A privacy-preserving task allocation method based on task content

Abstract

The present invention relates to a privacy-preserving task allocation method based on task content, and belongs to the field of mobile crowd-intelligence-aware privacy protection technology. The method uses polynomial functions and random matrix multiplication to realize privacy-preserving task allocation based on task content. Among them, by constructing a polynomial function to represent user task information, a random matrix to protect the task information of the data requester, and a privacy-preserving task allocation that performs task retrieval operations on the random matrix based on the matrix properties. The method supports task allocation functions based on privacy protection, threshold task allocation functions, task restoration functions, and task allocation functions based on access control. Compared with the prior art, the method has significant advantages in data privacy, retrieval efficiency, and ciphertext retrieval accuracy.

Description

Technical Field

[0001] The present invention relates to a privacy protection task allocation method based on task content, and belongs to the technical field of mobile crowd intelligence perception privacy protection. Background Art

[0002] With the development of 5G / 6G communication technology and the widespread popularity of various perception sensors, mobile crowd sensing has become an effective way to perceive and collect data.

[0003] Currently, there are many commercial mobile crowd-sensing platforms that can establish connections between data requesters and perception users, such as Amazon Mechanical Turk, WAZE, Be My Eyes, Witmart, Zhihu, etc., and have spawned a large number of applications such as environmental perception, health data collection, public opinion collection, text translation, etc. Since perception platforms usually have a large number of users and tasks, it is particularly important for perception platforms to assign different tasks to appropriate users in order to ensure the efficiency and accuracy of data collection and service provision.

[0004] In recent years, the "task allocation" problem in mobile crowd sensing has attracted widespread attention. In the existing mobile crowd sensing task allocation solution, each data requester sends his task requirements (usually expressed by task keywords) to the perception platform, and then the perception platform searches its own database to find perception users that meet the data requester's task requirements. Although this method is very effective, performing task retrieval on task plaintext may lead to serious privacy leakage problems. In actual mobile crowd sensing applications, the task information uploaded by users is often closely related to their identity, health status, location, hobbies and other sensitive information. For example, a driver's participation in a traffic information collection task can help other users better understand the current traffic conditions, but his own location, trajectory and other information may be inferred from the tasks he participates in; a patient with a certain disease can help medical institutions evaluate the therapeutic effect of new drugs by participating in new drug testing, but his participation in the task of new drug testing may disclose his health status or economic status. In addition, as a third-party platform, the perception platform is usually not fully trustworthy, and it may leak the user's task information to other institutions to obtain more benefits. If the task information is not effectively protected, users may be reluctant to send their data to the perception platform, which greatly affects the healthy development of mobile crowd sensing.

[0005] An intuitive way to achieve privacy-preserving task allocation is that the user first encrypts the task information, and then the perception platform performs task retrieval in ciphertext form. Therefore, searchable encryption is a promising technology for implementing privacy-preserving task allocation based on task content. However, most of the current solutions based on searchable encryption technology adopt the trusted data center model, that is, these solutions rely on a trusted data center to encrypt the data and distribute the search key to the search user. This implementation mode has some inherent defects and is therefore not suitable for mobile crowdsensing scenarios. First, the perception user needs to upload his or her data to the trusted data center first, and then the trusted data center encrypts it and uploads it to the perception platform. This form of data encryption and transmission will reduce the performance of the system and increase the risk of data leakage. Second, because the search users all have the same search key, some users holding the search key may restore the plaintext information from the search ciphertext of other users. Currently, there are also some public key searchable encryption schemes that can implement privacy-preserving task allocation based on task content without introducing a trusted data center. However, these schemes either cannot achieve a high degree of privacy protection or cannot achieve a high task retrieval efficiency.

[0006] Gong et al. proposed a task assignment scheme that protects user privacy. The scheme formally transforms task recommendation into an optimization problem of privacy, practicality and efficiency, and proposes a greedy algorithm to achieve an approximate optimal solution for task assignment. However, the scheme also does not strictly protect the task privacy of users and may leak sensitive information such as user identity, location, and hobbies. In order to achieve privacy protection of task information and task assignment based on task content, Shu et al. proposed several task assignment schemes that protect task privacy to solve problems such as single task matching, multi-task matching, and witch attacks in task assignment. In particular, the scheme designed by Shu et al. based on secure kNN achieves higher computational efficiency compared with existing schemes. It uses matrix decomposition to split the key of the secure kNN algorithm, uses polynomial functions to generate task indexes and query tokens, and uses random matrices to encrypt task information. Although the scheme can achieve higher computational efficiency, it assumes that all users are completely trustworthy, which is a strong security assumption in real mobile crowdsensing scenarios. In fact, some users may obtain private data of other users based on existing information such as keys.

[0007] In order to achieve more secure task allocation, Ni et al. designed a privacy-preserving task matching scheme based on random matrix multiplication technology. Compared with the scheme designed by Shu et al. based on secure kNN, this scheme can resist chosen plaintext attacks; compared with the bilinear mapping privacy-preserving task allocation scheme, this scheme has higher computational efficiency. However, in this scheme, the matrix dimension increases with the number of task types, which means that when there are many types of tasks, this scheme still faces large computational and communication overheads. Summary of the invention

[0008] The purpose of the present invention is to solve the security and privacy issues faced by privacy-preserving task allocation based on task content in mobile crowd intelligence perception scenarios in order to address the deficiencies in the prior art and creatively propose a privacy-preserving task allocation method based on task content. The method supports task allocation functions based on privacy protection, threshold task allocation functions, task restoration functions, and task allocation functions based on access control.

[0009] The innovation of this method is to use polynomial functions and random matrix multiplication to realize privacy-preserving task allocation based on task content. In particular, the privacy-preserving task allocation is constructed by constructing a polynomial function to represent user task information, protecting the task information of data requesters based on random matrices, and performing task retrieval operations on random matrices based on matrix properties.

[0010] In order to achieve the above object, the present invention adopts the following technical scheme.

[0011] First, the relevant concepts are explained.

[0012] Definition 1: Trusted Authority (TA)

[0013] The trusted center is responsible for generating relevant keys for entities participating in task allocation. When the system initialization is completed, the trusted center is offline or dormant.

[0014] Definition 2: Data Requester (DR)

[0015] Data requesters have task requirements. Limited by their own capabilities and data collection costs, data requesters upload tasks to the perception platform and find perception users who meet the task requirements through the perception platform.

[0016] Definition 3: Crowdsensing Platform (CP)

[0017] The perception platform is responsible for establishing connections between data requesters and perception users. The perception platform publishes tasks for data requesters and recruits perception users who meet their task requirements. Usually, the perception platform is a third-party platform operated by a company or organization.

[0018] Definition 4: Sensing User (SU)

[0019] To improve the efficiency of data collection, the perception user sends the task of interest to the perception platform. After receiving the task from the perception platform, the perception user uses the mobile device to collect the corresponding data and uploads it to the perception platform. The perception user is usually an individual holding a mobile device.

[0020] Definition 5: Passive Aggression

[0021] Passive attack means that the adversary can only obtain ciphertext information. This attack is similar to the ciphertext-only attack in cryptography. This attack may be launched when the adversary monitors a user's communication channel or the sensing platform receives the user's ciphertext data.

[0022] Definition 6: Active Attack

[0023] In an active attack, the adversary may obtain more information besides the ciphertext information.

[0024] A specific active attack mode is given from the perspective of attackers inside the perception platform and non-perception platform. For the perception platform, the perception platform may obtain the correspondence between some known tasks and ciphertexts. This attack is similar to the known plaintext attack in cryptography; for attackers inside the non-perception platform, they may try to obtain sensitive information of other users based on information such as task dictionaries and distribution keys. When the user's key is stored in a physical device (such as a smart card), once the adversary (such as a user with encryption or retrieval keys) obtains the smart card of other users, the adversary may choose special plaintext as input, observe the output ciphertext, and try to obtain the key information in the smart card based on these plaintext-ciphertext pairs. This attack is similar to the chosen plaintext attack in cryptography.

[0025] A privacy-preserving task allocation method based on task content includes initialization, task encryption, task re-encryption, trapdoor generation, trapdoor re-encryption and task retrieval.

[0026] Step 1: System initialization: The trusted center generates encryption keys and re-encryption keys for each entity participating in the system.

[0027] Given the highest power n of the polynomial function, the trusted center generates the key:

[0028] {M1,M2}

[0029] Among them, M1 represents master key 1, and M2 represents master key 2.

[0030] For the perceived user u k , the trusted center generates the encryption key:

[0031]

[0032]

[0033] For data requester u j , the trusted center generates the encryption key:

[0034]

[0035]

[0036] Step 2: Task encryption. The perception user determines the tasks of interest and encrypts them and sends them to the perception platform.

[0037] Suppose the task that the perceived user is interested in is Where m represents the task number, T k,m Represents the perceived user u k The value corresponding to the mth task in the task set in the task dictionary Task, l represents the perceived user u k The total number of tasks of interest.

[0038] Perceived user u k First generate the polynomial function Choose a random number r k ≠0, the main diagonal elements of (n+1)*(n+1) are (r k a0,r k a1,r k a2,…,r k a n ) of the lower triangular random matrix τ k , and encrypt it as:

[0039] E[τ k ]=A k,1 *τ k *A k,2

[0040] Step 3: Task re-encryption: After receiving the ciphertext information uploaded by the perception user, the perception platform performs the task re-encryption operation.

[0041] Given the encrypted task information E[τ k ], the perception platform uses the re-encryption key (A′ k,1 ,A′ k,2) re-encrypts the encrypted task to:

[0042] RE[τ k ]=A′ k,1 *E[τ k ]*A′ k,2

[0043] Step 4: Trapdoor generation: The data requester generates a trapdoor for the task information it needs.

[0044] Given a data requester u j Required tasks, u j Determine task information T j,m ←Task(m), select a random number r j,m ≠0 generates a (n+1)*(n+1) main diagonal element as The lower triangular random matrix τ j,m , and generate a trapdoor:

[0045] T[τ j,m ]=B j,1 *τ j,m *B j,2

[0046] Step 5: Trapdoor re-encryption. After receiving the trapdoor, the perception platform uses the re-encryption key to re-encrypt the trapdoor.

[0047] After receiving the trapdoor, the perception platform uses the re-encryption key (B′ j,1 ,B′ j,2 ) re-encrypts the trapdoor to:

[0048] RT[τ j,m ]=B′ j,1 *T[τ j,m ]*B′ j,2

[0049] Step 6: Task retrieval. The perception platform uses the re-encrypted trapdoor to perform task retrieval operations on the re-encrypted task information stored in its database.

[0050] Based on the re-encrypted tasks and trapdoors, the perception platform calculates:

[0051] Δ k,(j,m) =tr(RE[τ k ]*RT[τ j,m ])

[0052] If Δ k,(j,m) =0, then That is, the perceived user u k To data requester u j Task T j,mOtherwise, the perceived user u k Data requester u is not satisfied j mission requirements.

[0053] Finally, the perception platform connects data requesters and perception users.

[0054] Beneficial Effects

[0055] Compared with existing technologies, this method has significant advantages in data privacy, retrieval efficiency and ciphertext retrieval accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a schematic diagram for implementing the method of the present invention. DETAILED DESCRIPTION

[0057] The present invention is further described in detail below with reference to the accompanying drawings.

[0058] like Figure 1 As shown, a privacy-preserving task allocation method based on task content includes the following steps:

[0059] Step 1: System initialization: The trusted center generates encryption keys and re-encryption keys for each entity participating in the system.

[0060] Specifically, given the highest power n of the polynomial function, the trusted center generates two random (n+1)*(n+1) reversible matrices {M1, M2} as master keys, where M1 represents master key 1 and M2 represents master key 2.

[0061] For each registered perceptual user u k , the trusted center selects two random (n+1)*(n+1) reversible matrices {A′ k,1 ,A′ k,2}, generate an encryption key for it and Among them, A′ k,1 Represents the perceived user u k The encryption key process random matrix 1 is used as u k The re-encryption key 1, A′ k,2 Represents the perceived user u k The encryption key process random matrix 2 is used as u k The re-encryption key 2, Denotes a random matrix A′ k,1 The inverse matrix of Denotes a random matrix A′ k,2 The inverse matrix of k represents the perceived user u k Number, A k,1 Represents the perceived user u kEncryption Key 1, A k,2 Represents the perceived user u k Encryption key 2.

[0062] For each registered data requester u j , the trusted center selects two random (n+1)*(n+1) reversible matrices {B′ j,1 ,B′ j,2}, generate an encryption key for it and Among them, B′ j,1 Indicates the data requester u j The encryption key process random matrix 1 is used as u j The re-encryption key 1, B′ j,2 Indicates the data requester u j The encryption key process random matrix 2 is used as u j The re-encryption key 2, Denotes the random matrix B′ j,1 The inverse matrix of Denotes the random matrix B′ j,2 The inverse matrix of represents the inverse matrix of the master key M1, represents the inverse matrix of the master key M2, j represents the data requester u j Number, B j,1 Indicates the data requester u j The encryption key 1, B j,2 Indicates the data requester u j The encryption key 2.

[0063] Afterwards, the trusted center distributes the corresponding encryption key to each sensing user / data requester through a secure communication channel. k The re-encryption key (A′ k,1 ,A′ k,2 ) and data requester u j The re-encryption key (B′ j,1 ,B′ j,2 ) is sent to the perception center.

[0064] At the same time, the trusted center generates a task dictionary Task, which maps different tasks to different random numbers. Preferably, a secure hash function can be used to map tasks to random numbers, and the one-way and anti-collision properties of the secure hash function can be used to protect the privacy of the task. After generating the task dictionary, the trusted center sends the task dictionary to the perception user and the data requester through a secure communication channel.

[0065] After the system initialization is completed, the trusted center is in an offline or dormant state.

[0066] Step 2: Task Encryption. The sensing user determines the tasks it is interested in, encrypts them, and sends them to the sensing platform.

[0067] Specifically, let the sensing user be u k The set of tasks it is interested in is The sensing user u k Finds its corresponding value from the task dictionary Task, that is: where m represents the task number, and T k,m represents the value corresponding to the m-th task in the task set of the sensing user u k in the task dictionary Task, and l represents the total number of tasks the sensing user u k is interested in.

[0068] To maintain consistent dimensions, when l < n, the sensing user u k additionally selects several random numbers not in the dictionary Subsequently, the sensing user u k selects a non-zero random number a n ≠0 and generates a polynomial function to hide the required person information:

[0069]

[0070] where n represents the total number of tasks the sensing user u k is interested in; x represents the polynomial function variable, and x n represents the n-th power of the variable x; T k,n represents the value corresponding to the n-th task in the task set of the sensing user u k in the task dictionary Task; (a0, a1, a2, …) represents the coefficients of the polynomial function and a n represents the non-zero random number selected by the sensing user u k .

[0071] Then, extract the coefficients (a0, a1, a2, …, a n ), select a random number r k ≠0, and generate a lower triangular random matrix τ k with the main diagonal elements of (r k a0, r k a1, r k a2, …, r n a k ):

[0072]

[0073] Among them, r k Represents the perceived user u k The non-zero random number, τ, chosen to generate the triangular random matrix k Represents the lower triangular random matrix generated by the perceptual user. For ease of description, * is used to represent the random numbers in the lower triangular random matrix.

[0074] Then, the perceived user u k Using an encryption key (A k,1 ,A k,2 ) Encryption τ k for:

[0075] E[τ k ]=A k,1 *τ k *A k,2 (3)

[0076] Among them, E[τ k ] represents the encrypted ciphertext matrix.

[0077] After completing the above operations, the user u is perceived k The ciphertext E[τ k ]Sent to the perception platform.

[0078] Step 3: Task re-encryption: After receiving the ciphertext information uploaded by the perception user, the perception platform performs the task re-encryption operation.

[0079] Specifically, the perception platform uses the re-encryption key (A′ k,1 ,A′ k,2 )calculate:

[0080]

[0081] Among them, RE[τ k ] represents the perception platform's response to the received perception user u k The ciphertext E[τ k ]The ciphertext after re-encryption.

[0082] Step 4: Trapdoor generation: The data requester generates a trapdoor for the task information it needs.

[0083] Specifically, the data requester u j First, determine the keyword T of the required task from the task dictionary Task j,m ←Task(m). Where T j,m Indicates the data requester u j The value corresponding to the mth task in the task set in the task dictionary Task.

[0084] For each task Tj,m , data requester u j Choose a random number r j,m ≠0 generates a (n+1)*(n+1) main diagonal element as The lower triangular random matrix τ j,m .in, Indicates the data requester u j The nth power of the value corresponding to the mth task in the task dictionary Task, r j,m Indicates the data requester u j The nonzero random numbers chosen to generate the lower triangular random matrix.

[0085] Then, the data requester u j Using encryption key (B j,1 ,B j,2 ) Generate a trapdoor:

[0086] T[τ j,m ]=B j,1 *τ j,m *B j,2 (5)

[0087] Afterwards, the trapdoor T[τ j,m ] is sent to the perception platform. Among them, τ j,m Indicates the data requester u j The generated lower triangular random matrix, T[τ j,m ] indicates the data requester u j Generated trapdoor.

[0088] Step 5: Trapdoor re-encryption. After receiving the trapdoor, the perception platform uses the re-encryption key to re-encrypt the trapdoor.

[0089] Specifically, the perception platform uses the re-encryption key (B′ j,1 ,B′ j,2 )calculate:

[0090]

[0091]

[0092] Among them, RT[τ j,m ] indicates that the perception platform receives the book request from u j The ciphertext is re-encrypted using the trapdoor.

[0093] Step 6: Task retrieval. The perception platform uses the re-encrypted trapdoor to perform task retrieval operations on the re-encrypted task information stored in its database.

[0094] Specifically, for each task ciphertext RE[τk ], Perception Platform Computing:

[0095]

[0096] Here, tr(·) represents the matrix trace operation (i.e., the sum of the main diagonal elements of the matrix). Represents the perceived user u k Generated polynomial function, Δ k,(j,m) It represents the result calculated by the perception platform based on the task ciphertext matrix of the perception user and the ciphertext matrix of the data requester.

[0097] Due to r k 、r j,m is a random real number not equal to 0, if Δ k,(j,m) =0, then That is, the perceived user u k Tasks for data requesters T j,m Otherwise, the perceived user u k Data requester u is not satisfied j mission requirements.

[0098] Finally, the perception platform connects data requesters and perception users.

Claims

1. A privacy-preserving task allocation method based on task content, characterized in that: By using polynomial functions and random matrix multiplication, privacy-preserving task allocation based on task content is realized. In this method, a polynomial function is used to represent user task information, a random matrix is ​​used to protect the task information of data requesters, and task retrieval operations are performed on the random matrix based on matrix properties. The following steps are involved: Step 1: The trusted center generates encryption keys and re-encryption keys for each entity participating in the system; then, the trusted center distributes the corresponding encryption keys to each perception user and data requester through a secure communication channel; the trusted center sends the perception user's re-encryption key and the data requester's re-encryption key to the perception center through a secure communication channel; At the same time, the trusted center generates a task dictionary, which maps different tasks to different random numbers. After generating the task dictionary, the trusted center sends the task dictionary to the perception user and the data requester through a secure communication channel. The method for generating encryption keys and re-encryption keys is as follows: Given the highest power n of the polynomial function, the trusted center generates two random (n+1)*(n+1) reversible matrices {M1,M2} as master keys, where M1 represents master key 1 and M2 represents master key 2; For each registered perceptual user u k , the trusted center selects two random (n+1)*(n+1) reversible matrices {A′ k,1 ,A′ k,2 }, generate an encryption key for it and Among them, A′ k,1 Represents the perceived user u k The encryption key process random matrix 1 is used as u k The re-encryption key 1, A′ k,2 Represents the perceived user u k The encryption key process random matrix 2 is used as u k The re-encryption key 2, Denotes a random matrix A′ k,1 The inverse matrix of Denotes a random matrix A′ k,2 The inverse matrix of k represents the perceived user u k Number, A k,1 Represents the perceived user u k Encryption Key 1, A k,2 Represents the perceived user u k Encryption key 2; For each registered data requester u j , the trusted center selects two random (n+1)*(n+1) reversible matrices {B′ j,1 ,B′ j,2 }, generate an encryption key for it and Among them, B′ j,1 Indicates the data requester u j The encryption key process random matrix 1 is used as u j The re-encryption key 1, B′ j,2 Indicates the data requester u j The encryption key process random matrix 2 is used as u j The re-encryption key 2, Denotes the random matrix B′ j,1 The inverse matrix of Denotes the random matrix B′ j,2 The inverse matrix of represents the inverse matrix of the master key M1, represents the inverse matrix of the master key M2, j represents the data requester u j Number, B j,1 Indicates the data requester u j The encryption key 1, B j,2 Indicates the data requester u j The encryption key 2; Step 2: The perception user determines the task of interest and sends it to the perception platform after encryption; Assume that the perceived user u k The set of tasks of interest is Perceived user u k Find the corresponding value from the task dictionary Task, that is: Where m represents the task number, T k,m Represents the perceived user u k The value corresponding to the mth task in the task set in the task dictionary Task, l represents the perceived user u k The total number of tasks of interest; When l < n, the sensing user u k Additionally select several random numbers not in the dictionary Subsequently, the sensing user u k Select a random number a n ≠ 0 and generate a polynomial function Hide the required character information: Where n represents the perceived user u k The total number of tasks of interest; x represents the polynomial function variable, x n represents the variable x raised to the power of n; T k , n represents the perceived user u k The value corresponding to the nth task in the task set in the task dictionary Task; (a0, a1, a2, ...) represents a polynomial function The coefficient of n Represents the perceived user u k A non-zero random number is chosen; Then, extract The coefficients (a0, a1, a2, …, a n ), select a random number r k ≠0, the main diagonal elements of (n+1)*(n+1) are (r k a0,r k a1,r k a2,…,r k a n ) of the lower triangular random matrix τ k : Among them, r k Represents the perceived user u k The non-zero random number, τ, chosen to generate the triangular random matrix k Represents the perceived user u k The generated lower triangular random matrix; use * to represent the random numbers in the lower triangular random matrix; Then, the perceived user u k Using an encryption key (A k ,1,A k,2 ) Encryption τ k for: E[t k ]=A k,1 *t k *A k,2 Among them, E[τ k ] represents the encrypted ciphertext matrix; After completing the above operations, the user u is perceived k The ciphertext E[τ k ]Sent to the perception platform; Step 3: After receiving the ciphertext information uploaded by the perception user, the perception platform performs the task re-encryption operation Step 4: The data requester generates a trapdoor for the task information he needs; Data requester j First, determine the keyword T of the required task from the task dictionary Task j,m ←Task(m), where T j,m Indicates the data requester u j The value corresponding to the mth task in the task set in the task dictionary Task; For each task T j,m , data requester u j Choose a random number r j,m ≠0 generates a (n+1)*(n+1) main diagonal element as The lower triangular random matrix τ j,m ,in, Indicates the data requester u j The nth power of the value corresponding to the mth task in the task dictionary Task, r j,m Indicates the data requester u j The non-zero random numbers chosen to generate the lower triangular random matrix; Then, the data requester u j Using encryption key B j,1 , B j,2 Generate a trapdoor: T[t j,m ]=B j,1 *t j,m *B j,2 Afterwards, the trapdoor T[τ j,m ] is sent to the perception platform, where τ j,m Indicates the data requester u j The generated lower triangular random matrix, T[τ j,m ] indicates the data requester u j Generated trapdoors; Step 5: After receiving the trapdoor, the perception platform uses the re-encryption key to re-encrypt the trapdoor; Step 6: The perception platform uses the re-encrypted trapdoor to perform task retrieval operations on the re-encrypted task information stored in its database; finally, the perception platform establishes a connection between the data requester and the perception user.

2. A privacy protection task allocation method based on task content as claimed in claim 1, characterized in that: The task dictionary Task uses a secure hash function to map tasks to random numbers, and uses the one-way and collision-resistant properties of the secure hash function to protect the privacy of the tasks.

3. A privacy protection task allocation method based on task content as claimed in claim 1, characterized in that: Step 3 is implemented as follows: Given the encrypted task information E[τ k ], the perception platform uses the re-encryption key (A′ k,1 ,A′ k,2 ) re-encrypt the encrypted task to: RE[τ k ]0A′ k,1 *E[τ k ]*AM' k,2 Among them, RE[τ k ] represents the perception platform's response to the received perception user u k The ciphertext E[τ k ]The ciphertext after re-encryption; A′ k,1 Represents the perceived user u k The encryption key process random matrix 1 is used as u k The re-encryption key 1, A′ k,2 Represents the perceived user u k The encryption key process random matrix 2 is used as u k The re-encryption key 2.

4. A privacy protection task allocation method based on task content as claimed in claim 1, characterized in that: Step 5 is implemented as follows: After receiving the trapdoor, the perception platform uses the re-encryption key B′ j,1 , B′ j,2 Re-encrypt the trapdoor to: RT[t j,m ]=B′ j,1 *T[t j,m ]*B′ j,2 Among them, RT[τ j,m ] indicates that the perception platform receives the book request from u j The ciphertext after re-encryption with the trapdoor; B′ j,1 Indicates the data requester u j The encryption key process random matrix 1 is used as u j The re-encryption key 1, B′ j,2 Indicates the data requester u j The encryption key process random matrix 2 is used as u j The re-encryption key 2.

5. A privacy protection task allocation method based on task content as claimed in claim 1, characterized in that: Step 6 is implemented as follows: For each task ciphertext RE[τ k ], Perception Platform Computing: Among them, tr(·) represents the matrix trace operation, Represents the perceived user u k Generated polynomial function, Δ k,(j,m) represents the result calculated by the perception platform based on the task ciphertext matrix of the perception user and the ciphertext matrix of the data requester; τ k Represents the perceived user u k Generated lower triangular random matrix; Due to r k 、r j,m is a random real number not equal to 0, if Δ k,(j,m) =0, then That is, the perceived user u k Tasks for data requesters T j,m Interested, otherwise, perceive user u k Does not satisfy data requester u j mission requirements.

Images