Multi-label POI information encryption query method and system for privacy protection

By using homomorphic encryption and dual-cloud collaboration technology, an encrypted Voronoi graph and Vo-kd tree are constructed, which solves the privacy leakage problem in POI information query, realizes efficient and accurate multi-label POI data query, reduces dependence on trusted proxy servers, and improves user privacy protection and query efficiency.

CN118898085APending Publication Date: 2024-11-05XIDIAN UNIV
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202410930277.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-11
Publication Date
2024-11-05

AI Technical Summary

Technical Problem

Existing POI information query solutions fail to adequately protect user location privacy, posing a risk of information leakage, and rely on trusted proxy servers, making it difficult to achieve privacy-secure POI queries.

Method used

By employing homomorphic encryption and dual-cloud collaborative technology, a Voronoi diagram and Vo-kd tree are constructed by generating security parameters, public keys, and private keys to achieve encrypted processing of POI data. Collaborative queries are performed using a secure comparison protocol and a polygon range query protocol to avoid plaintext data transmission.

Benefits of technology

It effectively protects user location privacy, improves query efficiency and accuracy, reduces reliance on trusted proxy servers, and ensures data security and the accuracy of query results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118898085A_ABST
    Figure CN118898085A_ABST
Patent Text Reader

Abstract

The invention discloses a privacy-protected multi-label POI information encryption query method and system, relates to the technical field of information security, and solves the problem of information leakage caused by POI information query in the prior art. The method comprises the following steps: generating system security parameters, and generating security parameters required by a system and a homomorphic encryption algorithm by a POI data provider. System security parameters are distributed, and the POI data provider distributes the generated security parameters to the query user and the outsourcing cloud service provider; the query request is encrypted, a query user encrypts the position of the query user and a POI label needing to be queried by using a public key received from a POI data provider to form an encrypted query parameter, and carries the encrypted query parameter to send the query request to the CSP1 to obtain nearest neighbor target POI information; the encryption data is allowed to be calculated based on the homomorphic encryption method, plaintext circulation of user position information and POI data is avoided, and privacy leakage is prevented.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of information security technology, and in particular to a privacy-protected multi-label POI information encrypted query method and system. Background Art

[0002] As an important part of geographic information system, POI (point of interest) has been widely used in navigation, location services, commercial promotion and other fields. With the rapid development of mobile Internet and big data technology, the acquisition and application of POI data are becoming more and more convenient, which has brought great convenience to people's lives. Users can easily query various POI information such as nearby schools, subway entrances, hospitals, restaurants, etc. through mobile applications, online maps and other platforms, providing strong support for travel, shopping, entertainment and other activities. In the commercial field, POI data also plays an important role. Merchants can understand users' behavioral habits and consumption preferences by analyzing POI data, so as to formulate more accurate marketing strategies. At the same time, POI data can also be used in urban planning, traffic management and other fields, providing strong data support for government decision-making.

[0003] In location-based services and Geographic Information Systems (GIS), nearest neighbor search for POIs is an important task. It involves finding the points of interest that are closest to the user's location or query point within a given geographic range. In order to improve the efficiency and accuracy of POI nearest neighbor search, commonly used methods include Voronoi diagrams and KD trees. Voronoi diagrams are a mathematical structure for space segmentation that divides a plane into several regions, each corresponding to a specific POI. The distance from any point in each region of the Voronoi diagram to the corresponding POI is smaller than the distance to other POIs. Therefore, in POI nearest neighbor search, Voronoi diagrams can help quickly determine the region where the query point is located, so as to directly find the corresponding nearest POI. KD tree is a tree data structure for multidimensional spatial data, which is suitable for nearest neighbor search for high-dimensional data. For POI nearest neighbor search, KD tree can organize POI data by recursively dividing the space into hyperrectangles. The query process gradually narrows the search range by performing a series of region pruning operations in the tree, thereby efficiently finding the nearest POI.

[0004] There are many schemes for nearest neighbor search at present, among which information point query is to obtain category requirements and attribute requirements through demand identification, and then screen out various types of candidate POIs that match the location requirements in the map POI. And further determine the target POI of the query based on the category requirements, attribute requirements and the tag set corresponding to each candidate POI. Finally, a pan-POI query method that meets user needs is implemented in the map APP, enriching the functions of the map APP. However, the shortcomings of these schemes and inventions are that during the entire query process of the POI, the user's location data and POI data are circulated in plain text, the user's privacy information cannot be effectively protected, and there is a serious risk of privacy leakage.

[0005] In order to solve the problem of privacy leakage, there are also some methods based on anonymous technology. The user's current location information and query information are sent to the central anonymous server through the mobile device. The central anonymous server anonymously encrypts the current location information and query parameters based on the virtual trajectory generation algorithm, and sends the anonymous encrypted data to the LBS server for service query. The LBS server sends the query results to the central anonymous server. The central anonymous server filters out the query results required by the user from the query results returned by the LBS server, and returns the query results required by the user to the user's mobile device through encrypted communication. The shortcoming of this invention is that in such an anonymization scheme, a trusted proxy server is required to generate anonymous information for the user. However, in a practical environment, it is difficult to deploy a fully trusted proxy server. Therefore, it is difficult to meet the privacy and security POI query requirements in reality.

[0006] In summary, the problems of the existing technologies are as follows: (1) Some current POI information query solutions fail to fully protect the user's location privacy, resulting in the risk of leakage of user location information, which poses a potential threat to the user's personal privacy. (2) Although there are some privacy-preserving query solutions, these solutions often rely on a trusted proxy server to achieve anonymization. However, in real environments, it is extremely difficult to build a fully trusted proxy server, so it is difficult to achieve truly privacy-secure POI query. Summary of the invention

[0007] The present invention solves the problem of information leakage caused by POI information query in the prior art by providing a privacy-preserving multi-label POI information encryption query method and system, and realizes the calculation of encrypted data based on the homomorphic encryption method, thereby avoiding the plaintext flow of user location information and POI data, and effectively preventing privacy leakage.

[0008] In a first aspect, the present invention provides a privacy-preserving multi-label POI information encrypted query method, the method comprising:

[0009] The POI data provider generates a security parameter PP according to the data privacy protection requirements, and generates a public key PK and a private key SK required by the homomorphic encryption algorithm according to the security parameter;

[0010] The POI data provider distributes the security parameter PP and the public key PK to the querying user, sends the security parameter PP and the public key PK to CSP1 in the outsourced cloud service provider, and sends the security parameter PP and the private key SK to CSP2 in the outsourced cloud service provider;

[0011] The POI data provider generates Voronoi diagrams with different labels according to the labels of the POI data in the POI point set, and obtains a Voronoi diagram set VD(P); wherein the Voronoi diagram set VD(P) includes a plurality of Voronoi diagrams, and the Voronoi diagram includes a plurality of polygons;

[0012] The POI data provider calculates a corresponding KD tree of one of the POI data, and generates a Vo-kd tree based on the Vo-kd tree generation algorithm according to the Voronoi diagram set VD(P) and the KD tree;

[0013] The POI data provider encrypts each Voronoi diagram and the Vo-kd tree in the Voronoi diagram set VD(P) by homomorphic encryption, and sends the encrypted Voronoi diagram and the encrypted Vo-kd tree to CSP1 in the outsourced cloud service provider;

[0014] The query user generates an encrypted query request EQ(q, t) according to the public key PK, the coordinates of the current location point and the query tag, and sends the encrypted query request EQ(q, t) to the outsourced cloud service provider; wherein the query tag is one of the tags of the POI data;

[0015] CSP1 and CSP2 in the outsourced cloud service provider perform collaborative query according to the encrypted query request EQ(q,t) and utilize a secure comparison protocol and a secure polygon range query protocol to obtain a collaborative query result;

[0016] The query user obtains the collaborative query result, and decrypts the collaborative query result to obtain the nearest neighbor target POI information.

[0017] In combination with the first aspect, in a possible implementation, the POI data provider generates a security parameter PP according to data privacy protection requirements, and generates a public key PK and a private key SK required by a homomorphic encryption algorithm according to the security parameter, including:

[0018] Determine the homomorphic encryption security factor κ;

[0019] According to the homomorphic encryption security factor κ and the key generation function, a private key SK and a security parameter PP are obtained;

[0020] Use the private key SK to encrypt 0 and -1 respectively to obtain the public key PK.

[0021] In combination with the first aspect, in a possible implementation manner, the POI data provider generates Voronoi diagrams with different labels according to the labels of the POI data to obtain a Voronoi diagram set VD(P), including:

[0022] Determine the two-dimensional space corresponding to each type of label According to the two-dimensional space And the POI point set with the same label The initial Voronoi diagram corresponding to the same label generated;

[0023] Each POI point in each of the initial Voronoi diagrams is connected to the other adjacent POI points, and a perpendicular bisector is made for the line between two POI points. The initial Voronoi diagram is divided by a plurality of perpendicular bisectors to obtain a Voronoi diagram set VD(P).

[0024] In combination with the first aspect, in a possible implementation manner, the POI data provider calculates the KD tree corresponding to the i-th label in the POI data, and generates a Vo-kd tree based on the Vo-kd tree generation algorithm according to the Voronoi diagram set VD(P) and the KD tree, including:

[0025] Determine the POI point set with the same category label

[0026] Loop to determine the cutting dimension kd, and determine the cutting dimension kd in the POI point set The corresponding cutting point in the POI point set is used to Perform data space segmentation and separate the cutting points from the POI point set Remove until the POI point set is empty, and a KD tree is obtained; wherein the KD tree includes n+1 rectangular subspaces R={R1,…,R n+1};

[0027] Compare the areas where the Voronoi diagram set VD(P) and the KD tree intersect through the Vo-kd tree generation algorithm to obtain an area list

[0028] According to the area list Get the Vo-kd tree.

[0029] In combination with the first aspect, in a possible implementation manner, CSP1 and CSP2 in the outsourced cloud service provider perform collaborative queries on a security comparison protocol and a security polygon range query protocol according to the encrypted query request EQ(q, t) to obtain collaborative query results, including:

[0030] CSP1 in the outsourced cloud service provider receives the encrypted query request EQ(q,t), and filters and screens from the root node of the Vo-kd tree to obtain a list of candidate Voronoi diagram polygon areas for the kth label category POI point set.

[0031] The candidate Voronoi diagram polygon area list is compared by the first judgment condition in the safety comparison protocol. Make a judgment, if it meets the requirements, then get a list of alternative polygonal areas

[0032] The polygon area list is queried by the second judgment condition in the safe polygon range query protocol. The multiple polygonal areas in the query are screened to obtain the collaborative query result E(p i );

[0033] CSP1 in the outsourced cloud service provider selects a random number r1 and a random number r2, and according to the random number r1, the random number r2 and the collaborative query result E(p i ) calculates the first value E(p i ′), the first value E(p i ′) Sending to CSP2 in the outsourced cloud service provider, sending the random number r1 and the random number r2 to the querying user;

[0034] CSP2 in the outsourced cloud service provider receives the first value E(p i ′), and then use the private key SK to convert the first value E(p i ') decrypt to obtain second data, and send the second data to the querying user.

[0035] In combination with the first aspect, in a possible implementation manner, the query user obtains the collaborative query result, and decrypts the collaborative query result to obtain the nearest neighbor target POI information, including:

[0036] The collaborative query result is queried, and a disturbance recovery algorithm is executed to obtain the nearest neighbor target POI information.

[0037] In a second aspect, the present invention provides a privacy-protected multi-label POI information encrypted query system, the system comprising: a POI data provider, an outsourced cloud service provider, and a query user;

[0038] The POI data provider is used to collect POI data and process the POI data;

[0039] The outsourced cloud service provider is used to collaborate to implement the nearest neighbor POI query of the target tag under the encrypted data;

[0040] The query user is used to provide an encrypted user location and required query tags, and request a target tag nearest neighbor POI query service from an outsourced cloud service provider.

[0041] In conjunction with the second aspect, in a possible implementation, the POI data provider includes: a security parameter initialization module, a Voronoi diagram generation module, a KD tree generation module, a Vo-kd tree generation module, and a data encryption and transmission module;

[0042] The security parameter initialization module is used to select the security parameter PP according to the homomorphic encryption security requirements;

[0043] The Voronoi diagram generation module is used to generate Voronoi diagrams of POI data with different category labels;

[0044] The KD tree generation module is used to select POI data of a category label and create a KD tree according to the data set;

[0045] The Vo-kd tree generation module is used to further create a Vo-kd tree according to the Voronoi diagram generated by the Voronoi diagram generation module and the KD tree generated by the KD tree generation module;

[0046] The data encryption and sending module is used to encrypt the Voronoi diagram data and the Vo-kd tree data using a homomorphic encryption public key, and send the encrypted data to the outsourced cloud service provider.

[0047] In conjunction with the second aspect, in a possible implementation, the outsourced cloud service provider includes: a security parameter receiving module, an encrypted data receiving module, and a query collaboration module;

[0048] The security parameter receiving module includes: a CSP1 security parameter receiving module and a CSP2 security parameter receiving module, the CSP1 security parameter receiving module in the outsourced cloud service provider is used to receive the public key PK and the security parameter PP, and the CSP2 security parameter receiving module in the outsourced cloud service provider is used to receive the private key SK and the security parameter PP;

[0049] The encrypted data receiving module, the encrypted data receiving module of CSP1 in the outsourced cloud service provider, is used to receive the encrypted data provided by the POI data provider and store it for collaborative query;

[0050] The query collaboration module, the query collaboration module of CSP1 in the outsourced cloud service provider and the query collaboration module of CSP2 in the outsourced cloud service provider, is used to collaborate on secure queries under ciphertext according to homomorphic computing rules and secure computing protocols.

[0051] In conjunction with the second aspect, in a possible implementation, the query user includes: a user security parameter receiving module, a query request generating module, a query request encryption sending module and a query result receiving module;

[0052] The user security parameter receiving module is used to receive the public key PK and the security parameter PP;

[0053] The query request generation module is used to generate an encrypted query request EQ(q, t) according to the public key PK, the current location point coordinates and the query tag;

[0054] The query request encryption sending module is used to send the encrypted query request EQ(q, t) to the outsourced cloud service provider;

[0055] The query result receiving module is used to obtain the collaborative query result, and decrypt the collaborative query result to obtain the nearest neighbor target POI information.

[0056] One or more technical solutions provided in the present invention have at least the following technical effects or advantages:

[0057] (1) The present invention is based on homomorphic encryption and dual-cloud collaboration to achieve privacy protection of the multi-label POI database, query data, and query results during the query process of the nearest neighbor specific category POI. During the entire query process, the multi-label POI database is securely outsourced to a non-colluding outsourcing cloud service provider, and the query request data of the query user is securely encrypted, and the location privacy of the query user is effectively protected. At the same time, the query results are also scrambled, effectively ensuring the data privacy security in the entire query process;

[0058] (2) While ensuring the privacy and security of POI queries, the present invention effectively improves query efficiency by using Voronoi diagrams and KD trees. In addition, the present invention implements ciphertext processing of multi-label POI data, and can efficiently and accurately perform secure queries on nearest neighbor POI data for specific label categories. Through the combination of these technologies, the present invention significantly optimizes the performance and security of multi-label POI data queries. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A flowchart of the steps of the method for encrypted query of multi-label POI information with privacy protection provided by an embodiment of the present invention;

[0060] Figure 2 A Voronoi diagram provided by an embodiment of the present invention;

[0061] Figure 3 A 2-D tree diagram provided by an embodiment of the present invention;

[0062] Figure 4 The Vo-2d tree diagram provided by the embodiment of the present invention;

[0063] Figure 5 A schematic diagram of the structure of a privacy-protected multi-tag POI information encrypted query system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0064] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] The present invention provides a privacy-preserving multi-label POI information encrypted query method, such as Figure 1 As shown, the method includes the following steps S101 to S107.

[0066] S101, the POI data provider generates security parameters PP according to data privacy protection requirements, and generates the public key PK and private key SK required by the homomorphic encryption algorithm based on the security parameters.

[0067] Specifically, the POI data provider generates a security parameter PP according to the data privacy protection requirements, and generates a public key PK and a private key SK required for the homomorphic encryption algorithm according to the security parameter, including the following steps S1011 to S1013.

[0068] S1011, determine the homomorphic encryption security factor κ.

[0069] S1012, according to the homomorphic encryption security factor κ and the key generation function, obtain the private key SK and the security parameter PP.

[0070] S1013, use the private key SK to encrypt 0 and -1 respectively to obtain the public key PK.

[0071] Exemplarily, the POI data provider uses the SHE homomorphic encryption scheme, where the SHE homomorphic encryption algorithm supports a limited number of homomorphic additions and homomorphic multiplications and includes the following functions and homomorphic computing properties:

[0072] Key generation function, KeyGen(k0,k1,k2,k p ,k q )→(PP,SK). For a given security parameter (k0,k1,k2,k p ,k q ) needs to satisfy k0 <k1≤k2<k p ≤k q The key generation function selects a random number of k2 bits in length. and k p bit length prime number p, and then choose a series of k p Prime number of bit length And calculate The final key generation algorithm outputs public parameters and private key

[0073] Encryption function, Enc(SK,m)→E(m). For a given message and key SK, the encryption function is Where r, r′ are both random numbers and satisfy

[0074] Decryption function, Dec(SK,E(m))→m. For a given ciphertext E(m) and key SK, the decryption process is if It can be seen that m≥0,m=m′. Otherwise

[0075] Ciphertext homomorphism property, for two given ciphertexts E(m1) and E(m2), E(m1)·E(m2)→E(m1+m2), E(m1)·E(m2)→E(m1·m2), and the maximum multiplication depth for the multiplication is

[0076] Scalar homomorphism property, for a given ciphertext E(m1) and plaintext m2, E(m1)+m2→E(m1+m2), E(m1)·m2→E(m1·m2), and for the multiplication, m2>0 is satisfied, and the maximum multiplication depth is

[0077] In the initialization phase, the POI data provider determines a series of SHE homomorphic encryption security parameters κ = (k0, k1, k2, k p ,kq ). The larger the security factor κ is, the higher the security of the homomorphic encryption system is, and the greater the encryption overhead of the homomorphic encryption system is.

[0078] Based on the SHE homomorphic encryption security parameter κ, the POI data provider uses the key generation function KeyGen(κ) to obtain the public parameter PP and private key SK of the SHE homomorphic encryption system. The private key is used to encrypt 0 and -1 to obtain the public key PK = {E(0)1, E(0)2, E(-1)}. Among them, E(0)1 and E(0)2 in the public key PK are two different encrypted values ​​of the value 0, and E(-1) is the encrypted value of the value -1.

[0079] S102, the POI data provider distributes the security parameter PP and the public key PK to the querying user, sends the security parameter PP and the public key PK to CSP1 in the outsourced cloud service provider, and sends the security parameter PP and the private key SK to CSP2 in the outsourced cloud service provider.

[0080] Specifically, the POI data provider sends the public key PK and public parameter PP to CSP1, and sends the private key SK and public parameter PP to CSP2 for secure query collaboration between two different outsourced cloud service providers. All query users obtain the public key PK and public parameter PP from the POI data provider for encrypted query requests.

[0081] S103, the POI data provider generates Voronoi diagrams with different labels according to the labels of the POI data in the POI point set, and obtains a Voronoi diagram set VD(P); wherein the Voronoi diagram set VD(P) includes a plurality of Voronoi diagrams, and the Voronoi diagram includes a plurality of polygons.

[0082] Specifically, in step S103, the POI data provider generates Voronoi diagrams with different labels according to the labels of the POI data to obtain a Voronoi diagram set VD(P), which includes the following steps S1031 to S1032.

[0083] S1031, determine the two-dimensional space corresponding to each type of label According to the two-dimensional space And the POI point set with the same label The initial Voronoi diagram corresponding to the same label is generated.

[0084] S1032, connect each POI point in each initial Voronoi diagram with the other adjacent POI points, and make a perpendicular bisector for the line between two POI points. Multiple perpendicular bisectors divide the initial Voronoi diagram to obtain a Voronoi diagram set VD(P).

[0085] For example, Figure 2 As shown, given a two-dimensional space And the POI point set with the same label The generated Voronoi diagram will Divide into n regions. In order to generate the Voronoi diagram Voronoi diagram generation algorithm For each point First, point p j Connect the points next to it, and then make perpendicular bisectors of all the connecting lines to get n polygons. Perform the above operations on all POI categories to obtain the Voronoi diagram of different categories of POI Where m is the number of POI tags.

[0086] S104, the POI data provider calculates the KD tree corresponding to a label in the POI data, and generates a Vo-kd tree based on the Voronoi diagram set VD(P) and the KD tree and the Vo-kd tree generation algorithm.

[0087] Specifically, in step S104, the POI data provider calculates the KD tree corresponding to the i-th label in the POI data, and generates a Vo-kd tree based on the Voronoi diagram set VD(P) and the KD tree and the Vo-kd tree generation algorithm, including the following steps S1041 to S1044.

[0088] S1041, determine the POI point set with the same category label

[0089] S1042, loop to determine the cutting dimension kd, and determine the cutting dimension kd in the POI point set The corresponding cutting points in the POI point set are Perform data space segmentation and remove cutting points from the POI point set Remove until the POI point set is empty, and a KD tree is obtained; wherein the KD tree includes n+1 rectangular subspaces R={R1,…,R n+1};

[0090] S1043, compare the intersection area of ​​the Voronoi diagram set VD(P) and the KD tree through the Vo-kd tree generation algorithm to obtain a region list

[0091] S1044, according to the regional list Get the Vo-kd tree.

[0092] Exemplarily, the POI data provider selects a set of POI points with a certain type of label, such as Figure 3 As shown, a KD tree is constructed with the selected POI point set. The present invention divides KD tree nodes into three categories: complete internal nodes, incomplete internal nodes, and leaf nodes. Among them, a complete internal node has exactly two child nodes KN l and KN r , they constitute a left subtree and a right subtree respectively. An incomplete internal node has exactly one child node, and a leaf node does not contain any child nodes. And all nodes also contain a cutting dimension attribute kd and a cutting value attribute kv. The construction of the entire KD tree is a recursive process. The KD tree generation algorithm The process is as follows:

[0093] First, select a POI point set with a category label Used to construct a KD tree.

[0094] Select a cutting dimension kd from all dimensions, and then use the median of the kd dimension of all data records to split the multidimensional data space into two subspaces. Taking the generation of a KD tree in a two-dimensional space as an example, first select the x-axis dimension (or y-axis dimension) for splitting, and then Sort by the values ​​on the x-axis (or y-axis) and take the point p on the median mid As the segmentation basis point, point p mid The value on the x-axis (or y-axis) is the node's cut value attribute kv, and the selected point p mid Eliminate point set

[0095] For each subspace cut out, repeat the steps in b until the point set is empty, then the The KD tree partitioned by the benchmark contains n+1 rectangular subspaces R = {R1,…,R n+1}. Taking the implementation of two-dimensional space partitioning as an example, a 2-D tree is obtained.

[0096] In S103, The Voronoi diagram of POIs with different labels is obtained by division, which is specifically expressed as follows:

[0097]

[0098] in, Indicates the Nth POI of the i-th category label i The divided polygonal area represented by the POI points.

[0099] In S104, The partition results in a 2-d tree, which contains n+1 rectangular subspaces R = {R1,…,R n+1}, where n is the number of POI points contained in the selected POI category, and for each R j In order to further generate a Vo-2d tree based on the Voronoi diagram and the 2-d tree, for the Vo-kd tree generation algorithm GenVorKD(VD(P), KD(P)) of this embodiment, for each R i Find the Voronoi diagram formed by POIs with different category labels and the list of areas intersecting with them as follows:

[0100]

[0101] Among them, i represents the i-th rectangular subspace, j represents the j-th POI label, V(p k ) represents the Voronoi diagram polygon area represented by the k-th POI point in the j-th category.

[0102] The list of areas obtained by judging the overlap between the polygonal area of ​​the Voronoi diagram and each rectangular subspace divided by the 2-D tree is recorded as an information node And use this information node as the left subtree information node or right subtree information node of the incomplete internal node or leaf node in the 2-D tree, such as Figure 4 As shown, the new binary tree formed is the Vo-2d tree.

[0103] Therefore, in the Vo-2d tree, when any rectangular subregion R is found i When the j-th category label POI point set is formed, the Voronoi diagram with R i List of overlapping polygonal areas

[0104] S105, the POI data provider encrypts each Voronoi diagram and Vo-kd tree in the Voronoi diagram set VD(P) through homomorphic encryption, and sends the encrypted Voronoi diagram and the encrypted Vo-kd tree to CSP1 in the outsourced cloud service provider.

[0105] Exemplarily, the POI data provider encrypts all Voronoi diagrams and generated Vo-2d trees through the SHE homomorphic encryption method and sends them to CSP1 for storage and provision of secure outsourced query services.

[0106] S106, the query user generates an encrypted query request EQ(q, t) according to the public key PK, the coordinates of the current location point and the query tag, and sends the encrypted query request EQ(q, t) to the outsourced cloud service provider; wherein the query tag is one of the tags in the POI data.

[0107] Exemplarily, the querying user uses the obtained encryption public key to generate an encrypted query request according to the current location point and the required POI service category. The POI service category generates an encrypted query request specifically as follows:

[0108] EQ(q,t)=Enc(Query({q=(x,y),t∈{1,…m}}));

[0109] Where (x, y) is the location of the query user, and t is the nearest POI service category that the user needs to query.

[0110] The query user sends the encrypted query request EQ(q, t) to the outsourced cloud service provider CSP = {CSP1, CSP2} for secure service query.

[0111] S107, CSP1 and CSP2 in the outsourced cloud service provider perform collaborative query according to the encrypted query request EQ(q, t) and utilize the secure comparison protocol and the secure polygon range query protocol to obtain a collaborative query result.

[0112] Specifically, in step S107, CSP1 and CSP2 in the outsourced cloud service provider perform collaborative queries on the security comparison protocol and the security polygon range query protocol according to the encrypted query request EQ(q, t) to obtain collaborative query results, including the following steps S1071 to S1075.

[0113] S1071, CSP1 in the outsourced cloud service provider receives the encrypted query request EQ(q,t), and filters and selects from the root node of the Vo-kd tree to obtain a list of candidate Voronoi diagram polygon areas for the kth label category POI point set.

[0114] S1072, comparing the candidate Voronoi diagram polygon area list according to the first judgment condition in the security comparison protocol Make a judgment, if it meets the requirements, then get a list of alternative polygonal areas

[0115] S1073, the polygon area list is queried using the second judgment condition in the secure polygon range query protocol. The multiple polygonal areas in the query are screened to obtain the collaborative query result E(p i );

[0116] S1074, CSP1 in the outsourced cloud service provider selects random numbers r1 and r2, and then performs a query based on the random numbers r1, r2 and the collaborative query result E(p i ) calculates the first value E(p i ′), the first value E(p i ′) Send it to CSP2 in the outsourced cloud service provider, and send the random number r1 and the random number r2 to the querying user;

[0117] S1075, CSP2 in the outsourced cloud service provider receives the first value E(p i ′), the first value E(p i ′) decrypt to obtain second data, and send the second data to the querying user.

[0118] For example, CSP1 in the outsourced cloud service receives the encrypted query request EQ(q, t) from the query user, and obtains the encrypted Vo-2d tree root node E(Node root )={kd root ,E(kv root )} to filter. The filtering algorithm Filter() is as follows. First, assume is the value of point q in the kd dimension of the i-th node, and is obtained by the secure comparison protocol π in the outsourced cloud service provider. SCP Compare and kv i The value of . The query request continues to filter on the left subtree. The query request continues to filter on the right subtree until the final information node is found Where R j is the point q=(x q ,y q ) in the rectangular area, The Voronoi diagram and rectangle R formed by the k-th category POI j List of intersecting polygonal areas.

[0119] Outsourcing cloud service providers through security comparison protocol π SCP Collaboratively compare the sizes of E(t), E(t-1) and E(k). If π is satisfied SCP (E(t),E(k))=True and π SCP (E(t-1),E(k))=False, then select As a list of candidate polygon areas. Then run the safe polygon range query protocol π SPRQP ,Will All polygonal areas E(V(pi )) as the screening object, for all i = 1, …, N k , if there is π SPRQP (E(V(p i )), E(q)) = True, then E(p i ) is the target POI that the query user needs to query.

[0120] To ensure data security, the perturbation algorithm is executed. CSP1 selects two random numbers r1, r2, and generates them through homomorphic calculation and sends E(p i ′) to CSP2, and sends (r1, r2) to the query user.

[0121] After receiving E(p i ′), CSP2 decrypts it using the private key SK to obtain and sends to the query user.

[0122] Among them, the Security Comparison Protocol (π SCP ) is used to compare the sizes of a and b under ciphertext. CSP1 holds two values E(a) and E(b) under ciphertext, and CSP2 holds the private key SK of homomorphic encryption.

[0123] CSP1 obtains E(−b) through homomorphic calculation, and selects two random numbers r1, r2 such that r1 > r2. Calculate E(c) = r1·(E(a) + E(−b)) + r2 through homomorphic calculation again, and send E(c) to CSP2.

[0124] After receiving E(c), CSP2 decrypts it using SK to obtain c = r1·(a - b) + r2. If c >= 0, then a >= b, otherwise a < b. CSP2 returns the comparison result res ∈ {True, False} to CSP1, and a security comparison protocol is completed. During the comparison process, neither CSP1 nor CSP2 can obtain the original data values a, b, ensuring the privacy and security of the data, and at the same time being able to accurately give the comparison result under ciphertext.

[0125] The Security Polygon Range Query Protocol (π SPRQP ) is used to query whether a point is within the polygon range. CSP1 holds the Voronoi diagram polygon under ciphertext: The Voronoi diagram polygon is represented as: E(V(p)) = {E(c1), …, E(c l )} = {(E(x1), E(y1)), …, (E(x l ), E(yl ))} and a point E(q)=(E(x q ),E(y q )), where the list of vertices of the polygon is sorted in counterclockwise order, c i Represents the position coordinates of the i-th vertex of the polygon.

[0126] CSP1 selects 21 random numbers (r i ,…,r l ) and (s1,…,s l ), and it satisfies for any i=1,…l,r i >s i Always holds true, and i′=(i+1)modl. CSP1 performs homomorphic calculations under ciphertext to obtain:

[0127] E(A i )=E(x q )·E(y i )+E(y q )·E(x i ′)+E(x i )·E(y i′ )

[0128] E(B i )=E(x q )·E(y i′ )+E(y q )·E(x i )+E(x i′ )·E(y i )

[0129] E(D i )=r i ·(E(A i )+E(-1)·E(B i ))+s i

[0130] Then CSP1 converts E(D)={E(D1),…,E(D l )} is sent to CSP2.

[0131] After receiving E(D), CSP2 uses the key SK to decrypt and obtain D = {D1, ..., D l}, if for any i=1,…l, there is D i >=0, then point q is within the range of polygon V(p), otherwise point q is not within the range of polygon V(p). CSP2 returns the range comparison result res∈{True,False} to CSP1, and a safe polygon range query protocol is completed.

[0132] S108, the query user obtains the collaborative query result, and decrypts the collaborative query result to obtain the nearest neighbor target POI information.

[0133] Specifically, in step S108, the query user obtains the collaborative query result, and decrypts the collaborative query result to obtain the nearest neighbor target POI information, including: querying the collaborative query result, executing the disturbance recovery algorithm, and obtaining the nearest neighbor target POI information.

[0134] For example, the query user receives (r1, r2) and After that, execute the disturbance recovery algorithm and do subtraction to get That is, the querying user successfully obtains the nearest neighbor POI results that meet the requirements of his query request EQ(q,t).

[0135] The present invention strengthens privacy protection: by combining dual-cloud collaboration and homomorphic encryption technology, the security of POI data during storage, transmission and processing is ensured. Based on the homomorphic encryption method, calculations can be performed on encrypted data, thereby avoiding the plaintext flow of user location information and POI data, effectively preventing privacy leakage.

[0136] The present invention eliminates the reliance on a trusted proxy server: the present invention does not rely on a completely trusted proxy server to complete anonymization processing, but instead implements privacy-secure POI query through cloud collaboration and encryption technology. This reduces the difficulty and cost of deploying and maintaining a trusted proxy server in a practical environment.

[0137] The present invention improves query efficiency and accuracy: By using structures such as Voronoi polygons and kd trees, the present method achieves efficient and accurate POI information query while protecting user privacy. This provides users with faster and more accurate query results, improving user experience.

[0138] The present invention provides multi-label POI information query: the present invention can provide query for POI information with specific labels, so that the query result is more accurate, and can provide users with more accurate and diversified query services.

[0139] The privacy-protected multi-label POI information encrypted query method proposed in the present invention can effectively protect the data security of POI data during use. All original POI data will be sent to the outsourced cloud service provider after encryption, and the cloud service provider cannot steal the original data. Secondly, in order to realize the query of the nearest neighbor POI of the target label under encryption, the present invention constructs two security protocols. Through the collaboration of dual clouds, data comparison and the determination of the position relationship between points and polygons can be realized under ciphertext. In addition, the present invention further protects the privacy security of query requests and query results. On the one hand, the query request uploaded by the query user is encrypted, and on the other hand, the query result obtained by the cloud service provider will not be directly decrypted, but decrypted after scrambling, and the decrypted result after disturbance and scrambling will be returned to the query user. In the entire query process, the cloud service provider will not obtain any valid plaintext POI information, realizing the privacy security of the entire process of POI data outsourcing and query. And through the label comparison under ciphertext, the present invention can realize the encrypted query of multi-label POI information, can accurately query POIs of specific label categories, and can effectively provide more accurate and diverse query services for query users.

[0140] The privacy-protected multi-label POI information encrypted query method proposed in the present invention is highly efficient. Before data outsourcing, the present invention generates a Voronoi diagram for POI data with different labels in advance, selects a POI point set data with one label to construct a KD tree, and finally constructs a Vo-2d tree for query search. The tree structure can effectively improve the query efficiency, quickly lock the potential Voronoi diagram polygon area list, and then use the properties of the Voronoi diagram to quickly obtain the final query result by traversing the results after screening, which effectively improves the query efficiency of the present invention.

[0141] The present invention provides a privacy-protected multi-label POI information encrypted query system, such as Figure 5 As shown, the system includes: a POI data provider, an outsourced cloud service provider and a query user.

[0142] POI data provider, used to collect POI data and process POI data. POI data provider includes: security parameter initialization module, Voronoi diagram generation module, KD tree generation module, Vo-kd tree generation module and data encryption and transmission module;

[0143] Security parameter initialization module, used to select security parameters PP according to homomorphic encryption security requirements;

[0144] Voronoi diagram generation module, used to generate Voronoi diagrams of POI data with different category labels;

[0145] The KD tree generation module is used to select POI data of a category label and create a KD tree based on the data set;

[0146] A Vo-kd tree generation module, used for further creating a Vo-kd tree according to the Voronoi diagram generated by the Voronoi diagram generation module and the KD tree generated by the KD tree generation module;

[0147] The data encryption and sending module is used to encrypt the Voronoi diagram data and the Vo-kd tree data using the homomorphic encryption public key, and send the encrypted data to the outsourced cloud service provider.

[0148] An outsourced cloud service provider is used to collaboratively implement a target tag nearest neighbor POI query under encrypted data; the outsourced cloud service provider includes: a security parameter receiving module, an encrypted data receiving module and a query collaboration module;

[0149] A security parameter receiving module, the security parameter receiving module of CSP1 in the outsourced cloud service provider is used to receive the public key PK and the security parameter PP, and the security parameter receiving module of CSP2 in the outsourced cloud service provider is used to receive the private key SK and the security parameter PP;

[0150] An encrypted data receiving module, the encrypted data receiving module of CSP1 in the outsourced cloud service provider, is used to receive the encrypted data provided by the POI data provider and store it for collaborative query;

[0151] The query collaboration module, the query collaboration module of CSP1 in the outsourced cloud service provider and the query collaboration module of CSP2 in the outsourced cloud service provider, is used to collaborate on secure queries under ciphertext according to homomorphic computing rules and secure computing protocols.

[0152] The query user is used to provide the encrypted user location and the required query tags, and request the nearest neighbor POI query service of the target tag from the outsourced cloud service provider. The query user includes: a security parameter receiving module, a query request generating module, a query request encryption sending module and a query result receiving module;

[0153] A security parameter receiving module, used for receiving a public key PK and a security parameter PP;

[0154] A query request generation module is used to generate an encrypted query request EQ(q,t) according to the public key PK, the current location point coordinates and the query tag;

[0155] A query request encryption sending module, used for sending the encrypted query request EQ(q,t) to the outsourced cloud service provider;

[0156] The query result receiving module is used to obtain the collaborative query result and decrypt the collaborative query result to obtain the nearest neighbor target POI information.

[0157] The present invention can outsource multi-label POI data to a cloud service provider under the premise of privacy protection, and realize encrypted query of multi-label POI information with the help of a constructed security protocol, and uses the Voronoi diagram and KD tree to accelerate the query process, further improving the efficiency of the solution, making it suitable for use in real scenarios.

[0158] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some or all of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the present invention.

Claims

1. A privacy-preserving multi-label POI information encrypted query method, characterized in that: include: The POI data provider generates a security parameter PP according to the data privacy protection requirements, and generates a public key PK and a private key SK required by the homomorphic encryption algorithm according to the security parameter; The POI data provider distributes the security parameter PP and the public key PK to the querying user, sends the security parameter PP and the public key PK to CSP1 in the outsourced cloud service provider, and sends the security parameter PP and the private key SK to CSP2 in the outsourced cloud service provider; The POI data provider generates Voronoi diagrams with different labels according to the labels of the POI data in the POI point set, and obtains a Voronoi diagram set VD(P); wherein the Voronoi diagram set VD(P) includes a plurality of Voronoi diagrams, and the Voronoi diagram includes a plurality of polygons; The POI data provider calculates a KD tree corresponding to a label in the POI data, and generates a Vo-kd tree based on the Vo-kd tree generation algorithm according to the Voronoi diagram set VD(P) and the KD tree; The POI data provider encrypts each Voronoi diagram and the Vo-kd tree in the Voronoi diagram set VD(P) by homomorphic encryption, and sends the encrypted Voronoi diagram and the encrypted Vo-kd tree to CSP1 in the outsourced cloud service provider; The query user generates an encrypted query request EQ(q, t) according to the public key PK, the coordinates of the current location point and the query tag, and sends the encrypted query request EQ(q, t) to the outsourced cloud service provider; wherein the query tag is one of the tags of the POI data; CSP1 and CSP2 in the outsourced cloud service provider perform collaborative query according to the encrypted query request EQ(q,t) and utilize a secure comparison protocol and a secure polygon range query protocol to obtain a collaborative query result; The query user obtains the collaborative query result, and decrypts the collaborative query result to obtain the nearest neighbor target POI information.

2. The privacy-preserving multi-tag POI information encrypted query method according to claim 1 is characterized in that: The POI data provider generates a security parameter PP according to the data privacy protection requirements, and generates a public key PK and a private key SK required by the homomorphic encryption algorithm according to the security parameter, including: Determine the homomorphic encryption security factor κ; According to the homomorphic encryption security factor κ and the key generation function, a private key SK and a security parameter PP are obtained; Use the private key SK to encrypt 0 and -1 respectively to obtain the public key PK.

3. The privacy-preserving multi-tag POI information encrypted query method according to claim 1, characterized in that: The POI data provider generates Voronoi diagrams with different labels according to the labels of the POI data in the POI point set, and obtains a Voronoi diagram set VD(P), including: Determine the two-dimensional space corresponding to each type of label According to the two-dimensional space And the POI point set with the same label The initial Voronoi diagram corresponding to the same label generated; Each POI point in each of the initial Voronoi diagrams is connected to the other adjacent POI points, and a perpendicular bisector is made for the line between two POI points. The initial Voronoi diagram is divided by a plurality of perpendicular bisectors to obtain a Voronoi diagram set VD(P).

4. The privacy-preserving multi-tag POI information encrypted query method according to claim 1, characterized in that: The POI data provider calculates a KD tree corresponding to a label in the POI data, and generates a Vo-kd tree based on the Voronoi diagram set VD(P) and the KD tree and a Vo-kd tree generation algorithm, including: Determine the POI point set with the same category label Loop to determine the cutting dimension kd, and determine the cutting dimension kd in the POI point set The corresponding cutting point in the POI point set is used to Perform data space segmentation and separate the cutting points from the POI point set Remove until the POI point set is empty, and a KD tree is obtained; wherein the KD tree includes n+1 rectangular subspaces R={R1,…,R n+1 }; Compare the areas where the Voronoi diagram set VD(P) and the KD tree intersect through the Vo-kd tree generation algorithm to obtain an area list According to the area list Get the Vo-kd tree.

5. The privacy-preserving multi-tag POI information encrypted query method according to claim 4 is characterized in that: The CSP1 and CSP2 in the outsourced cloud service provider perform collaborative query on the security comparison protocol and the security polygon range query protocol according to the encrypted query request EQ(q, t) to obtain collaborative query results, including: CSP1 in the outsourced cloud service provider receives the encrypted query request EQ(q,t), and filters and screens from the root node of the Vo-kd tree to obtain a list of candidate Voronoi diagram polygon areas for the kth label category POI point set. The candidate Voronoi diagram polygon area list is compared by the first judgment condition in the safety comparison protocol. Make a judgment, if it meets the requirements, then get a list of alternative polygonal areas The polygon area list is queried by the second judgment condition in the safe polygon range query protocol. The multiple polygonal areas in the query are screened to obtain the collaborative query result E(p i ); CSP1 in the outsourced cloud service provider selects a random number r1 and a random number r2, and performs a query based on the random number r1, the random number r2 and the collaborative query result E(p i ) calculates the first value E(p i ′), the first value E(p i ′) Sending to CSP2 in the outsourced cloud service provider, sending the random number r1 and the random number r2 to the querying user; CSP2 in the outsourced cloud service provider receives the first value E(p i ′), and then use the private key SK to convert the first value E(p i ') decrypt to obtain second data, and send the second data to the querying user.

6. The privacy-preserving multi-tag POI information encrypted query method according to claim 1, characterized in that: The query user obtains the collaborative query result, and decrypts the collaborative query result to obtain the nearest neighbor target POI information, including: The collaborative query result is queried, and a disturbance recovery algorithm is executed to obtain the nearest neighbor target POI information.

7. A privacy-preserving multi-label POI information encrypted query system, characterized in that: include: POI data providers, outsourced cloud service providers and query users; The POI data provider is used to collect POI data and process the POI data; The outsourced cloud service provider is used to collaborate to implement the nearest neighbor POI query of the target tag under the encrypted data; The query user is used to provide an encrypted user location and required query tags, and request a target tag nearest neighbor POI query service from an outsourced cloud service provider.

8. The privacy-preserving multi-label POI information encrypted query system according to claim 7, characterized in that: The POI data provider includes: a security parameter initialization module, a Voronoi diagram generation module, a KD tree generation module, a Vo-kd tree generation module and a data encryption and transmission module; The security parameter initialization module is used to select the security parameter PP according to the homomorphic encryption security requirements; The Voronoi diagram generation module is used to generate Voronoi diagrams of POI data with different category labels; The KD tree generation module is used to select POI data of a category label and create a KD tree according to the data set; The Vo-kd tree generation module is used to further create a Vo-kd tree according to the Voronoi diagram generated by the Voronoi diagram generation module and the KD tree generated by the KD tree generation module; The data encryption and sending module is used to encrypt the Voronoi diagram data and the Vo-kd tree data using a homomorphic encryption public key, and send the encrypted data to the outsourced cloud service provider.

9. The privacy-preserving multi-label POI information encrypted query system according to claim 7, characterized in that: The outsourced cloud service provider includes: a security parameter receiving module, an encrypted data receiving module and a query collaboration module; The security parameter receiving module includes: a CSP1 security parameter receiving module and a CSP2 security parameter receiving module, the CSP1 security parameter receiving module in the outsourced cloud service provider is used to receive the public key PK and the security parameter PP, and the CSP2 security parameter receiving module in the outsourced cloud service provider is used to receive the private key SK and the security parameter PP; The encrypted data receiving module, the encrypted data receiving module of CSP1 in the outsourced cloud service provider, is used to receive the encrypted data provided by the POI data provider and store it for collaborative query; The query collaboration module, the query collaboration module of CSP1 in the outsourced cloud service provider and the query collaboration module of CSP2 in the outsourced cloud service provider, is used to collaborate on secure queries under ciphertext according to homomorphic computing rules and secure computing protocols.

10. The privacy-protected multi-tag POI information encrypted query system according to claim 7, characterized in that: The query user includes: a user security parameter receiving module, a query request generating module, a query request encryption sending module and a query result receiving module; The user security parameter receiving module is used to receive the public key PK and the security parameter PP; The query request generation module is used to generate an encrypted query request EQ(q, t) according to the public key PK, the current location point coordinates and the query tag; The query request encryption sending module is used to send the encrypted query request EQ(q, t) to the outsourced cloud service provider; The query result receiving module is used to obtain the collaborative query result, and decrypt the collaborative query result to obtain the nearest neighbor target POI information.

Citation Information

Cited By

  • Visualization method and device for coupling place name tag and TreeMap subdivision, and electronic equipment

    CN120144686A