triple pool offline and pre-distributes it to the two fog servers;

[0008] Step 2: The task requester generates a random scrambling rule, performs differential privacy perturbation on the real location to obtain the perturbation location of the task; performs a bitwise XOR secret sharing on the required skill vector to obtain two required shares; encrypts the first required share and the scrambling rule with the first fog server's public key to obtain the first required ciphertext; encrypts the second required share and the scrambling rule with the second fog server's public key to obtain the second required ciphertext; uploads the task identifier, the total number of required skills, the perturbation location, the distance threshold, and the two required ciphertexts to the cloud platform.

[0009] Step 3: Each worker performs differential privacy perturbation on their real location to obtain the worker's perturbed location. The skill vector is XORed bitwise and secretly shared to obtain two skill shares. The first skill share is encrypted with the first fog server's public key to obtain the first skill ciphertext. The second skill share is encrypted with the second fog server's public key to obtain the second skill ciphertext. The worker identifier, perturbed location, and two skill ciphertexts are uploaded to the cloud platform.

[0010] Step 4: The cloud platform forms a candidate worker set for workers whose perturbation distance between the task's perturbation location and the worker's perturbation location is less than a threshold; initializes the coverage state and XORs it with secret to share it into two shares; sends the worker identifier, corresponding skill ciphertext, and coverage state share in the candidate worker set to two fog servers.

[0011] Step 5: The two fog servers decrypt to obtain their respective shares; for each candidate worker, secure bitwise AND and secure bitwise OR calculations are performed sequentially using the triplet share to iteratively update the coverage state share, and after being scrambled in accordance with the scrambling rules, it is sent to the cloud platform.

[0012] Step 6: The cloud platform performs a bitwise XOR operation to restore the scrambled coverage share and calculates the Hamming weight of the coverage share as the current coverage quantity. If the current coverage quantity is greater than or equal to the total number of required skills, the selected workers are used as a collaboration set to complete the task allocation; otherwise, return to Step 5.

[0013] Furthermore, in steps 2 and 3, a planar Laplace mechanism is used to perform differential privacy perturbation on the real location, specifically as follows:

[0014] By adding random noise that follows a planar Laplace distribution to the true location, a perturbed location can be obtained.

[0015] Furthermore, in step 2, the random scrambling rule is a randomly generated sorting from the complete skill set index to the scrambled index.

[0016] Furthermore, in step 2, the process of generating the two demand shares is as follows:

[0017] The task requester randomly generates the first demand share, specifically as follows: Where m is the total number of skills;

[0018] The second demand share is generated by XORing the first demand share with the demand skill vector, specifically as follows:

[0019] ;

[0020] In the formula, It represents the second largest share of demand; This represents the required skills vector. It represents the largest share of demand.

[0021] Furthermore, in step 3, the process of generating the two skill shares is as follows:

[0022] Workers are randomly assigned a first skill share, specifically: Where i is the worker number; m is the total number of skills;

[0023] The second skill share is generated by XORing the first skill share with the skill vector.

[0024] ;

[0025] In the formula, For the second skill share; For skill vectors; The first skill share.

[0026] Furthermore, in step 5, the specific process of the secure bitwise AND calculation is as follows:

[0027] Each of the two fog servers calculates its own correction value, using the following formula:

[0028] ; ;

[0029] ; ;

[0030] The two fog servers exchange correction values, each calculating its own XOR correction value. The specific formula is as follows:

[0031] ;

[0032] ;

[0033] The specific formula for calculating the output share of two fog servers is as follows:

[0034] ;

[0035] ;

[0036] The formula for calculating the result of the security bitwise AND operation is:

[0037] ;

[0038] In the formula, , These are the correction values ​​calculated by the two fog servers to remove the random mask of worker skills; , These are the correction values ​​calculated for the two fog servers to eliminate the random mask required for the task; , Each has two skill shares; , These represent two demand shares; , , , , , For triples The elements in ; This is the XOR correction value used to eliminate the random mask of worker skills; This is the XOR correction value used to eliminate the random mask required by the task; , The share of output for the two fog servers; For XOR operation; For operations; s is the skill vector; r is the required skill vector.

[0039] Furthermore, in step 5, the formula for the result of the secure bitwise OR calculation is:

[0040] ;

[0041] In the formula, a and b are the two binary input bits to be performed in the secure bitwise OR operation; For XOR operation; For operation.

[0042] Furthermore, in step 6, the specific process for selecting collaborative workers includes: for each candidate worker not selected in the candidate worker set, the two fog servers obtain the candidate coverage state after adding the candidate worker based on the current coverage state share, the secure bitwise AND calculation result, and the secure bitwise OR calculation result; the cloud platform calculates the marginal coverage gain of the candidate worker based on the Hamming weight of the candidate coverage state, and selects the candidate worker with the largest marginal coverage gain to join the collaborative set; when multiple candidate workers have the same marginal coverage gain, the candidate worker with the smaller perturbation distance is selected to join the collaborative set.

[0043] Furthermore, the Beaver triple pool is managed using a synchronous indexing method. When the two fog servers perform a secure bitwise AND calculation each time, they call the Beaver triple share at the same index position and mark the corresponding Beaver triple as used after the call. For secure bitwise AND calculations of different skill positions, different candidate workers, or different iteration rounds, unused Beaver triple shares are called to avoid privacy leaks caused by the reuse of Beaver triples.

[0044] This invention also provides a privacy-preserving task allocation system for mobile crowd sensing, comprising:

[0045] Trusted institution, used for system initialization, key generation, offline generation and distribution of Beaver triple pool;

[0046] The task requester is used to perform differential privacy perturbation on the real location of the task, XOR the required skill vector into two shares, generate random scrambling rules, and encrypt and upload them to the cloud platform.

[0047] Workers use differential privacy perturbation to determine their real location, XOR their skills into two shares, and encrypt and upload them to the cloud platform.

[0048] The first fog server and the second fog server are used to decrypt the shares. Based on the pre-stored Beaver triples, they perform secure bitwise AND and secure bitwise OR calculations, iteratively update the overlay state shares, and use random scrambling rules to consistently scramble the overlay state shares before uploading them to the cloud platform.

[0049] The cloud platform is used to filter candidate workers based on the location of the disturbance, forward the share ciphertext to two fog servers, receive the scrambled coverage status share uploaded by the fog servers, and complete the selection of collaborative workers and task allocation based on the coverage quantity and the iteration termination condition.

[0050] The beneficial effects of this invention are:

[0051] This invention addresses the problem of sensitive skill information in multi-skill collaborative tasks, where platforms struggle to select collaboration partners without revealing the plaintext of skills. It proposes a privacy-preserving task allocation framework based on dual-fog server collaborative computing. By using XOR secret sharing, the task requirement skills and worker skills are separated and stored on two fog servers, making it impossible for any single server to recover the plaintext of skills.

[0052] This invention combines the Beaver triplet construction secure Boolean computation protocol to complete skill matching, coverage judgment and collaborative worker selection without restoring the plaintext of skills, effectively reducing the risk of the platform inferring the skill structure during the iterative selection process.

[0053] This invention introduces a planar Laplace mechanism to perform differential privacy perturbations on the positions of tasks and workers, protecting position privacy while preserving sufficient position availability for candidate selection; at the same time, it introduces a random scrambling rule, so that the cloud platform can only obtain the scrambled coverage state and its Hamming weight, and cannot know the semantic information of specific skill bits, thereby further enhancing skill privacy protection. Attached Figure Description

[0054] The features and advantages of the invention will be more clearly understood by referring to the accompanying drawings, which are schematic and should not be construed as limiting the invention in any way. In the drawings:

[0055] Figure 1 This is a flowchart of a specific embodiment of the present invention;

[0056] Figure 2 This is a timing diagram of a specific embodiment of the present invention. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] The present invention will be further illustrated below with reference to specific embodiments. Those skilled in the art should understand that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Modifications to the present invention in various equivalent forms all fall within the scope defined by the appended claims.

[0059] like Figure 1 , Figure 2 As shown, this invention provides a privacy-preserving task allocation method for mobile crowd sensing, comprising the following steps:

[0060] Step 1: The trusted authority (TA) generates a master key and uses the RSA-2048 asymmetric encryption algorithm to generate two asymmetric key pairs (K) for the fog servers. A ,k A ) and (K B ,k B The public keys are distributed to the first fog service (FSA) and the second fog service (FSB) respectively, and are made public. TA generates Beaver triples offline. ,in, It is a random bit, and Each triple is split into two parts by XOR secret sharing (u A ,u B ), (v A ,v B ), (w A ,w B ), which are pre-distributed to the FSA and FSB respectively;

[0061] Step 2: The task requester (TR) generates a task identifier. Real location Skills Requirement Vector Total number of skills required and distance threshold ,in, This represents the Hamming weight function, used to count the number of bits with a value of 1 in a binary vector; TR randomly generates a permutation As a rule for disorder.

[0062] TR to Perform a plane Laplace perturbation to obtain the perturbation position for TR. ;

[0063] TR randomly generates the first demand share. :

[0064] ;

[0065] First demand share With demand skills vector The second demand share is generated after XORing. Specifically:

[0066] ;

[0067] Encrypted with FSA public key Encrypted using the FSB public key ;

[0068] The generated task information is as follows:

[0069] ;

[0070] Upload the task information to the cloud platform CSP.

[0071] Step 3: Each worker Holding the real position and skill vectors Workers Performing a plane Laplace perturbation yields the target Disturbance location .

[0072] Workers are randomly assigned a first skill share, specifically: Where i is the worker number; m is the total number of skills;

[0073] First skill share With skill vectors Generating a second skill share after XOR Specifically:

[0074] ;

[0075] Encrypted with FSA public key Encrypted using the FSB public key ;

[0076] The generated worker information is as follows:

[0077] ;

[0078] in, (·) indicates the use of the first fog server's public key. The executed asymmetric encryption function; (·) indicates the use of the second fog server public key. The executed asymmetric encryption function; Let i be the ID number of the i-th worker;

[0079] Worker information is uploaded to the cloud platform CSP.

[0080] Step 4: Location of disturbances in CSP computing tasks on the cloud platform Disturbance position of workers Distance between disturbances:

[0081] ;

[0082] in, The Euclidean distance function;

[0083] choose The workers constitute the candidate set Sort by distance in ascending order.

[0084] CSP Initialize Coverage Status Share: Randomly generate the first coverage state share The first coverage share Share of initial coverage status XOR generates cover state shares Specifically:

[0085] ;

[0086] in, Used to record the skill coverage of currently selected collaborating workers, when At the beginning, all skill slots were not covered;

[0087] Subsequently, CSP will:

[0088] ;

[0089] ;

[0090] as well as Send to FSA;

[0091] Will:

[0092] ;

[0093] ;

[0094] as well as Send to FSB;

[0095] in, OR operation; , , as well as The information packet sent by the cloud platform CSP to the AB server; For splicing operations;

[0096] In addition, the CSP needs to send the candidate worker ID list to the two fog servers in order.

[0097] Step 5: The two fog servers decrypt to obtain their respective shares; for each candidate worker, secure bitwise AND and secure bitwise OR calculations are performed sequentially using the triplet share to iteratively update the coverage state share, and after being scrambled in accordance with the scrambling rules, it is sent to the cloud platform.

[0098] Safety bitwise operations and calculation process:

[0099] With the first Taking a wheel as an example, the subscript Indicates the first For the i-th worker in the round, let the current coverage state be denoted as . Its share is The execution process of the two fog servers is as follows. For ease of description, the following formulas all use the binary bits of any candidate worker at any skill level as the calculation object:

[0100] Safety bitwise operations and calculation process:

[0101] Each of the two fog servers calculates its own correction value, using the following formula:

[0102] ; ;

[0103] ; ;

[0104] The two fog servers exchange correction values, each calculating its own XOR correction value. The specific formula is as follows:

[0105] ;

[0106] ;

[0107] The specific formula for calculating the output share of two fog servers is as follows:

[0108] ;

[0109] ;

[0110] The formula for calculating the result of the security bitwise AND operation is:

[0111] ;

[0112] In the formula, , These are the correction values ​​calculated by the two fog servers to remove the random mask of worker skills; , These are the correction values ​​calculated for the two fog servers to eliminate the random mask required for the task; , Each has two skill shares; , These represent two demand shares; , , , , , For triples The elements in ; This is the XOR correction value used to eliminate the random mask of worker skills; This is the XOR correction value used to eliminate the random mask required by the task; , The share of output for the two fog servers; For XOR operation; For operation; For the first The skill vector of the i-th worker in the round; r is the required skill vector;

[0113] The formula for calculating the result of a secure bitwise OR operation is:

[0114] ;

[0115] in, OR operation;

[0116] Using the identity transformation, we get:

[0117] ;

[0118] in, ;

[0119] The share is obtained by calling the secure bitwise AND protocol again. ;

[0120] The two fog servers perform local computation:

[0121] ;

[0122] ;

[0123] Scrambling and Uploading Process:

[0124] The two fog servers follow the scrambling rules right and Perform a consistent bit permutation to obtain and Each of these is uploaded to the CSP.

[0125] Step 6: CSP receives and Then, the coverage share is restored by bitwise XOR, with the specific formula as follows:

[0126] ;

[0127] Calculate coverage status share Hamming weight The specific formula for determining the current coverage quantity is as follows:

[0128] ;

[0129] when At that time, then order Continue evaluating the next candidate worker; otherwise, terminate the iteration, take the selected worker set as collaborating workers, send task instructions to these workers, and complete the task assignment.

[0130] This invention also provides a privacy-preserving task allocation system for mobile crowd sensing, comprising:

[0131] Trusted institution, used for system initialization, key generation, offline generation and distribution of Beaver triple pool;

[0132] The task requester is used to perform differential privacy perturbation on the real location of the task, XOR the required skill vector into two shares, generate random scrambling rules, and encrypt and upload them to the cloud platform.

[0133] Workers use differential privacy perturbation to determine their real location, XOR their skills into two shares, and encrypt and upload them to the cloud platform.

[0134] The first fog server and the second fog server are used to decrypt the shares. Based on the pre-stored Beaver triples, they perform secure bitwise AND and secure bitwise OR calculations, iteratively update the overlay state shares, and use random scrambling rules to consistently scramble the overlay state shares before uploading them to the cloud platform.

[0135] The cloud platform is used to filter candidate workers based on the location of the disturbance, forward the share ciphertext to two fog servers, receive the scrambled coverage status share uploaded by the fog servers, and complete the selection of collaborative workers and task allocation based on the coverage quantity and the iteration termination condition.

[0136] Security analysis shows that, under a semi-honest and non-colluding threat model, this invention can protect the privacy of task-required skills and worker skills, while protecting location information through location differential privacy perturbation and preventing cloud platforms from inferring the semantic structure of skills through random scrambling.

[0137] Experimental results show that when the skill dimension is 40 and the number of workers is 400, the task allocation success rate of this invention can reach over 95%, the online running time of a single allocation is in the millisecond range, and the communication overhead is reduced by about 85% compared with similar schemes based on asymmetric encryption.

[0138] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A privacy-preserving task allocation method for mobile crowd sensing, characterized in that, Includes the following steps: Step 1: The trusted organization generates system parameters and asymmetric key pairs for the two fog servers, and generates a Beaver triple pool offline and pre-distributes it to the two fog servers; Step 2: The task requester generates a random scrambling rule, performs differential privacy perturbation on the real location to obtain the perturbation location of the task; performs a bitwise XOR secret sharing on the required skill vector to obtain two required shares; encrypts the first required share and the scrambling rule with the first fog server's public key to obtain the first required ciphertext; encrypts the second required share and the scrambling rule with the second fog server's public key to obtain the second required ciphertext; uploads the task identifier, the total number of required skills, the perturbation location, the distance threshold, and the two required ciphertexts to the cloud platform. Step 3: Each worker performs differential privacy perturbation on their real location to obtain the worker's perturbed location. The skill vector is XORed bitwise and secretly shared to obtain two skill shares. The first skill share is encrypted with the first fog server's public key to obtain the first skill ciphertext. The second skill share is encrypted with the second fog server's public key to obtain the second skill ciphertext. The worker identifier, perturbed location, and two skill ciphertexts are uploaded to the cloud platform. Step 4: The cloud platform forms a candidate worker set for workers whose perturbation distance between the task's perturbation location and the worker's perturbation location is less than a threshold; initializes the coverage state and XORs it with secret to share it into two shares; sends the worker identifier, corresponding skill ciphertext, and coverage state share in the candidate worker set to two fog servers. Step 5: The two fog servers decrypt to obtain their respective shares; for each candidate worker, secure bitwise AND and secure bitwise OR calculations are performed sequentially using the triplet share to iteratively update the coverage state share, and after being scrambled in accordance with the scrambling rules, it is sent to the cloud platform. Step 6: The cloud platform performs a bitwise XOR operation to restore the scrambled coverage state share. Based on the coverage state share, it selects collaborating workers from the candidate workers and calculates the Hamming weight of the coverage state share as the current coverage quantity. When the current coverage quantity is greater than or equal to the total number of required skills, the selected collaborating workers are formed into a collaborating set to complete the task allocation; otherwise, it returns to Step 5.

2. The privacy-preserving task allocation method for mobile crowd sensing as described in claim 1, characterized in that, In steps 2 and 3, a planar Laplace mechanism is used to perform differential privacy perturbation on the real location, specifically as follows: By adding random noise that follows a planar Laplace distribution to the true location, a perturbed location can be obtained.

3. The privacy-preserving task allocation method for mobile crowd sensing as described in claim 1, characterized in that, In step 2, the random scrambling rule is a randomly generated sorting from the complete skill set index to the scrambled index.

4. The privacy-preserving task allocation method for mobile crowd sensing as described in claim 1, characterized in that, In step 2, the process of generating the two demand shares is as follows: The task requester randomly generates the first demand share, specifically as follows: Where m is the total number of skills; The second demand share is generated by XORing the first demand share with the demand skill vector, specifically as follows: ； In the formula, It represents the second largest share of demand; This represents the required skills vector. It represents the largest share of demand.

5. The privacy-preserving task allocation method for mobile crowd sensing as described in claim 1, characterized in that, In step 3, the process of generating the two skill shares is as follows: Workers are randomly assigned a first skill share, specifically: Where i is the worker number; m is the total number of skills; The second skill share is generated by XORing the first skill share with the skill vector. ； In the formula, For the second skill share; For skill vectors; The first skill share.

6. The privacy-preserving task allocation method for mobile crowd sensing as described in claim 1, characterized in that, In step 5, the specific process of the secure bitwise AND calculation is as follows: Each of the two fog servers calculates its own correction value, using the following formula: ； ； The two fog servers exchange correction values, each calculating its own XOR correction value. The specific formula is as follows: ； ； The specific formula for calculating the output share of two fog servers is as follows: ； ； The formula for calculating the result of the security bitwise AND operation is: ； In the formula, , These are the correction values ​​calculated by the two fog servers to remove the random mask of worker skills; , These are the correction values ​​calculated for the two fog servers to eliminate the random mask required for the task; , Each has two skill shares; , These represent two demand shares; , , , , , For triples The elements in ; This is the XOR correction value used to eliminate the random mask of worker skills; This is the XOR correction value used to eliminate the random mask required by the task; , The share of output for the two fog servers; For XOR operation; For operation; s is the skill vector; r is the required skill vector.

7. The privacy-preserving task allocation method for mobile crowd sensing as described in claim 1 or 6, characterized in that, In step 5, the formula for the result of the secure bitwise OR calculation is: ； In the formula, a and b are the two binary input bits to be performed in the secure bitwise OR operation; For XOR operation; For operation.

8. The privacy-preserving task allocation method for mobile crowd sensing as described in claim 1, characterized in that, In step 6, the specific process of selecting collaborative workers includes: for each candidate worker not selected in the candidate worker set, the two fog servers obtain the candidate coverage state after adding the candidate worker based on the current coverage state share, the secure bitwise AND calculation result, and the secure bitwise OR calculation result; the cloud platform calculates the marginal coverage gain of the candidate worker based on the Hamming weight of the candidate coverage state, and selects the candidate worker with the largest marginal coverage gain to join the collaborative set; when multiple candidate workers have the same marginal coverage gain, the candidate worker with the smaller perturbation distance is selected to join the collaborative set.

9. The privacy-preserving task allocation method for mobile crowd sensing as described in claim 1 or 6, characterized in that, The Beaver triple pool is managed using a synchronous indexing method. When the two fog servers perform a secure bitwise AND calculation, they call the Beaver triple share at the same index position and mark the corresponding Beaver triple as used after the call. For secure bitwise AND calculations of different skill positions, different candidate workers, or different iteration rounds, unused Beaver triple shares are called.

10. A privacy-preserving task allocation system for mobile crowd sensing, applicable to the privacy-preserving task allocation method for mobile crowd sensing as described in claim 1, characterized in that, The system includes: Trusted institution, used for system initialization, key generation, offline generation and distribution of Beaver triple pool; The task requester is used to perform differential privacy perturbation on the real location of the task, XOR the required skill vector into two shares, generate random scrambling rules, and encrypt and upload them to the cloud platform. Workers use differential privacy perturbation to determine their real location, XOR their skills into two shares, and encrypt and upload them to the cloud platform. The first fog server and the second fog server are used to decrypt the shares. Based on the pre-stored Beaver triples, they perform secure bitwise AND and secure bitwise OR calculations, iteratively update the overlay state shares, and use random scrambling rules to consistently scramble the overlay state shares before uploading them to the cloud platform. The cloud platform is used to filter candidate workers based on the location of the disturbance, forward the share ciphertext to two fog servers, receive the scrambled coverage status share uploaded by the fog servers, and complete the selection of collaborative workers and task allocation based on the coverage quantity and the iteration termination condition.

Images