CN109740376A

CN109740376A - Location privacy protection method, system, equipment and medium based on NN Query

Info

Publication number: CN109740376A
Application number: CN201811570419.5A
Authority: CN
Inventors: 秦阳; 王轩; 王晗; 蒋琳
Original assignee: Shenzhen Graduate School Harbin Institute of Technology
Current assignee: Shenzhen Graduate School Harbin Institute of Technology
Priority date: 2018-12-21
Filing date: 2018-12-21
Publication date: 2019-05-10
Anticipated expiration: 2038-12-21
Also published as: CN109740376B

Abstract

The present invention discloses a kind of location privacy protection method based on NN Query, system, equipment and medium.This method includes the following steps of LBS client executing: being based on target position and interest vertex type, inquiry is in advance based on the Z-order tree of quaternary tree and the creation of Z-order curve, obtains destination node and target subtree；The leaf node in target subtree is traversed, destination node truncation coordinate is obtained and coordinate is truncated in adjacent node；Based on private cipher key, computations are carried out using improved privacy homomorphic encryption algorithm, destination node encryption data and adjacent node encryption data is obtained and is sent to LBS server, receive the secret range data that LBS server returns；Based on private cipher key, secret range data is decrypted using improved privacy homomorphic decryption algorithm, obtain the distance between destination node and adjacent node respectively and is ranked up, neighbour's point of interest corresponding with neighbour's number is obtained.This method can fast implement quickly positioning and inquiry, and computing cost is low and secrecy effect is good.

Description

Location privacy protection method, system, equipment and medium based on NN Query

Technical field

The present invention relates to position enquiring technical field more particularly to a kind of location privacy protection sides based on NN Query Method, device, equipment and medium.

Background technique

The fast development of mobile communication and space orientation technique is promoted based on location-based service (location based Services, LBS) rise.LBS (Location Based Service, i.e. location based service), is moved by telecommunications The radio communication network (such as GSM net, CDMA net) of dynamic operator or external positioning method (such as GPS) obtain LBS client Location information (geographical coordinate or geodetic coordinates), in GIS-Geographic Information System (foreign language abbreviation: GIS, foreign language full name: Geographic Information System) platform support under, provide a kind of value-added service of respective service for user.The LBS client It is mountable but be not limited to various personal computers, laptop, smart phone, tablet computer and portable wearable device On equal mobile terminals.Since the location information of LBS client includes the privacy informations such as user identity and behavior pattern, in order to avoid The leakage of privacy information needs first to carry out secret protection to the location information of LBS client, so as to be submitted to the position of LBS server Confidence breath is the location information being hidden, and this LBS is referred to as the LBS based on encryption.

K neighbour (i.e. approximate k nearest neighbor, Approximate k Nearest Neighbor, hereinafter referred to as kNN) inquiry is position The important inquiry business for setting service refers to that LBS client is a recently according to k of its location information near LBS server inquiry Point of interest (Point of Interest, hereinafter referred to as POIs), in GIS-Geographic Information System, POIs can be retail shop, public transport It stands, dining room or gas station etc..In general, its location information is submitted to LBS server by LBS client, and LBS server passes through The location information and neighbouring the distance between POIs for comparing LBS client, find and return to immediate k POIs.In kNN In query process, need to the location information to LBS client largely calculated with possible POIs, if calculate error it is big when meeting The accuracy of k returned POIs is seriously affected, therefore, need to guarantee the efficiency and accuracy of kNN inquiry.Based on encryption In LBS, the location information that LBS client is submitted is encrypted, and all POIs prestored in LBS server are with certain spy Mode tissue is determined, so that kNN query process can not ensure its efficiency and accuracy rate.

Realize that kNN inquiry was specifically included such as the next stage in the LBS based on encryption: the first stage, LBS client is from LBS The corresponding serial number of its location information is obtained in the organizational form of server publishes.Second stage, LBS client is by acquired sequence It number is encrypted, obtains ciphering sequence number, and ciphering sequence number is sent to LBS server；LBS server is carried out based on ciphering sequence number It calculates, obtains the POIs information of encryption, and the POIs information of the encryption is sent to LBS client.Phase III, LBS client End is decrypted and sorts to the POIs information of encryption, obtains k POIs of arest neighbors.It is to be appreciated that in order to guarantee that kNN is looked into The accuracy rate of inquiry, the quantity that the POIs of LBS server need to be arranged is the multiple of k.

It is using space filling curve is based on that POIs all on map are linear by orientation tissue in LBS, with realization pair The storage of location information.Space filling curve be it is a kind of can not have to intersect in two-dimensional space pass through all region or Say the curve of the hypercube of a multidimensional, including but not limited to hibert curve and Moore curve.

Hibert curve is famous with the ability that it can partially retain the adjacent adjoining of initial data, has preferable cluster Attribute.Fig. 1 (a) shows that the hibert curve of first three sequence, the hibert curve of N sequence can pass through 2^N×2^NA cell Domain.One number can be converted the POIs to be inquired to using hibert curve, the method for Applied cryptography is facilitated to carry out Linear search is encrypted and directly carried out when calculating the position of neighbour, but is changing two-dimensional space dress into one-dimensional linear space Later, it can be inevitably lost a part of azimuth information, when so that the range of inquiry is bigger, inquired more inaccurate.

Moore curve is a kind of variant version of hibert curve, is combined with the hibert curve as four Get up, is the loop version of hibert curve, as shown in Fig. 1 (b) to make endpoint coincidence be formed by curve.It is every in this way A POI has POIs adjacent thereto in the both direction of Moore curve, and common hibert curve can be overcome to inquire When the more bigger more inaccurate disadvantage of k.LBS server first by all POIs with Moore curve mapping at the two dimension of a bracing cable Table, first is classified as the serial number of POI, and second is classified as corresponding POI information.Include the following steps: (1) in the LBS based on encryption When the service of LBS client request, the excursion matrix of encryption can be sent to LBS server.Wherein, the offset of excursion matrix The actual value of t only has LBS client to know, and excursion matrix can be added with the homomorphism of any support multiplication and additive homomorphism Decryption method encryption.(2) LBS server excursion matrix carries out matrix multiplication calculating with preset map datum in LBS server, Since excursion matrix is secrecy for LBS server, so the result after calculating is also secrecy for LBS server , therefore obtained the new tables of data of the secrecy after a transformation movement.(3) LBS client requirements LBS server is returned through inclined The information of a certain serial number of physical location after shifting.Since LBS server is not aware that the mobile offset t's of data table transform Actual value, therefore when the data of a certain column of LBS client request, LBS server does not know which this column data is really yet A position, only user, which are known, oneself makes the tables of data in LBS server offsets by how many, so that LBS client can be with Desired location information is obtained in the case where the position that do not stick one's chin out is to LBS server.This side based on Moore curve Formula can also return to correct service while although user location privacy flexibly can be protected in primary request, due to LBS server will move up and down POIs each time, be equivalent to and do primary encryption matrix operation in entire map datum, Keep its efficiency lower.

Summary of the invention

The embodiment of the present invention provides a kind of location privacy protection method based on NN Query, device, equipment and medium, with Expense is big when solving the problem of to be currently based on kNN inquiry in the LBS of encryption and can not ensure the efficiency and accuracy of inquiry.

A kind of location privacy protection method based on NN Query, including LBS client executing following steps:

NN Query request is generated, the NN Query request includes target position, interest vertex type, neighbour's number and private There is key；

Based on the target position and the interest vertex type, inquiry is in advance based on quaternary tree and the creation of Z-order curve Z-order tree corresponding with the interest vertex type, obtain corresponding destination node and target subtree；

The leaf node in the target subtree is traversed, the corresponding node truncation coordinate of the leaf node is obtained, it is described Node truncation coordinate includes destination node truncation coordinate and adjacent node truncation coordinate；

Based on the private cipher key, coordinate and institute are truncated to the destination node using improved privacy homomorphic encryption algorithm It states adjacent node truncation coordinate and carries out computations, obtain destination node encryption data and adjacent node encryption data；

The destination node encryption data and the adjacent node encryption data are sent to LBS server, and receive institute State the homomorphism of LBS server return carried out in ciphertext to the destination node encryption data and the adjacent node encryption data Addition and multiplication calculate secret range data obtained；

Based on the private cipher key, the secret range data is solved using improved privacy homomorphic decryption algorithm It is close, the distance between destination node and adjacent node are obtained respectively；

It is ranked up, is obtained corresponding with neighbour's number based on the distance between the destination node and adjacent node Neighbour's point of interest.

A kind of location privacy protection method based on NN Query, including LBS server execute following steps:

Destination node encryption data and adjacent node encryption data that client is sent are received, the destination node encrypts number According to be respectively with the adjacent node encryption data using improved privacy homomorphic encryption algorithm to destination node truncation coordinate and Adjacent node is truncated coordinate and carries out data acquired in computations；

Is carried out by homomorphism addition in ciphertext and is multiplied for the destination node encryption data and the adjacent node encryption data Method calculates, and obtains secret range data；

The secret range data is sent to LBS client, so that the LBS client is based on private cipher key, is used The secret range data is decrypted in improved privacy homomorphic decryption algorithm, obtain respectively destination node and adjacent node it Between distance, and be ranked up based on the distance between the destination node and adjacent node, acquisition and neighbour's number phase Corresponding neighbour's point of interest.

A kind of location privacy protection system based on NN Query, including LBS client and LBS server；

The LBS client executing following steps:

It is ranked up, is obtained corresponding with neighbour's number based on the distance between the destination node and adjacent node Neighbour's point of interest；

The LBS server executes following steps:

A kind of computer equipment, including memory, processor and storage are in the memory and can be in the processing The computer program run on device, the processor realize the above-mentioned position based on NN Query when executing the computer program The step of method for secret protection.

A kind of computer readable storage medium, the computer-readable recording medium storage have computer program, the meter The step of calculation machine program realizes the above-mentioned location privacy protection method based on NN Query when being executed by processor.

The above-mentioned location privacy protection method based on NN Query, device, equipment and medium are based on quaternary tree and Z- In the Z-order tree of order curve creation, so that the tree of Z-order tree supports distributed storage, has easy segmentation And the characteristics of invariance.Although Z-order curve is than the poor continuity of hibert curve, it is more advised than hibert curve Then and positioning is easy, and Z-order curve is used together with quaternary tree, positioning can not only be easy, and can also be overcome continuous Property difference caused by inquiry inaccuracy disadvantage, and by analytical proof its can guarantee certain inquiry accuracy rate, therefore need by Quaternary tree and Z-order curve, which combine, forms Z-order tree, is used for storage location information.Using improved privacy homomorphic algorithm It is encrypted and decrypted, it can be with the safety of effective guarantee user privacy information；Moreover, LBS client is by encrypted target section Point encryption data and adjacent node encryption data are sent to LBS server, so that LBS server calculates secret range data, with So that LBS server is shared the computing cost of LBS client, to guarantee that the expense of LBS client is lower, guarantees its normal work Make.

Detailed description of the invention

In order to illustrate the technical solution of the embodiments of the present invention more clearly, below by institute in the description to the embodiment of the present invention Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the invention Example, for those of ordinary skill in the art, without any creative labor, can also be according to these attached drawings Obtain other attached drawings.

Fig. 1 is a schematic diagram of hibert curve and Moore curve；

Fig. 2 is the application scenarios signal of the location privacy protection method in one embodiment of the invention based on NN Query Figure；

Fig. 3 is a flow chart of the location privacy protection method in one embodiment of the invention based on NN Query；

Fig. 4 is the flow chart that Z-order tree is created in one embodiment of the invention；

Fig. 5 is the schematic diagram using one region of quadtree decomposition；

Fig. 6 is a schematic diagram of Z-order curve；

Fig. 7 is to carry out the schematic diagram that second decomposition is formed by grid map to a certain region；

Fig. 8 is a schematic diagram for the Z-order tree formed based on the grid schematic diagram in Fig. 7；

Fig. 9 is a schematic diagram of the target subtree being truncated out based on Z-order tree in Fig. 8；

Figure 10 is the trend chart that CPU calculates the time in PCQP, PIR-NN and DPIR-NN；

Figure 11 is a trend chart of Network Transmission Delays in PCQP, PIR-NN and DPIR-NN；

Figure 12 is the trend chart that kNN inquires accuracy rate on Sequoia data set；

Figure 13 is the trend chart that kNN inquires accuracy rate on Uniform data set；

Figure 14 is a schematic diagram of computer equipment in one embodiment of the invention.

Specific embodiment

Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete Site preparation description, it is clear that described embodiments are some of the embodiments of the present invention, instead of all the embodiments.Based on this hair Embodiment in bright, every other implementation obtained by those of ordinary skill in the art without making creative efforts Example, shall fall within the protection scope of the present invention.

Location privacy protection method provided in an embodiment of the present invention based on NN Query, should the position based on NN Query Method for secret protection can be using in application environment as shown in Figure 2.It specifically, should the location privacy protection side based on NN Query Method is applied in LBS system (hereinafter referred to as LBS system), should include such as Fig. 1 based on the location privacy protection system of NN Query Shown in LBS client and LBS server, LBS client is communicated with LBS server by network, for realizing in base Inquiry kNN inquiry is carried out in the LBS of encryption, and guarantees the efficiency and accuracy rate of kNN inquiry.Wherein, LBS client is also known as used Family end, refers to corresponding with LBS server, provides the program of local service for client.LBS client it is mountable but be not limited to On the mobile terminals such as various personal computers, laptop, smart phone, tablet computer and portable wearable device.LBS Server can realize with the LBS server cluster of independent LBS server either multiple LBS servers composition, this reality It applies in example, wherein LBS client is to issue the terminal of inquiry request or other requests.LBS server includes destination server Or intermediate server.Destination server is the LBS server for answering all inquiry requests.Intermediate server is that setting exists Retransmission unit between LBS client and destination server can cache the destination server content of a part, and to LBS visitor The device that the component requests that family end is sent are answered.It is to be appreciated that for the processing of support distribution formula, due to intermediate server Middle a part of destination server content of caching, therefore, intermediate server can be to relevant to this partial target server content Request is answered.

In one embodiment, as shown in figure 3, a kind of location privacy protection method based on NN Query is provided, with the party Method is illustrated for applying LBS client and LBS server in Fig. 2.

The location privacy protection method based on NN Query includes LBS client executing following steps:

S11:LBS client generates NN Query request, and NN Query request includes target position, interest vertex type, close Adjacent number and private cipher key.

S12:LBS client is based on target position and interest vertex type, and inquiry is in advance based on quaternary tree and Z-order curve The Z-order tree corresponding with interest vertex type of creation, obtains corresponding destination node and target subtree.

S13:LBS client traverses the leaf node in target subtree, obtains the corresponding node of leaf node and coordinate is truncated, Node truncation coordinate includes destination node truncation coordinate and adjacent node truncation coordinate.

S14:LBS client is based on private cipher key, is truncated and is sat to destination node using improved privacy homomorphic encryption algorithm Mark and adjacent node truncation coordinate carry out computations, obtain destination node encryption data and adjacent node encryption data.

Destination node encryption data and adjacent node encryption data are sent to LBS server by S15:LBS client, and are connect That receives that LBS server returns carries out homomorphism addition in ciphertext and multiplies to destination node encryption data and adjacent node encryption data Method calculates secret range data obtained.

S16:LBS client be based on private cipher key, using improved privacy homomorphic decryption algorithm to secret range data into Row decryption, obtains the distance between destination node and adjacent node respectively.

S17:LBS client is based on the distance between destination node and adjacent node and is ranked up, and obtains and neighbour's number Corresponding neighbour's point of interest.

The location privacy protection method based on NN Query further includes that LBS server executes following steps:

S21:LBS server receives the destination node encryption data and adjacent node encryption data that client is sent, target Node encrytion data and adjacent node encryption data are that destination node is truncated using improved privacy homomorphic encryption algorithm respectively Coordinate and adjacent node truncation coordinate carry out data acquired in computations.

S22:LBS server carries out the homomorphism addition in ciphertext to destination node encryption data and adjacent node encryption data It is calculated with multiplication, obtains secret range data.

Secret range data is sent to LBS client by S23:LBS server, so that LBS client is based on private cipher key, Secret range data is decrypted using improved privacy homomorphic decryption algorithm, obtain respectively destination node and adjacent node it Between distance, and be ranked up based on the distance between destination node and adjacent node, obtain it is corresponding with neighbour's number closely Adjacent point of interest.

Before executing step as shown in Figure 3 in LBS system, LBS server need to initialize Z-order tree in advance, and will Z-order tree is distributed to LBS client, so that LBS client can carry out quick kNN inquiry.It is based on quaternary tree and Z- The Z-order tree corresponding with interest vertex type of order curve creation, the Z-order tree are created based on Z-order curve, It is set to be convenient to calculate neighbour, to be conducive to save the computing cost of NN Query；And quaternary tree can support distributed storage, Help to carry out NN Query based on the Z-order tree.It is to be appreciated that created based on quaternary tree and Z-order curve with The corresponding Z-order tree of interest vertex type can be pre-created in LBS server, and the Z-order tree created is stored In LBS client, so that LBS client can be used directly, which once creates in LBS server, can be multiple It uses, does not count the computing cost of LBS system.As shown in figure 4, created based on quaternary tree and Z-order curve and point of interest The corresponding Z-order tree of type, specifically comprises the following steps:

S401: the corresponding original point of interest of same interest vertex type is labeled in area map.

Area map refers to some region of electronic map.The setting in the region can be carried out by administrative regions such as provinces and cities districts It divides.Specifically, by corresponding original point of interest (the i.e. original POIs) mark of same interest vertex type in area map, with Make to show the corresponding all original points of interest of the interest vertex type in the area map.For example, all hotels, Shenzhen are marked Note is completed the corresponding original point of interest of this interest vertex type of hotel being labeled in area map in Shenzhen's map Operation.

In order to guarantee that area map can store all original points of interest with quaternary tree storage result, each area map can be made It is square map, i.e., is square the center of map with the center in region, making whole region map includes all positions in region It sets.It is to be appreciated that since practical map is not the square map of rule, so that the original interest of two edges of regions positions It o'clock is labeled at least two area maps, to guarantee the accuracy of subsequent NN Query.

S402: according to the labeling position of original point of interest, area map is decomposed using quadtree decomposition rule, is obtained Corresponding grid map is taken, grid map includes at least one grid.

In LBS system, need to facilitate the institutional framework for calculating neighbour to store with a kind of POIs all in region in advance, To save the computing cost of LBS client during subsequent NN Query, and to support distributed storage, one is needed to have It is easy the storage organization of segmentation and invariance.Quaternary tree (quad-tree) this be in such a suitable storage two-dimensional map The data structure of POIs information.Quaternary tree is one kind of tree, and each internal node has 4 children, and four children are distinguished Northwest (NW), northeast (NE), southwestern orientation (SW) and the southeast (SE) of corresponding map, as shown in Figure 5.Specifically, according to same The labeling position of the corresponding all original points of interest of interest vertex type, divides area map using quadtree decomposition rule Solution, obtains corresponding grid map.The grid map is to store all original points of interest based on quaternary tree storage organization to be formed Map.The grid map includes at least one grid, and each grid can correspond to root node in quaternary tree storage organization, interior Portion's node or leaf node, wherein root node is the node of top layer in quaternary tree storage organization, no superior node.Leaf Node is the node of the bottom in quaternary tree storage organization, no downstream site.Internal node is between root node and leaf section Node between point, both having there is superior node, there is also downstream sites.

In the present embodiment, uses quadtree decomposition rule to carry out decomposition to area map and specifically include: firstly, with region The center of figure is the root node of its corresponding target quaternary tree, is by northwest, northeast, the corresponding region in southwest and southeast orientation Intermediate node carries out first time decomposition, is decomposed into tetra- regions A, B, C and D in (b) in Fig. 7.Then, judge each intermediate node Corresponding region whether there is POIs corresponding with interest vertex type (child i.e. in Z-order tree)；If it exists with interest The corresponding POIs of vertex type obtains four new areas then to being decomposed again in the corresponding area map of the intermediate node The corresponding intermediate node in domain repeats and judges the corresponding region of each intermediate node with the presence or absence of corresponding with interest vertex type POIs, reach preset times or all areas until decomposing number POIs corresponding with interest vertex type be not present； POIs corresponding with interest vertex type if it does not exist, then be determined as leaf node for intermediate node.As shown in fig. 7, using Quadtree decomposition rule carries out primary decompose to area map corresponding to Fig. 7 (a) and is formed by grid map as Fig. 7 (b), adopts Area map corresponding to Fig. 7 (a) is carried out second decomposition to be formed by grid map being Fig. 7 (c) with quadtree decomposition rule.

S403: processing, acquisition and interest are attached at least one grid in grid map using Z-order curve The corresponding Z-order tree of vertex type, Z-order tree include at least one node, and each node corresponding one is original for storing The number field of point of interest quantity.

Specifically, processing is attached at least one grid in grid map using Z-order curve, with formed with it is emerging The corresponding Z-order tree of interesting vertex type, it is opposite that this Z-order tree with quaternary tree storage organization stores same interest vertex type All POIs answered make it support distributed storage；Moreover, connecting between all POIs using Z-order curve, keep it real Now facilitate and calculate neighbour, can effectively save computing cost during NN Query.

As shown in fig. 7, the corresponding dot of each number indicates a POI in Fig. 7, the P19 table that P1, P2 is respectively adopted ... Show, as area map is overstepping the bounds of propriety thinner, the grid that POI is in grid map is not also identical, such as Fig. 7 (b) and (c).Based on Fig. 7 (c) the Z-order tree that corresponding grid map is formed is as shown in figure 8, the child of each node is marked as 00 in Z-order tree (upper left), 01 (upper right), 10 (lower-lefts) and 11 (bottom rights), i.e., first is y value, 0 be on, under 1 is；Second is x value, and 0 is A left side, 1 is right.Since kNN inquiry is related to the process of a sequence, so it is inadequate for only obtaining POIs all in certain area 's.However, although Z-order curve has discontinuity, i.e., in serial number it is neighbouring be not necessarily it is neighbouring on geographical location, But kNN search can be carried out on the Z-order tree of generation in conjunction with its arrangement regulation (i.e. by the ascending order sequence of node Z value), To save the computing cost during NN Query.

In kNN inquiry, sort again if being directly based upon all POI and being calculated with target position, to obtain k neighbour It is huge that the operation of point of interest (i.e. neighbour POI) will lead to its computing cost, and therefore, kNN inquiry should consider related interests point Density on map.Its reason is that the difference of density of certain original points of interest on map is very huge, such as 10 May have upper hundred restaurants in kilometer range, but hospital may only one even without.In the present embodiment, it is being formed by Z- Include at least one node in order, and make each node corresponding one for storing the number field of original point of interest quantity, such as schemes Shown in number in 8 in the corresponding circle of each node.It is to be appreciated that the number field for store in its corresponding node with The quantity of the corresponding all original points of interest of interest vertex type, for reflecting that it stores the density of the original point of interest, so as to It carries out kNN inquiry in the number field, to save computing cost.

In kNN query process, when the number in the number field in non-leaf nodes is greater than k, user can be truncated Z-order tree, only using calculative subtree as target subtree.At this point, by being deposited in leaf nodes all in target subtree The POI of storage is truncated with target position and comes out, to carry out kNN inquiry based on target subtree.In general, required for search Target subtree height is not high, this is all no small reduction for user institute byte overhead to be encrypted and communication overhead, thus Achieve the purpose that save expense.

In one embodiment, as shown in fig. 6, step S403, that is, use Z-order curve at least one in grid map A grid is attached processing, obtains Z-order tree corresponding with interest vertex type, specifically includes:

S4031: the corresponding mesh coordinate (x, y) of each grid in grid map is obtained.It is formed by grid map In two-dimensional space, using the upper left corner as origin, space coordinates are formed as unit of the corresponding grid of leaf node, and obtain each Mesh coordinate of the grid in the space coordinates.

S4032: to mesh coordinate (x, y) carry out Binary Conversion, obtain corresponding x-axis binary system coordinate and y-axis two into Coordinate processed.For example, acquired x-axis binary system coordinate is x=(3) when carrying out Binary Conversion to mesh coordinate (3,4)₁₀= (011)₂=(x₁x₂x₃), y-axis binary system coordinate is y=(4)₁₀=(100)₂=(y₁y₂y₃)。

S4033: position crossover operation is carried out to x-axis binary system coordinate and y-axis binary system coordinate, obtains node Z value.For example, Position crossover operation is carried out to x-axis binary system coordinate and y-axis binary system coordinate to refer to the 1st digit of inverse of x-axis binary system coordinate The 1st bit digital of inverse of binary node Z value after word intersects as position, the 1st bit digital of inverse of y-axis binary system coordinate is made The 2nd bit digital of inverse of binary node Z value after intersecting for position, using the 2nd bit digital of inverse of x-axis binary system coordinate as position The 3rd bit digital of inverse of binary node Z value after intersection is intersected the 2nd bit digital of inverse of y-axis binary system coordinate as position The 4th bit digital of inverse of binary node Z value afterwards, two after the 3rd bit digital of inverse of x-axis binary system coordinate is intersected as position The 5th bit digital of inverse of the node Z value of system, binary system after the 3rd bit digital of inverse of y-axis binary system coordinate is intersected as position Node Z value the 6th bit digital of inverse, then its obtain node Z value be Z ((011)₂, (100)₂)=(011010)₂= (26)₁₀。

S4034: according to the ascending order sequence of node Z value, processing is attached at least one grid, is obtained and point of interest class The corresponding Z-order tree of type.

Since the corresponding grid Z value of each grid is by its x-axis binary system coordinate of cross arrangement and y-axis binary system coordinate All grids can be attached processing according to the ascending order sequence of node Z value, to obtain Z-order by the numerical value calculated Curve forms Z-order tree corresponding with interest vertex type based on the Z-order tree and grid map.Due to Z-order song Line is the space filling region got up according to the ascending order sequential connection of node Z value, this feature may be implemented in tree structure Comparison distance is not needed to determine which node is more specific orientation be representated by, so that it may directly quickly navigate to and be searched Target.

Below in conjunction with the Z-order tree that step S401-S403 is created, to LBS client shown in Fig. 3 and LBS service Each step that device executes is described in detail:

Wherein, NN Query request is that the request of NN Query processing is carried out for triggering LBS client.The NN Query Request is the request that user triggers in LBS client, for inquiring the target point of interest of understanding needed for user (hereinafter referred to as Target POIs).Target position is the position that user clearly knows.Interest vertex type (hereinafter referred to as POI type) is user's needs The type of the point of interest of inquiry.For example, user can input " safety financial center in the input frame at LBS client-side search interface Hotel near mansion ", and click search button, that is, NN Query request is produced, then " safety financial center mansion " is Target position, and " hotel " is interest vertex type.Neighbour's number is the number for the neighbour for needing LBS client to return, the neighbour Number can independently be set by user, can also use the number of LBS system default.Private cipher key is being used for of being independently arranged of user The key of encryption or decryption process is carried out, which only has user's understanding, to reach secrecy effect, to ensure that user is hidden The safety of personal letter breath.

In the present embodiment, Z-order tree is the storage of space storage result and a certain interest vertex type phase using quaternary tree Corresponding all points of interest (i.e. POIs), and filling processing in space is carried out using Z-order curve and is formed by tree.Z- Order curve refers to a kind of space filling curve formed according to Z collating sequence, as shown in Figure 6.Destination node refers to Z- Node corresponding with target position in order tree.Target subtree is the son including destination node based on Z-order tree determination Tree.

In one embodiment, step S12, that is, be based on target position and interest vertex type, inquiry be in advance based on quaternary tree and The Z-order tree corresponding with interest vertex type of Z-order curve creation, obtains corresponding destination node and target subtree, Specifically comprise the following steps:

S121: be based on target position and interest vertex type, inquiry be in advance based on quaternary tree and Z-order curve creation with The corresponding Z-order tree of interest vertex type obtains destination node of the target position in Z-order tree.

Specifically, be based on target position and interest vertex type, inquiry be in advance based on quaternary tree and Z-order curve creation The Z-order tree that Z-order tree corresponding with interest vertex type, i.e. inquiry are created using step S401-S403 in advance, will The corresponding node in target position corresponding position in Z-order tree corresponding with interest vertex type is determined as destination node. In Fig. 7 and Fig. 8, if target position is P9, the mesh coordinate (00,10) in P9 grid map shown in Fig. 7 is obtained, it will The mesh coordinate (00,10) is converted into binary node Z value (1000) in Z-order curve, and node Z value (1000) is in Fig. 8 The path of middle direct search " 10 " → " 00 " can determine destination node.

S122: the original point of interest quantity of the corresponding number field of the superior node of destination node is successively determined as number of targets Amount, if destination number is greater than neighbour's number, the father's node for obtaining superior node is the target subtree of root node.

In the present embodiment, step S122 specifically comprises the following steps: (1) first by superior node (i.e. his father of destination node Close node) the original point of interest quantity of corresponding number field is determined as destination number.(2) judge whether destination number is greater than neighbour Number.(3) if destination number is greater than neighbour's number, by father's node of the superior node (i.e. its father's node) of destination node (i.e. the grandparent node of destination node) forms target subtree, the root of the target subtree as truncation node, based on the truncation node Node is the grandparent node of destination node.Destination number is greater than father's node of the superior node of neighbour's number as target The root node of subtree can fall into the boundary of father's node region to avoid target position and influence to inquire accuracy rate.(4) if The corresponding destination number of father's node is less than neighbour's number, then obtains superior node (the i.e. mesh of the superior node of destination node Mark the grandparent node of node) the original point of interest quantity of corresponding number field is determined as destination number, repeat step (2)- (4), to obtain target subtree.

As shown in figure 8, when passage path " 10 " → " 00 " positions P9, when searching father's node of P9, father's node Number field in number be 2, it is bigger than k=1, at this time by father's node (i.e. the grandparent node of P9) conduct of father's node of P9 The subtree of root node is determined as target subtree.At this point, the Z-order of all leaf node P8, P9, P14, P13 of target subtree Curve binary system coordinate is truncated into " 00 " " 00 " " 01 " and " 10 " respectively, be reduced into position coordinates (x, y) be (0,0) (0, 0) (1,0) and (0,1).It is to be appreciated that using the grandparent node of P9 rather than father's node as target subtree root node (or It is truncation node that person, which says), it is possible to prevente effectively from target position falls into the boundary of father's node region and influences the standard inquired True rate.

LBS client can traverse the target subtree after obtaining the target subtree (as shown in Figure 9) where target position In leaf node, obtain each leaf node in target subtree node truncation coordinate, so as to based on the node truncation sit Mark carries out NN Query calculating, obtains corresponding distance.Specifically, coordinate is truncated in destination node truncation coordinate and adjacent node. Destination node truncation coordinate refers to coordinate of the destination node in target subtree, and adjacent node truncation coordinate is target subtree In coordinate of each adjacent node in target subtree, the adjacent node be specially be truncated after target subtree in leaf section Point.In Fig. 9, the adjacent node in target subtree is P8, P9, P14 and P13, the adjacent node truncation in target subtree Coordinate is respectively (0,0) (0,0) (1,0) and (0,1), is indicated with binary system coordinate, respectively 00,00,01 and 10, and target It is (0,0) that coordinate, which is truncated, in node, is to indicate with binary system coordinate, is 00, so that the truncation that LBS client obtains 5 nodes is sat Mark.

Wherein, improved privacy homomorphic encryption algorithm be to simple privacy homomorphism (privacy homomorphisms, Hereinafter referred to as PH) Encryption Algorithm improve gained algorithm.Wherein, PH is a kind of encryption conversion, it is by one group on plaintext Operation is mapped to the operation in another group of ciphertext.In form, they are encryption function E_k: T' → T, it allows do not decrypting Function D_kIn the case where encryption data is operated.For example, p and q is two secret big prime numbers.M=p × q is public It opens.Collection is combined into T'=Z in plain text_m={ 0,1 ..., m-1 }, one group of plaintext operation that definition is closed in plaintext collection is F'= {+_m,-_m,×_m, it is the addition on mould m, subtraction and multiplication respectively.Ciphertext collection is T=Z_q×Z_q.Operational set F in ciphertext It is the F' calculated with componentwise, i.e., when carrying out operation to two ciphertexts, by the number of t corresponding position in two ciphertexts point It carry out not addition, subtraction and multiplying.Define private cipher key k=(p, q) and encryption function E_k(a)=[amodp, amodq]. Decryption uses Chinese remainder theorem: for k relatively prime positive integer n₁,...,n_k, in Z_N(N=n₁n₂...n_k) in existence anduniquess X, meet x ≡ a_i modn_i,Wherein, by using the division algorithm (Euclidean of extension Algorithm), decryption person can be easily found the x uniquely met in polynomial time.Obviously, add in simple HP In close algorithm, encryption is the PH under the operation that F' and F are defined, because of m=p × q.However, this cipher mode will receive it is " bright The influence of text " attack (known-plaintext attacks), that is to say, that if opponent knows a pair of corresponding plaintext and close Text, so that it may extrapolate the value of p and q.

Domingo-Ferrer improves above-mentioned simple PH Encryption Algorithm, and under the operation of same cleartext-ciphertext, Be exactly modulo addition, mould subtraction and modular multiplication, propose a privacy homomorphism method can verify that and safe, i.e., it is improved hidden Private homomorphic encryption algorithm.

In improved privacy homomorphic encryption algorithm, (t, m) is Public Key, i.e., Public Key includes one greater than 2 Positive integer t and integer m one big, m are to close number, that is, include at least one factor in addition to 1 and m.T represents a plaintext quilt Be divided into how many parts (in simple PH Encryption Algorithm, t=2；And t > 2 in improved privacy homomorphic encryption algorithm).m There should be many factors smaller than as t.Further, there is integer much smaller than m is that mould m is reversible, that is to say, thatThere is corresponding r^-1Make r^-1×r≡1modm.Wherein, private cipher key is (r, m'), including r and m', wherein r is integer Set Z_mIn a numerical value,There are corresponding r^-1Make r^-1×r≡1modm；M' is a factor greater than 1 of m, It is expressed as k=(r, m'), i.e., appointing a first factor greater than 1 in the numerous factors of m is m'.

Set is T'=Z in plain text_m′.The set of ciphertext is a t tuple, that is, T=(Z_m)^t.Plaintext operational set F' It is made of addition, subtraction and the multiplication on T', i.e. a ciphertext has the tuple of t number, and each number belongs to integer field collection M is closed, addition, subtraction and multiplication herein is the operation of modulus m.It is similar to simple PH Encryption Algorithm, ciphertext operational set F It is addition, subtraction and the multiplication calculated with componentwise on T.Finally, as follows based on improved privacy homomorphic encryption algorithm:

At random by plaintext a ∈ Z_m'It is divided into the part t, a₁,...,a_t, makea_j∈Z_m, after segmentation, Plaintext a becomes following form:

E_k(a)=(a₁rmodm,a₂r²modm,...,a_tr^t modm) (1)

Wherein, Z_m'For the integer set that 0 to m'-1 total m' integer is constituted, a is integer set Z_m'In a numerical value, It is a plaintext, Z_mFor the integer set that 0 to m-1 total m integer is constituted, (t, m) is Public Key, and t is just whole greater than 2 Number, m are to close number, and (r, m') is private cipher key, and m' is a factor of m, and r is integer set Z_mIn a numerical value, There are corresponding r^-1Make r^-1×r≡1modm。

In this step, LBS client is after getting destination node truncation coordinate and adjacent node truncation coordinate, by it Plaintext a is encrypted as plaintext a, and using above-mentioned formula (1), at this point, plaintext a is a lot of number, by this A lot of digital random is divided into the part t, it is made to meet formula (1), to obtain corresponding destination node encryption data and adjacent Node encrytion data.It is to be appreciated that due in private cipher key (r, m') r be the customized numerical value of user, use formula (1) Coordinate is truncated to destination node and after adjacent node truncation coordinate is encrypted, only knows that the specific value of r just can be with The location privacy information in destination node encryption data and adjacent node encryption data is known, to play privacy functions.

Specifically, LBS client is after obtaining destination node truncation coordinate and adjacent node truncation coordinate, if direct base Coordinate is truncated in destination node and adjacent node truncation coordinate calculates distance between destination node and adjacent node, LBS can be made The computing cost of client is bigger, influences the normal work of LBS client.If coordinate and adjacent directly is truncated in destination node Node truncation coordinate is sent to LBS server and carries out apart from calculating, and may cause user privacy information, (such as where is user, thinks Will be where) leakage therefore coordinate and adjacent segments first need to be truncated to destination node using improved privacy homomorphic encryption algorithm Point truncation coordinate is encrypted, so that encrypted destination node encryption data and adjacent node encryption data reach privacy secrecy Purpose.Encrypted destination node encryption data and adjacent node encryption data are sent to LBS service again by LBS client Device so that LBS server based on destination node encryption data and adjacent node encryption data carry out homomorphism addition in ciphertext and Multiplication calculates, and to obtain the secret range data of distance between reflection destination node and adjacent node, saves LBS visitor to reach The purpose of the computing cost at family end.

Specifically, LBS server can receive the destination node encryption data and adjacent node encryption number that LBS client is sent According to since the destination node encryption data and adjacent node encryption data are using improved privacy homomorphic encryption algorithm pair respectively Coordinate is truncated in destination node and adjacent node truncation coordinate carries out data acquired in computations, and improved privacy homomorphism adds R is the customized numerical value of user in private cipher key used by close algorithm (r, m'), so that encrypted destination node encrypts number It can only be decrypted based on private cipher key according to adjacent node encryption data, to play the mesh for ensureing user privacy information safety , to avoid destination node encryption data and adjacent node encryption data in LBS client and quilt in LBS server transmission process Intercept and cause the leakage of user privacy information.

Specifically, LBS server is after receiving destination node encryption data and adjacent node encryption data, to target Node encrytion data and adjacent node encryption data carry out homomorphism addition and multiplication calculating in ciphertext, i.e., encrypt to destination node Data and adjacent node encryption data carry out homomorphism addition by the operation of ciphertext operational set F' setting and calculate with multiplication, to obtain The distance between destination node and each adjacent node for being indicated using ciphertext form, i.e., secret range data.In the present embodiment, Ciphertext operational set F' includes following content:

One is the identical component of r degree carries out corresponding subtraction or add operation when homomorphism addition and subtraction, i.e. operation.

The second is homomorphism multiplication, i.e., all items are in Z_mMiddle multiplication cross, that is to say, that spend for t₁Xiang Yudu is t₂Item Be multiplied can degree of obtaining be t₁+t₂Item.The item with identical r degree is added up later.

Although improved privacy homomorphic algorithm (including Encryption Algorithm and decipherment algorithm) can directly hold on cryptogram Row addition, subtraction and multiplying, but there is still a need for a large amount of calculating of cost for these operations, if directly in LBS client It is calculated, so that the expense in LBS client is larger, influences its normal performance.For example, allowing η₊Indicate component of degree n n summation Computing cost, η_×Indicate the expense of this component multiplication.So, the expense of three kinds of operations is t η respectively in ciphertext operational set₊、 tη₊And t²η_×+(t²-t)η₊.It should be noted that multiplication can make the length of ciphertext become 2t component from t component.

Since each component requires multiplication operation.So the computing cost of encryption is t η_×.Decrypt expense and encryption Expense compare, only in finally mostly accumulation operations, therefore its expense is t²η_×+(t-1)η₊.It is worth noting that, after encryption The length of ciphertext be t times of length of plaintext before encrypting.Because each component is Z_mIn a positive integer, so ciphertext Size be tl (m), wherein l (m) indicate m binary length.

In practical applications, since the add operation expense of each component and the multiplying expense comparison of each component are It is not worth mentioning, thus addition and multiplication all in the presence of, can not consider the computing overhead of addition.For the multiplication fortune on component For calculation, multiplying especially on large modulus, if using MontgomeryReduction, operation can become Very efficiently, can achieve 10^-5Second grade.

It is to be appreciated that being carried out in ciphertext in LBS server to destination node encryption data and adjacent node encryption data Homomorphism addition and multiplication calculate, to be taken in the encryption LBS that obtains the process of secret range data and realized based on Moore curve Be engaged in device calculating process it is similar, but in the present embodiment, LBS server only need to after truncation destination node and adjacent node carry out It calculates, so that its number of computations is less, efficiency is higher.

Specifically, calculated secret range data is sent to LBS client by LBS server, due to the secret distance Data are to be based on private cipher key (r, m') using improved privacy homomorphic encryption algorithm in advance to carry out encrypted data progress together State addition and multiplying numerical value obtained, and r is the customized numerical value of user, other illegal users do not grasp private cipher key (r, m') can not just be decrypted, so that secret range data can ensure the safety of its user privacy information when intercepted.LBS clothes Secret range data is sent to LBS client by business device, so that LBS client is decrypted and determines corresponding neighbour's interest Point to achieve the purpose that NN Query, and reduces the computing cost of LBS client.

Specifically, use improved privacy homomorphic decryption algorithm as follows:

R is calculated using r^-1, by ciphertext (e₁,e₂,...,e_t) be calculated as follows, to obtain plaintext a

Wherein, (t, m) is Public Key, and t is the positive integer greater than 2, and m is to close number, and (r, m') is private cipher key, m' m A factor, r be integer set Z_mIn a numerical value,There are corresponding r^-1Make r^-1× r ≡ 1modm, Z_mIt is 0 The integer set constituted to the total m integer of m-1, e are the secret range data of ciphertext, and a is plaintext, specially destination node and phase The distance between neighbors.I.e. in the present embodiment, e is that LBS server encrypts number to destination node encryption data and adjacent node Secret range data obtained is calculated according to the homomorphism addition and multiplication that carry out in ciphertext, a is that LBS client is based on secret distance The distance between destination node and adjacent node that data obtain after being decrypted.

The improved privacy homomorphic algorithm (including Encryption Algorithm and decipherment algorithm) can resist plaintext attack.Its reason exists In due to the set sizes exponentially s-n grades of growth (s=log of the possible key pair of the known cleartext-ciphertext pair of n_m'm)。 Even if the key pair that he is derived also still may belong to arbitrarily large collection this means that attacker knows corresponding cleartext-ciphertext It closes, therefore plaintext attack can be resisted.

It is applied in improved privacy homomorphic algorithm (including Encryption Algorithm and decipherment algorithm) privately owned on insincere data cloud In querying method, by it is true the experiment proves that improved privacy homomorphic algorithm efficiency.In experiment, the range of plaintext be [0, 10⁶], due to addition to be executed, multiplication, so the size of plaintext domain m' must have 4 × 10¹².Finally experiment setting m ∈ [2 × 10¹⁷, 8×10¹⁷],m′∈[2×10¹⁵,4×10¹⁵], r ∈ [10¹⁵,2×10¹⁵], t=3.Encryption key is (m', r).In processor For Dual 4-cores Intel Xeon X5570 2.93GHz CPU, 32GB RAM, GNU/Linux operating system is run On IBM eserver xSeries 335, the data deciphering CPU overhead of a 20bits be about 2ms (and in paper, 40 The CPU overhead of data is 80ms), therefore can effectively reduce expense.

Specifically, LBS client is ranked up based on the distance between destination node and adjacent node, obtains sequence knot Fruit, according to position of the destination node in the ranking results and neighbour's number, can quick obtaining it is corresponding with neighbour's number close Adjacent point of interest.In above-described embodiment, after being ranked up according to the distance between destination node and adjacent node, the sequence of acquisition As a result the sequence of (P8, P9, P13, P14) can be obtained, because of k=1, therefore it is exactly P8 or P9 that user, which can inquire neighbour's point of interest, (the reason of cannot further navigating to P9 is because the division of leaf node is not thin enough).In practical applications, division should Guarantee only one POI in each leaf node, the corresponding closest node pair of destination node can thus be accurately positioned The neighbour's point of interest answered.

In location privacy protection method based on NN Query provided by the present embodiment, it is based on quaternary tree and Z-order Curve creation Z-order tree in so that Z-order tree tree support distributed storage, have be easy divide with not The characteristics of denaturation.Although Z-order curve than the poor continuity of hibert curve, its it is more more regular than hibert curve and Positioning is easy, and Z-order curve is used together with quaternary tree, and positioning can not only be easy, can also overcome poor continuity The disadvantage of caused inquiry inaccuracy, and by analytical proof, it can guarantee certain inquiry accuracy rate and efficiency, therefore need Quaternary tree and Z-order curve are combined and form Z-order tree, is used for storage location information.It is calculated using improved privacy homomorphism Method is encrypted and decrypted, can be with the safety of effective guarantee user privacy information；Moreover, LBS client is by encrypted target Node encrytion data and adjacent node encryption data are sent to LBS server, so that LBS server calculates secret range data, So that LBS server shares the computing cost of LBS client, to guarantee that the expense of LBS client is lower, guarantee its normal work Make.

The location privacy protection method provided by the present embodiment based on NN Query is carried out below in conjunction with specific experiment Performance evaluation.In experiment, designed all kNN inquiry experiment is all to be used as most to terminate using the average result of 10 experiments Fruit can greatly eliminate randomness in this way, guarantee the correctness of data result.In the present embodiment, this will be tested and assessed KNN vlan query protocol VLAN designed by text is contribute and is realized with kNN lookup code by C++, and runs in our experimental situation, real It is as shown in Table 1 to test environment.It wherein, is sex work of LBS server due to contributing, in later inquiry also no longer again It contributes.Therefore this part can be not belonging to user overhead information as being system initialization expense.

In terms of experimental data set, there is employed herein Sequoia data sets comprising 62556 U.S. add the state Li Fuliya True place name and coordinate.According to Sequoia data set, first by all places be divided into 9 classes (School, Lake, Bridge, Creek, Farm, Mine, Golf Course, Hospital, Campground), establish institute's POIs data set to be used. To obtain the accuracy that corresponding correct inquiry data set is used to experiment with computing scheme according to every one kind POI, this experiment is not surpassing 1000 user query coordinates have been randomly generated in the ground point range crossed in Sequoia data set, and have been calculated using Euclidean distance The neighbor coordinate set of 50 arest neighbors of every one kind POI inquiry out.Just obtained in this way this experiment institute it is to be used be used to pair The correct query result data set of ratio.Accuracy rate calculation formula is accuracy rate=(| R ∩ G |)/(| G |), and wherein G is The actual neighborhood calculated using Euclidean distance, R are the neighborhoods that experiment obtains.For the justice for guaranteeing Experimental comparison Property, construct the map grid size and DPIR-NN (Distributed private information of hibert curve Retrieval-nearest neighbor, i.e., distributed specific information retrieval-neighbour, is as based on provided by the present embodiment The corresponding NN Query strategy of the location privacy protection method of NN Query) size of the leaf node of tree that is constructed in inquiry All meetIn addition to this, an equally distributed data set has been synthesized herein, there are 65536 POIs, it Be evenly distributed in square area.These data processing works are completed by python.

One experimental situation of table

Used ASM-PH encryption parameter is similar with document in experiment, is set as plaintext domain m ' ∈ [2 × 10¹⁵,4× 10¹⁵], encryption key is (m', r), wherein m ∈ [2 × 10¹⁷,8×10¹⁷], r ∈ [10¹⁵,2×10¹⁵], component number t=3.

Firstly, to PCQP (Private Circular Query Protocol, i.e. privacy cyclic query agreement), PIR- NN (private information retrieval-nearest neighbor, i.e. specific information retrieval-neighbour), DPIR- (Distributed private information retrieval-nearest neighbor, that is, be distributed by NN and DPIR-NN Location privacy protection method based on NN Query provided by formula specific information retrieval-neighbour, as the present embodiment is corresponding NN Query strategy) user's computing cost is assessed in scheme, and PCQP, PIR-NN and DPIR-NN are shown in Figure 10 These three schemes with k variation when, the CPU runing time of LBS client, in Figure 11 show PCQP, PIR-NN and DPIR- The Network Transmission Delays of these three schemes of NN with k variation when, the Network Transmission Delays of LBS client.

It can be seen that runing time is increasing when k is bigger, changes comparatively fast in [1,10] section, change gradually become later It is slow.This is because map area diameter and k are quadratic relationships where k neighbours, and the diameter of map area is bigger, is wanted in tree The range of lookup is bigger, then the expense of user encryption decryption is more.So this k more large overhead increase occur more not Obvious phenomenon.Exist with PCQP (Private Circular Query Protocol, i.e. privacy cyclic query agreement) method Win7OS, Intel i5-2400 3.1GHz processor, on 8GB RAM PC, encryption needed for user in a kNN inquiry For decryption time is 1.7286s, not in a magnitude.This is because PCQP method needs to add using Paillier Close system encryption contains the offset vector of 62556 (POI number in data set) a elements to LBS server.Regardless of how DPIR-NN need to only use a bit simple improved privacy homomorphic algorithm by the user encryption part position POI.

Although in DPIR-NN, user's CPU time is short, since LBS server needs to be calculated in ciphertext, and PIR-NN (private information retrieval-nearest neighbor, i.e. specific information retrieval-neighbour) with In PCQP, LBS server need to only take out the adjacent position number in a certain position or so from corresponding hibert curve, Therefore its LBS server end computing cost is larger.Wherein, the main computing cost of PCQP method is all consumed in Paillier in fact In encryption.KNN inquiry, LBS server needs calculate n in Paillier ciphertext_pSecondary power operation and n_p- 1 multiplication fortune It calculates, is about 28.2 × k (s) the time required to the LBS server of a monokaryon.PIR-NN need to be n in ciphertext domain_pSecondary multiplication fortune It calculates, dependent on the size of ciphertext domain, generally requires 20-60 (s).And DPIR-NN, the work of LBS server include two parts.The A part is to calculate all leaf nodes under submitted root node to encrypt coordinates at a distance from user location encryption coordinate；Second Part is that DPIR inquiry details are further used according to the serial number of the k neighbour obtained.In first part, user is not considered The case where be sent to LBS server is subtree, LBS server use all 62556 the leaf node of ASM-PH encryption Coordinate all with user coordinates calculate a distance, operation time also less than 1s, so LBS server computing cost is substantially all wave Take in the inquiry of further information.If the detail time of the k neighbours POI found all needs to know, then in DPIR-NN The calculating time of LBS server be about the calculating time of LBS server in k times and PCQP of PIR-NN scheme it is similar.

Figure 11 simulates address using ns-3, and environmental parameter is with shown in above-mentioned table one, PCQP, PIR-NN for being shown by Figure 11 Comparison with DPIR-NN these three Network Transmission Delays is it is found that since DPIR-NN includes communication process twice, so that its biography Defeated delay will be high than PIR and PCQP method.

Then, proposed kNN inquiry privacy is tested using Sequoia data set and Uniform data set to protect The kNN protected between algorithm DPIR-NN and two kinds of correlation techniques PCQP and PIR-NN inquires accuracy rate (result is as shown in figure 12), this Two methods all fill target map using hibert curve, and find nearest neighbours.Based on PIR-NN's and PCQP Feature inquires a kNN, and the size (or quantity of disclosed POIs) for the result set that LBS server returns is respectivelyWith 3k, n_pIt is the number of all POI in map.

In figure 12 it can be seen that compared with PCQP, when the sum for returning to POIs is within 20, DPIR-NN ratio PCQP It is good.Even if the sum for returning to POIs is increased to k=50, the accuracy rate of DPIR-NN be still also it is equally matched with PCQP, at most Also only poor 2% accuracy rate.Although PIR-NN method executes very well in lesser situation, the case where increasing Under, the precision property of PIR-NN declines rapidly, this is because PIR-NN's is designed only for inquiry arest neighbors.Although also, I Although method when k is gradually increased accuracy rate reduce always, these variations are acceptables because accuracy rate decline Less than 5%.The reason of analyzing DPIR-NN decline, it should be because of the increase with k, the more farther neighbours of inquiring position Node has exceeded query sub tree root node range, causes to lack, and corresponding accuracy rate can reduce.

As seen from Figure 13, DPIR-NN is better than PCQP in the accuracy rate of synthesis being evenly distributed on data set, this is Because data point is equally distributed, it means that the tree constructed compares balance, therefore can see on more uniform data set The superiority of DPIR-NN out.

In this experiment, the kNN query strategy of a suitable DPIR is devised, Z-order curve and quaternary tree knot is utilized The Z-order tree being configured to carrys out storage location information, and user position location is made to become simple and quick；It has used and has supported the hidden of multiplication Private homomorphic encryption algorithm (i.e. improved privacy homomorphic encryption algorithm) inquires content to maintain secrecy, and makes LBS server and intermediary service Device is only responsible for calculating, so as to which in the case where LBS server is unaware of location information, user obtains k neighbor information；Due to Tree construction is still tree after decomposing, and naturally supports distributed storage, so as to be used in combination with DPIR, facilitates user into one Step inquiry.By using real data set (i.e. Sequoia data set) and generated data collection (i.e. Uniform data set) two numbers DPIR-NN strategy is tested according to collection, the experimental results showed that, DPIR-NN user's calculating time is few and inquiry accuracy rate is high.

It should be understood that the size of the serial number of each step is not meant that the order of the execution order in above-described embodiment, each process Execution sequence should be determined by its function and internal logic, the implementation process without coping with the embodiment of the present invention constitutes any limit It is fixed.

In one embodiment, a kind of location privacy protection system based on NN Query is provided, it should be based on NN Query Location privacy protection method in location privacy protection system and above-described embodiment based on NN Query corresponds.It should be based on close The location privacy protection system of neighbour's inquiry, including LBS client and LBS server.

LBS client executing following steps:

NN Query request is generated, NN Query request includes target position, interest vertex type, neighbour's number and privately owned close Key.

Based on target position and interest vertex type, inquiry is in advance based on quaternary tree and the creation of Z-order curve and interest The corresponding Z-order tree of vertex type, obtains corresponding destination node and target subtree.

The leaf node in target subtree is traversed, the corresponding node of leaf node is obtained and coordinate is truncated, coordinate is truncated in node Coordinate is truncated including destination node and coordinate is truncated in adjacent node.

Based on private cipher key, coordinate is truncated to destination node using improved privacy homomorphic encryption algorithm and adjacent node is cut Disconnected coordinate carries out computations, obtains destination node encryption data and adjacent node encryption data.

Destination node encryption data and adjacent node encryption data are sent to LBS server, and receives LBS server and returns The homomorphism addition carried out in ciphertext to destination node encryption data and adjacent node encryption data returned is obtained with multiplication calculating Secret range data.

Based on private cipher key, secret range data is decrypted using improved privacy homomorphic decryption algorithm, is obtained respectively Take the distance between destination node and adjacent node.

It is ranked up based on the distance between destination node and adjacent node, it is emerging to obtain neighbour corresponding with neighbour's number Interesting point.

LBS server executes following steps:

Receive client send destination node encryption data and adjacent node encryption data, destination node encryption data and Adjacent node encryption data is that coordinate and adjacent node is truncated to destination node using improved privacy homomorphic encryption algorithm respectively Coordinate is truncated and carries out data acquired in computations.

Homomorphism addition and multiplication calculating in ciphertext are carried out to destination node encryption data and adjacent node encryption data, obtained Take secret range data.

Secret range data is sent to LBS client, so that LBS client is based on private cipher key, using improved hidden Secret range data is decrypted in private homomorphic decryption algorithm, obtains the distance between destination node and adjacent node respectively, and It is ranked up based on the distance between destination node and adjacent node, obtains neighbour's point of interest corresponding with neighbour's number.

Specific restriction about the location privacy protection system based on NN Query may refer to above for based on close The restriction of the location privacy protection method of neighbour's inquiry, details are not described herein.The above-mentioned location privacy protection system based on NN Query Modules in system can be realized fully or partially through software, hardware and combinations thereof.Above-mentioned each module can be in the form of hardware It is embedded in or independently of the storage that in the processor in computer equipment, can also be stored in a software form in computer equipment In device, the corresponding operation of the above modules is executed in order to which processor calls.

In one embodiment, a kind of computer equipment is provided, which can be LBS client can also be with LBS server, internal structure chart can be as shown in figure 14.The computer equipment includes the processing connected by system bus Device, memory, network interface and database.Wherein, the processor of the computer equipment is for providing calculating and control ability.It should The memory of computer equipment includes non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operation System, computer program and database.The built-in storage is operating system and computer program in non-volatile memory medium Operation provide environment.The database of the computer equipment is for executing the location privacy protection method process based on NN Query The data of middle use or formation, such as Z-order tree.The network interface of the computer equipment is used to pass through with external terminal Network connection communication.To realize a kind of location privacy protection side based on NN Query when the computer program is executed by processor Method.

In one embodiment, a kind of computer equipment is provided, including memory, processor and storage are on a memory And the computer program that can be run on a processor, processor are realized in above-described embodiment when executing computer program based on neighbour The step of location privacy protection method of inquiry, such as step shown in Fig. 3, to avoid repeating, which is not described herein again.

In one embodiment, a computer readable storage medium is provided, meter is stored on the computer readable storage medium Calculation machine program, the computer program realize the location privacy protection in above-described embodiment based on NN Query when being executed by processor The step of method, such as step shown in Fig. 3, to avoid repeating, which is not described herein again.

Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein, To any reference of memory, storage, database or other media used in each embodiment provided herein, Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM (PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms, Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..

It is apparent to those skilled in the art that for convenience of description and succinctly, only with above-mentioned each function Can unit, module division progress for example, in practical application, can according to need and by above-mentioned function distribution by different Functional unit, module are completed, i.e., the internal structure of described device is divided into different functional unit or module, more than completing The all or part of function of description.

Embodiment described above is merely illustrative of the technical solution of the present invention, rather than its limitations；Although referring to aforementioned reality Applying example, invention is explained in detail, those skilled in the art should understand that: it still can be to aforementioned each Technical solution documented by embodiment is modified or equivalent replacement of some of the technical features；And these are modified Or replacement, the spirit and scope for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution should all It is included within protection scope of the present invention.

Claims

1. a kind of location privacy protection method based on NN Query, which is characterized in that the following step including LBS client executing It is rapid:

Generate NN Query request, NN Query request includes target position, interest vertex type, neighbour's number and privately owned close Key；

Based on the target position and the interest vertex type, inquiry be in advance based on quaternary tree and the creation of Z-order curve with The corresponding Z-order tree of the interest vertex type, obtains corresponding destination node and target subtree；

The leaf node in the target subtree is traversed, the corresponding node truncation coordinate of the leaf node, the node are obtained Truncation coordinate includes destination node truncation coordinate and adjacent node truncation coordinate；

Based on the private cipher key, coordinate and the phase are truncated to the destination node using improved privacy homomorphic encryption algorithm Neighbors is truncated coordinate and carries out computations, obtains destination node encryption data and adjacent node encryption data；

The destination node encryption data and the adjacent node encryption data are sent to LBS server, and receive the LBS The homomorphism addition destination node encryption data and the adjacent node encryption data carried out in ciphertext that server returns Secret range data obtained is calculated with multiplication；

Based on the private cipher key, the secret range data is decrypted using improved privacy homomorphic decryption algorithm, point It Huo Qu not the distance between destination node and adjacent node；

It is ranked up, is obtained corresponding with neighbour's number close based on the distance between the destination node and adjacent node Adjacent point of interest.

2. as claimed in claim 2 based on the location privacy protection method of NN Query, which is characterized in that described based on described Target position and the interest vertex type, inquiry be in advance based on quaternary tree and Z-order curve creation with the point of interest class The corresponding Z-order tree of type, obtains corresponding destination node and target subtree, comprising:

Based on the target position and the interest vertex type, inquiry be in advance based on quaternary tree and the creation of Z-order curve with The corresponding Z-order tree of the interest vertex type, obtains destination node of the target position in the Z-order tree；

The original point of interest quantity of the corresponding number field of the superior node of the destination node is successively determined as destination number, if The destination number is greater than neighbour's number, then the father's node for obtaining the superior node is the target subtree of root node.

3. as described in claim 1 based on the location privacy protection method of NN Query, which is characterized in that described using improvement Privacy homomorphic encryption algorithm include:

At random by plaintext a ∈ Z_m'It is divided into the part t, a₁,...,a_t, makea_j∈Z_m, after segmentation, plaintext a Become following form:

E_k(a)=(a₁rmodm,a₂r²modm,...,a_tr^tmodm)

Wherein, Z_m'For the integer set that 0 to m'-1 total m' integer is constituted, a is integer set Z_m'In a numerical value, be one In plain text, Z_mFor the integer set that 0 to m-1 total m integer is constituted, (t, m) is Public Key, and t is the positive integer greater than 2, and m is to close Number, (r, m') are private cipher key, and m' is a factor of m, and r is integer set Z_mIn a numerical value,There are correspondences R^-1Make r^-1×r≡1modm。

4. as described in claim 1 based on the location privacy protection method of NN Query, which is characterized in that described using improvement Privacy homomorphic decryption algorithm include:

Wherein, (t, m) is Public Key, and t is the positive integer greater than 2, and m is to close number, and (r, m') is private cipher key, and m' is the one of m A factor, r are integer set Z_mIn a numerical value,There are corresponding r^-1Make r^-1× r ≡ 1modm, Z_mIt is 0 to m- The integer set that 1 total m integer is constituted, e are the secret range data of ciphertext, and a is plaintext, specially destination node and adjacent segments The distance between point.

5. a kind of location privacy protection method based on NN Query, which is characterized in that the following step executed including LBS server It is rapid:

Receive client send destination node encryption data and adjacent node encryption data, the destination node encryption data and The adjacent node encryption data is that coordinate and adjacent is truncated to destination node using improved privacy homomorphic encryption algorithm respectively Node is truncated coordinate and carries out data acquired in computations；

The homomorphism addition and multiplication meter in ciphertext are carried out to the destination node encryption data and the adjacent node encryption data It calculates, obtains secret range data；

The secret range data is sent to LBS client, so that the LBS client is based on private cipher key, using improvement Privacy homomorphic decryption algorithm the secret range data is decrypted, obtain between destination node and adjacent node respectively Distance, and be ranked up based on the distance between the destination node and adjacent node, it obtains corresponding with neighbour's number Neighbour's point of interest.

6. as claimed in claim 5 based on the location privacy protection method of NN Query, which is characterized in that in the reception visitor Before destination node encryption data and adjacent node encryption data that family end is sent, the location privacy based on NN Query is protected Maintaining method includes:

The corresponding original point of interest of the same interest vertex type is labeled in area map；

According to the labeling position of the original point of interest, the area map is decomposed using quadtree decomposition rule, is obtained Corresponding grid map is taken, the grid map includes at least one grid；

Processing is attached to grid described at least one of described grid map using Z-order curve, obtain with it is described The corresponding Z-order tree of interest vertex type, the Z-order tree include at least one node, and each node corresponding one is used for Store the number field of original point of interest quantity.

7. as claimed in claim 6 based on the location privacy protection method of NN Query, which is characterized in that described to use Z- Order curve is attached processing to grid described at least one of described grid map, obtains and the interest vertex type Corresponding Z-order tree, comprising:

Obtain the corresponding mesh coordinate (x, y) of each grid in the grid map；

Binary Conversion is carried out to the mesh coordinate (x, y), obtains corresponding x-axis binary system coordinate and y-axis binary system coordinate；

Position crossover operation is carried out to the x-axis binary system coordinate and the y-axis binary system coordinate, obtains node Z value；

According to the ascending order sequence of the node Z value, processing is attached at least one described grid, is obtained and the point of interest The corresponding Z-order tree of type.

8. a kind of location privacy protection system based on NN Query, which is characterized in that including LBS client and LBS server；

The LBS client executing following steps:

It is ranked up, is obtained corresponding with neighbour's number close based on the distance between the destination node and adjacent node Adjacent point of interest；

The LBS server executes following steps:

9. a kind of computer equipment, including memory, processor and storage are in the memory and can be in the processor The computer program of upper operation, which is characterized in that the processor realized when executing the computer program as claim 1 to The step of location privacy protection method described in 7 any one based on NN Query.

10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, and feature exists In realization is as described in any one of claim 1 to 7 based on the position of NN Query when the computer program is executed by processor The step of method for secret protection.