Position privacy approximate continuous Skyline query processing method based on public places under road network

Abstract

The invention discloses a position privacy approximate continuous Skyline query processing method based on public places under a road network, and aims to quickly and continuously search results meeting Skyline definition while protecting the position privacy of a query user in the road network and improve the influence of non-space factors in road network Skyline query. At present, research on continuous inquiry of a road network does not consider the problem of location privacy of a user in an inquiry process, and the method constructs hidden location information of a public safety approximate area for the inquiry user based on a large number of public places in the road network, and further provides a new dominance, namely public safety approximate dominance. The influence of the spatial distance dimension on the non-spatial dominating relation is reduced, and the pruned data set is quickly filtered through the PWDG-Tree index and the weak data points. According to the method, the approximate continuous Skyline result meeting the requirement of protecting the location privacy of the user can be quickly obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of data query, and in particular to a method for processing location privacy approximate continuous skyline queries based on public places in a road network. Background Art

[0002] Due to the growth of mobile devices such as smartphones, as well as the availability of cheap location locators and network bandwidth, location-based services are becoming more and more popular. Continuous Skyline Query of Road Network is an effective method to solve the Skyline query problem in road networks with constantly changing locations. It continuously updates and optimizes the Skyline result set of the road network according to the changing location of the query object, and has wide applications in real-time navigation systems, intelligent transportation systems, and urban planning.

[0003] Continuous skyline query of road network requires continuous acquisition of location information of query points, which can easily lead to location leakage of query points. Existing k-anonymity and range query location protection methods do not fully consider the characteristics of road networks, and do not consider the update problem of location protection under continuous skyline query; traditional continuous skyline query of road network rarely considers data privacy leakage problem, and rarely considers the problem of adding differential privacy to the continuously updated result set; skyline query under traditional road network strictly queries according to the size of distance dimension, and does not consider the importance of non-spatial dimension when the difference of distance dimension is too small. Summary of the invention

[0004] In order to solve the above problems, the purpose of the present invention is to provide a method for processing location privacy approximate continuous skyline queries based on public places in a road network. Based on a large number of public places in the road network, the public safety approximate area hidden location information is constructed for querying users, and a new dominance, namely public safety approximate dominance, is further proposed to weaken the influence of the spatial distance dimension on the non-spatial dominance relationship, and quickly filter the pruned data set through the PWDG-Tree index and weak data points. The present invention can quickly obtain an approximate continuous skyline result that satisfies the protection of user location privacy.

[0005] The technical solution of the present invention is:

[0006] 1. A method for processing location privacy approximate continuous skyline queries based on public places in a road network. Given a query point q, a set of road network segments E, a set of data points P in a k-dimensional space on the road network, and a set of public place points P on the road network, r , the road network environment is abstracted as an undirected and weighted G = (V, E, W), where V is the vertex of the road in the road network, E is the set of roads in the road network, W is the weight of the road in the road network, that is, the distance of the road, d εTo construct the road network distance threshold of the public safety approximate area, λ is the public place point threshold for constructing the public safety approximate area. First, a PWDG-Tree index is constructed. Starting from the query point q, the public safety approximate area is first queried and constructed. The initial global road network safety approximate continuous skyline query result set is calculated. When the query point moves, if the query point reaches any safe end point, the public safety approximate area is updated. The data set is filtered based on the filtering algorithm of the PWDG-Tree index. The filtered data set is queried for the road network safety approximate continuous skyline, and the updated result set is returned. If the query point does not reach any safe end point, no update is required, and the previous query result is returned, including the following steps:

[0007] (1) Establish a PWDG-Tree index. The G-tree index is an efficient index for the road network environment. This paper adopts the idea of ​​G-tree for graph partitioning. In order to adapt to the public safety approximate area construction algorithm and public safety approximate dominated skyline query proposed in this paper and reduce the computational overhead, a new index structure PWDG-Tree index is proposed based on the G-tree index to construct the public safety approximate area and perform continuous skyline query to speed up the operation efficiency of the algorithm. The PWDG-Tree index structure stores the degree of nodes while storing boundary points. It stores the corresponding data points in the graph partition while storing the graph partition, and stores the corresponding weak data points. The data points are stored in layers according to the non-spatial dominated relationship. The PWDG-Tree index stores the distance result of the last query between each boundary point and the data point for fast filtering of the data set. The PWDG-Tree index structure stores the degree of each boundary point on the road network and the distance dbefore of the last continuous query from the last query point for each boundary point. Each graph partition stores the weak data points in the partition. The leaf nodes of the PWDG-Tree index structure store all road network nodes and all data points in the subgraph partition.

[0008] (2) Construction of public safety approximate area: Given a road network graph G and a set of public places P r , query point q, distance threshold d ε As well as the public place point threshold λ, starting from the query point q, the road network nodes of each path in all directions along the road network are traversed in a breadth-first manner to find all safe end points that meet the public safety approximate area, and the public safety approximate area is constructed through the found safe end points; all the nodes with a distance less than 3d from the query point q are screened. ε The public safety approximate dominated skyline query is performed on the filtered data points, and the non-skyline points in the public safety approximate area are pruned;

[0009] The safe endpoint can be a public point set P rA common point in the network can be a road network node or any point on the road network that satisfies the public safety approximate area.

[0010] The screening rules for the safety endpoint are as follows:

[0011] a. Starting from the query point q, a breadth-first search is performed along all directions of the road network. If there is a node whose number of visits is equal to the degree of this node, then this node is an internal node of the public safety approximate area, and the end points of all edges connected to this node will not be safe end points;

[0012] b. Starting from the query point q, perform a breadth-first search along all directions of the road network. If there are two endpoints s1 and s2, and s1 and s2 are directly searched from the same node v, and if s1 is on the path between s2 and v, then endpoint s1 is not a safe endpoint;

[0013] c. Starting from the query point q, perform a breadth-first search along all directions of the road network. If there are two safe endpoints s1 and s2, and s1 and s2 are endpoints directly searched from nodes v1 and v2, respectively, and if s1 is on the path between s2 and v2, and s2 is on the path between s1 and v1, then endpoints s1 and s2 are not safe endpoints;

[0014] d. Starting from the query point q, a breadth-first search is performed along all directions of the road network. If a node is visited multiple times and the number of visits is less than the degree of the node, the visit with the largest remaining road network distance is selected as the optimal visit. The endpoint generated by the optimal visit is the safe endpoint;

[0015] Starting from the query point q, perform a breadth-first traversal of the adjacent nodes and initialize the current remaining distance of the query point to d ε , the number of visits to the common point is 0, and the road network nodes are searched breadth-first starting from the query point q. Each time, it is determined whether the search distance is equal to d ε First, find a safe endpoint with the number of common points on the path equal to λ. If it does not exist, find an endpoint that satisfies the public safety approximate area in terms of distance. When visiting the next node v from the query point q, if v to v i The distance of node v is equal to the remaining distance of v and node v i If v is not visited, i is a safe endpoint, if v to v i The distance of node v is less than the remaining distance of node v. If the current search is not the optimal search, end the current search. Otherwise, move node v to the optimal search point. i Enter the breadth-first queue, regardless of the node v i Whether it is a safe endpoint, move the node v to v i All safe endpoints on the path are deleted; if v to v iThe distance of s is greater than the remaining distance of v, find the end point s, satisfying the distance from s to node v and the maximum remaining search distance of v, delete all safe end points on the search path of other end points. Finally, return the set of safe end points.

[0016] (3) Filter the data set based on the PWDG-Tree index. First, partition the graph G where the query point q does not exist in the region. i The weak data points in are marked as non-Skyline points of the road network safety approximate continuous Skyline query, and these weak data points are added to the queue PQ, which is the set of skyline points and non-Skyline points under the road network safety approximate continuous Skyline query screened in advance. Then, the skyline points and non-Skyline points in each partition are screened according to the proposed filtering method and added to the queue PQ. The non-Skyline points screened by the above operation are deleted from the data set P. Finally, the data points in the data set P that are dominated by the public safety approximation of the data points in the queue PQ are deleted and added to the queue PQ until there are no dominated data points.

[0017] The proposed filtering method is:

[0018] a. In a non-initial road network safe approximate continuous skyline query, there is a data point p∈P, which is a skyline point in the previous query. When the query point q and the public safety approximate area constructed by it are updated, if the road network distance between the query point q and the data point p is equal to the distance of the last query of the data point p minus the distance interval between the two queries, that is, dist(q,p)=p.dbefore-dist(q,q′), and point q′ is the last query point, then the data point p is still a skyline point of the road network safe approximate continuous skyline query;

[0019] b. In a non-initial road network safe approximate continuous skyline query, there is a data point p∈P, which is not a skyline point in the previous query. When the query point q and the public safety approximate area constructed by it are updated, if the road network distance between the query point q and the data point p is equal to the distance of the data point p in the previous query plus the distance interval between the two query points, that is, dist(q,p)=p.dbefore+dist(q,q′), and point q′ is the previous query point, then the data point p is still not a skyline point in the road network safe approximate continuous skyline query;

[0020] c. In the road network safety approximate continuous skyline query, there is a non-starting query point q that is not in the graph partition G. i In the road network space, query point q to the graph partition G iThe road network distance of all boundary points is equal to the distance of the last query minus the distance interval between the two query points, then the graph partition G i Among all the data points above, the data points whose last query was for the skyline point of the road network safety approximate continuous skyline query are still the skyline points;

[0021] d. In the road network safe approximate continuous skyline query, there is a non-starting query point q that is not in the graph partition G i In the road network space, query point q to the graph partition G i The road network distance of all boundary points is equal to the distance of the last query plus the distance interval between the two query points, then the graph partition G i Among all the data points above, the data points that were not skyline points of the road network safety approximate continuous skyline query in the last query are still not skyline points.

[0022] (4) Based on the public safety approximate control, the filtered data set is subjected to continuous skyline query to quickly obtain the approximate continuous skyline results that meet the requirements of protecting the user's location privacy.

[0023] 2. A method for processing location privacy approximate continuous skyline queries based on public places in a road network according to 1, characterized in that the public safety approximate area and the safety end point specifically include: given a road network graph G, a public place point set P r , query point q, distance threshold d ε and the common point threshold λ, starting from the query point q, the end point q of each path in each direction along the road network z One of the following two conditions is met:

[0024] 1) End point q z ∈P r , end point q z The road network distance to the query point q is less than d ε ; and query point q to end point q z There are at least λ public places on the shortest path;

[0025] 2) End point q z The road network distance to the query point q is equal to d ε ;

[0026] The closed area surrounded by all endpoints is the public safety approximate area of ​​the query point q. z It is the safe end point of the public safety approximate area. The result set of the public safety approximate dominated skyline in the public safety approximate area remains unchanged, which is the same as the result set of the query point q.

[0027] 3. A method for processing location privacy approximate continuous skyline queries based on public places in a road network according to 1, characterized in that the public safety approximate dominance specifically includes: given a road network distance threshold d for constructing a public safety area ε , given a query point q and data points p1 and p2 in a road network environment, d(q,p1) is used to represent the road network distance between the query point q and the data point p1, if one of the two conditions is met:

[0028] 1) p2 does not spatially dominate p1, that is And satisfy d(q,p1)-2d ε ≤d(q,p2);

[0029] 2) Satisfaction And d(q,p1)-2d ε <d(q,p2);

[0030] Then the public safety of p2 is said to approximately dominate p1, denoted by

[0031] 4. A method for processing location privacy approximate continuous skyline queries based on public places in a road network according to 1, characterized in that the weak data points specifically include: there is a data point p1∈P, and p1∈G i , G i is a subgraph of graph G, if for each graph G i The boundary point is used as the query point, and there is another data point p2, p2∈G i ,satisfy Then the data point p1 is called graph G i Weak data points.

[0032] Beneficial effects of the method of the present invention: Based on a large number of public places in the road network, the present invention constructs public safety approximate regional hidden location information for querying users, further proposes a new dominance, namely public safety approximate dominance, which weakens the influence of the spatial distance dimension on the non-spatial dominance relationship, and quickly filters the pruned data set through the PWDG-Tree index and weak data points. The present invention can quickly obtain an approximate continuous skyline result that satisfies the protection of user location privacy. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a flow chart of the steps of a method for processing location privacy approximate continuous skyline queries based on public places in a road network according to the present invention.

[0034] Figure 2 It is a schematic diagram of the approximate public safety area of ​​the present invention.

[0035] Figure 3It is a schematic diagram of the road network environment and points of interest information of the present invention.

[0036] Figure 4 It is a schematic diagram of the zoning diagram of the road network environment map of the present invention.

[0037] Figure 5 It is a schematic diagram of the PWDG-Tree index of the present invention. DETAILED DESCRIPTION

[0038] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in each embodiment of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0039] Reference Figure 1 , Figure 2 , Figure 3 , Figure 4 and Figure 5 The present invention provides a method for processing location privacy approximate continuous skyline queries based on public places in a road network, the method comprising the following steps:

[0040] 1. A method for processing location privacy approximate continuous skyline queries based on public places in a road network. Given a query point q, a set of road network segments E, a set of data points P in a k-dimensional space on the road network, and a set of public place points P on the road network, r , the road network environment is abstracted as an undirected and weighted G = (V, E, W), where V is the vertex of the road in the road network, E is the set of roads in the road network, W is the weight of the road in the road network, that is, the distance of the road, d ε is the road network distance threshold for constructing the public safety approximate area, λ is the public place point threshold for constructing the public safety approximate area, first construct the PWDG-Tree index, start from the query point q to query and construct the public safety approximate area for the first time, calculate the initial global road network safety approximate continuous skyline query result set, when the query point moves, if the query point reaches any safe end point, update the public safety approximate area, filter the data set based on the filtering algorithm of the PWDG-Tree index, perform a road network safety approximate continuous skyline query on the filtered data set, and return the updated result set. If the query point does not reach any safe end point, no update is required, and the previous query result is returned, such as Figure 3As shown in Figure 1, three dimensions including the spatial dimension are given for seven road network data points. The distance dimension represents their distance from the road network nodes. For example, the distance of data point p1 is (v1,1)(v3,2). The distance from data point p1 to road network node v1 is 1, and the distance to node v3 is 2. Their positions in the road network are shown in Figure 1. Figure 1 As shown in the figure, for the convenience of explanation, we first assume that query point q1 and query point q2 are two consecutive queries passing through the public safety approximate area, and construct the road network distance threshold d of the public safety area. ε is 1. The distance from query point q1 to node v9 is 1, and the distance from query point q2 to node v9 is 2. Figure 3 and Figure 4 As shown in the figure, the public safety approximate skyline dominance result set of the user at the query point q1 is {p1, p3, p4, p6}, and the non-skyline set is {p2, p5, p7}. When the user query point moves to q2, the public safety approximate area is updated. At the same time, the distance between the user and the data points p1, p3, p5, p8 increases, and the distance between the user and the data points p2, p4, p6, p7 decreases. The distance between the query point q2 and p3 is 9, recorded as d(q2, p3) = 9, and the distance between the query point q2 and p7 is 8, recorded as d(q2, p7) = 8. The data points p3 and p7 do not satisfy the road network dominance relationship, but satisfy d(q2, p3)-2d ε ≤d(q2,p7), so the public safety approximation of data point p3 dominates p7, and the public safety approximation skyline dominance result set of the user at query point q2 is {p1, p3, p4, p6}.

[0041] A method for processing location privacy approximate continuous skyline queries based on public places in a road network comprises the following steps:

[0042] (1) Establish a PWDG-Tree index. The G-tree index is an efficient index for the road network environment. This paper adopts the idea of ​​G-tree for graph partitioning. In order to adapt to the public safety approximate area construction algorithm and public safety approximate dominated skyline query proposed in this paper and reduce the computational overhead, a new index structure PWDG-Tree index is proposed based on the G-tree index to construct the public safety approximate area and perform continuous skyline query to speed up the operation efficiency of the algorithm. The PWDG-Tree index structure stores the degree of nodes while storing boundary points. It stores the corresponding data points in the graph partition while storing the graph partition, and stores the corresponding weak data points. The data points are stored in layers according to the non-spatial dominated relationship. The PWDG-Tree index stores the distance result of the last query between each boundary point and the data point for fast filtering of the data set. The PWDG-Tree index structure stores the degree of each boundary point on the road network and the distance dbefore of the last continuous query from the last query point for each boundary point. Each graph partition stores the weak data points in the partition. The leaf nodes of the PWDG-Tree index structure store all road network nodes and all data points in the subgraph partition.

[0043] like Figure 4 and Figure 5 As shown in the figure, partition G0 is the entire road network graph, and each graph in the lower layer is a sub-partition of the graph partition G0. Each graph corresponds to the distance matrix of the graph partition. The distance matrix of the non-leaf partition is the shortest distance matrix of the boundary points, and the distance matrix of the leaf partition is the shortest distance matrix between the boundary points and all the nodes in the graph partition and all the data points in the graph partition. The data points p1, p2, p3, p4, p5, p6, p7, p8 are in the graph partition G0, p1, p3, p5, p8 are in the graph partition G1, and so on; the boundary points of the graph partition G0 are v3, v4, v5, v9, v 11 , v 12 , v 13 , the boundary points of the graph partition G1 are v3, v9, v 11 , and so on; each graph partition stores the weak data points in the partition, for example, the weak data point of graph partition G1 is p8, and the leaf nodes of the PWDG-Tree index structure store all road network nodes in the subgraph partition and all data points in the partition.

[0044] (2) Construction of public safety approximate area: Given a road network graph G and a set of public places P r , query point q, distance threshold d εAs well as the public place point threshold λ, starting from the query point q, the road network nodes of each path in all directions along the road network are traversed in a breadth-first manner to find all safe end points that meet the public safety approximate area, and the public safety approximate area is constructed through the found safe end points; all the nodes with a distance less than 3d from the query point q are screened. ε The public safety approximate dominated skyline query is performed on the filtered data points, and the non-skyline points in the public safety approximate area are pruned;

[0045] The safe endpoint can be a public point set P r A common point in the network can be a road network node or any location point on the road network that meets the public safety approximate area.

[0046] The screening rules for the safety endpoint are as follows:

[0047] a. Starting from the query point q, a breadth-first search is performed along all directions of the road network. If there is a node whose number of visits is equal to the degree of this node, then this node is an internal node of the public safety approximate area, and the end points of all edges connected to this node will not be safe end points;

[0048] b. Starting from the query point q, perform a breadth-first search along all directions of the road network. If there are two endpoints s1 and s2, and s1 and s2 are directly searched from the same node v, and if s1 is on the path between s2 and v, then endpoint s1 is not a safe endpoint;

[0049] c. Starting from the query point q, perform a breadth-first search along all directions of the road network. If there are two safe endpoints s1 and s2, and s1 and s2 are endpoints directly searched from nodes v1 and v2, respectively, and if s1 is on the path between s2 and v2, and s2 is on the path between s1 and v1, then endpoints s1 and s2 are not safe endpoints;

[0050] d. Starting from the query point q, a breadth-first search is performed along all directions of the road network. If a node is visited multiple times and the number of visits is less than the degree of the node, the visit with the largest remaining road network distance is selected as the optimal visit. The endpoint generated by the optimal visit is the safe endpoint;

[0051] like Figure 2 As shown in (a), this is a distance threshold d ε And the common point threshold λ = 2, the safe area Safe1 constructed from the query point q, where p r1 , p r4 , p r5 is the point set P of public places r The common points in all satisfy the road network distance to the query point q is less than d ε; and there are at least λ public places on the shortest path to the query point q; points s1 and s2 are such that the distance from the query point q is d ε So Figure 2 The safe end point of the query point q shown in (a) is {p r1 , p r4 , p r5 , s1, s2}.

[0052] like Figure 2 As shown in (c), the end point s2 is the end point generated by searching the query point q through nodes v2, v3, and v5. The end point p r2 The endpoints s2 and p are generated by searching the query point q through nodes v2, v4, and v5. r2 Satisfy the public safety approximate area, end point p r2 On the path between node v5 and end point s2, end point p r2 It's not a safe destination. Figure 2 The safe endpoint of query point q in (c) is {s1, s2}.

[0053] like Figure 2 As shown in (b), the end point s4 is the end point generated by searching the query point q through nodes v2 and v4, and the end point s3 is the end point generated by searching the query point q through nodes v2, v3, and v5. The end point s4 is the end point directly searched by node v4, and the end point s3 is the end point directly searched by v5. The end point s4 is on the path between node v5 and the end point s3. The end point s4 is not a safe end point. Similarly, s3 is not a safe end point either. Figure 2 The safe endpoint of query point q in (b) is {s1, s2}.

[0054] Starting from the query point q, perform a breadth-first traversal of the adjacent nodes and initialize the current remaining distance of the query point to d ε , the number of visits to the common point is 0, and the road network nodes are searched breadth-first starting from the query point q. Each time, it is determined whether the search distance is equal to d ε First, find a safe endpoint with the number of common points on the path equal to λ. If it does not exist, find an endpoint that satisfies the public safety approximate area in terms of distance. When visiting the next node v from the query point q, if v to v i The distance of node v is equal to the remaining distance of v and node v i If v is not visited, i is a safe endpoint, if v to v i The distance of node v is less than the remaining distance of node v. If the current search is not the optimal search, end the current search. Otherwise, move node v to the optimal search point. i Enter the breadth-first queue, regardless of the node v i Whether it is a safe endpoint, move the node v to v iAll safe endpoints on the path are deleted; if v to v i The distance of s is greater than the remaining distance of v, find the end point s, satisfying the distance from s to node v and the maximum remaining search distance of v, delete all safe end points on the search path of other end points. Finally, return the set of safe end points.

[0055] (3) Filter the data set based on the PWDG-Tree index. First, partition the graph G where the query point q does not exist in the region. i The weak data points in are marked as non-Skyline points of the road network safety approximate continuous Skyline query, and these weak data points are added to the queue PQ, which is the set of skyline points and non-Skyline points under the road network safety approximate continuous Skyline query screened in advance. Then, the skyline points and non-Skyline points in each partition are screened according to the proposed filtering method and added to the queue PQ. The non-Skyline points screened by the above operation are deleted from the data set P. Finally, the data points in the data set P that are dominated by the public safety approximation of the data points in the queue PQ are deleted and added to the queue PQ until there are no dominated data points.

[0056] The proposed filtering method is:

[0057] a. In a non-initial road network safe approximate continuous skyline query, there is a data point p∈P, which is a skyline point in the previous query. When the query point q and the public safety approximate area constructed by it are updated, if the road network distance between the query point q and the data point p is equal to the distance of the last query of the data point p minus the distance interval between the two queries, that is, dist(q,p)=p.dbefore-dist(q,q′), and point q′ is the last query point, then the data point p is still a skyline point of the road network safe approximate continuous skyline query;

[0058] b. In a non-initial road network safe approximate continuous skyline query, there is a data point p∈P, which is not a skyline point in the previous query. When the query point q and the public safety approximate area constructed by it are updated, if the road network distance between the query point q and the data point p is equal to the distance of the data point p in the previous query plus the distance interval between the two query points, that is, dist(q,p)=p.dbefore+dist(q,q′), and point q′ is the previous query point, then the data point p is still not a skyline point in the road network safe approximate continuous skyline query;

[0059] c. In the road network safety approximate continuous skyline query, there is a non-starting query point q that is not in the graph partition G. iIn the road network space, query point q to the graph partition G i The road network distance of all boundary points is equal to the distance of the last query minus the distance interval between the two query points, then the graph partition G i Among all the data points above, the data points whose last query was for the skyline point of the road network safety approximate continuous skyline query are still the skyline points;

[0060] d. In the road network safe approximate continuous skyline query, there is a non-starting query point q that is not in the graph partition G i In the road network space, query point q to the graph partition G i The road network distance of all boundary points is equal to the distance of the last query plus the distance interval between the two query points, then the graph partition G i Among all the data points above, the data points that were not skyline points of the road network safety approximate continuous skyline query in the last query are still not skyline points.

[0061] (4) Based on the public safety approximate control, the filtered data set is subjected to continuous skyline query to quickly obtain the approximate continuous skyline results that meet the requirements of protecting the user's location privacy.

[0062] 2. A method for processing location privacy approximate continuous skyline queries based on public places in a road network according to 1, characterized in that the public safety approximate area and the safety end point specifically include: given a road network graph G, a public place point set P r , query point q, distance threshold d ε and the common point threshold λ, starting from the query point q, the end point q of each path in each direction along the road network z Meet one of the following two conditions:

[0063] 1) End point q z ∈P r , end point q z The road network distance to the query point q is less than d ε ; and query point q to end point q z There are at least λ public places on the shortest path;

[0064] 2) End point q z The road network distance to the query point q is equal to d ε ;

[0065] The closed area surrounded by all endpoints is the public safety approximate area of ​​the query point q. zIt is the safe end point of the public safety approximate area. The result set of the public safety approximate dominated skyline in the public safety approximate area remains unchanged, which is the same as the result set of the query point q.

[0066] 3. A method for processing location privacy approximate continuous skyline queries based on public places in a road network according to 1, characterized in that the public safety approximate dominance specifically includes: given a road network distance threshold d for constructing a public safety area ε , given a query point q and data points p1 and p2 in a road network environment, d(q,p1) is used to represent the road network distance between the query point q and the data point p1, if one of the two conditions is met:

[0067] 1) p2 does not spatially dominate p1, that is And satisfy d(q,p1)-2d ε ≤d(q,p2);

[0068] 2) Satisfaction And d(q,p1)-2d ε <d(q,p2);

[0069] Then the public safety of p2 is said to approximately dominate p1, denoted by

[0070] 4. A method for processing location privacy approximate continuous skyline queries based on public places in a road network according to 1, characterized in that the weak data points specifically include: there is a data point p1∈P, and p1∈G i , G i is a subgraph of graph G, if for each graph G i The boundary point is used as the query point, and there is another data point p2, p2∈G i ,satisfy Then the data point p1 is called graph G i Weak data points.

[0071] The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A method for processing location privacy approximate continuous skyline queries based on public places in a road network, characterized in that: Given a query point q, a set of road network segments E, a set of data points P in a k-dimensional space on the road network, and a set of public place points P on the road network r , the road network environment is abstracted as an undirected and weighted G = (V, E, W), where V is the vertex of the road in the road network, E is the set of roads in the road network, W is the weight of the road in the road network, that is, the distance of the road, d ε To construct the road network distance threshold of the public safety approximate area, λ is the public place point threshold for constructing the public safety approximate area. First, a PWDG-Tree index is constructed. Starting from the query point q, the public safety approximate area is first queried and constructed. The initial global road network safety approximate continuous skyline query result set is calculated. When the query point moves, if the query point reaches any safe end point, the public safety approximate area is updated. The data set is filtered based on the filtering algorithm of the PWDG-Tree index. The filtered data set is queried for the road network safety approximate continuous skyline, and the updated result set is returned. If the query point does not reach any safe end point, no update is required, and the previous query result is returned, including the following steps: (1) Establish a PWDG-Tree index. Based on the G-tree index, a new index structure PWDG-Tree index is proposed to construct public safety approximate areas and perform continuous skyline queries to speed up the operation efficiency of the algorithm. The PWDG-Tree index structure stores the degree of nodes while storing boundary points. It stores the corresponding data points in the graph partitions while storing graph partitions, and stores the corresponding weak data points. The data points are stored in layers according to non-spatial dominance relationships. The PWDG-Tree index stores the distance result of the last query between each boundary point and the data point for fast filtering of the data set. The PWDG-Tree index structure stores the degree of each boundary point on the road network and the distance dbefore of the last continuous query from the last query point for each boundary point. Each graph partition stores the weak data points in the partition. The leaf nodes of the PWDG-Tree index structure store all road network nodes and all data points in the subgraph partition. (2) Construction of public safety approximate area: Given a road network graph G and a set of public places P r , query point q, distance threshold d ε As well as the public place point threshold λ, starting from the query point q, the road network nodes of each path in all directions along the road network are traversed in a breadth-first manner to find all safe end points that meet the public safety approximate area, and the public safety approximate area is constructed through the found safe end points; all the nodes with a distance less than 3d from the query point q are screened. ε The public safety approximate dominated skyline query is performed on the filtered data points, and the non-skyline points in the public safety approximate area are pruned; (3) Filter the data set based on the PWDG-Tree index. First, partition the graph G where the query point q does not exist in the region. i The weak data points in are marked as non-Skyline points of the road network safety approximate continuous Skyline query, and these weak data points are added to the queue PQ, which is the set of skyline points and non-Skyline points under the road network safety approximate continuous Skyline query screened in advance. Then, the skyline points and non-Skyline points in each partition are screened according to the proposed filtering method and added to the queue PQ. The non-Skyline points screened by the above operation are deleted from the data set P. Finally, the data points in the data set P that are dominated by the public safety approximation of the data points in the queue PQ are deleted and added to the queue PQ until there are no dominated data points. The proposed filtering method is: a. In a non-initial road network safe approximate continuous skyline query, there is a data point p∈P, which is a skyline point in the previous query. When the query point q and the public safety approximate area constructed by it are updated, if the road network distance between the query point q and the data point p is equal to the distance of the last query of the data point p minus the distance interval between the two queries, that is, dist(q,p)=p.dbefore-dist(q,q′), and point q′ is the last query point, then the data point p is still a skyline point of the road network safe approximate continuous skyline query; b. In a non-initial road network safe approximate continuous skyline query, there is a data point p∈P, which is not a skyline point in the previous query. When the query point q and the public safety approximate area constructed by it are updated, if the road network distance between the query point q and the data point p is equal to the distance of the data point p in the previous query plus the distance interval between the two query points, that is, dist(q,p)=p.dbefore+dist(q,q′), and point q′ is the previous query point, then the data point p is still not a skyline point in the road network safe approximate continuous skyline query; c. In the road network safety approximate continuous skyline query, there is a non-starting query point q that is not in the graph partition G. i In the road network space, query point q to the graph partition G i The road network distance of all boundary points is equal to the distance of the last query minus the distance interval between the two query points, then the graph partition G i Among all the data points above, the data points whose last query was for the skyline point of the road network safety approximate continuous skyline query are still the skyline points; d. In the road network safe approximate continuous skyline query, there is a non-starting query point q that is not in the graph partition G i In the road network space, query point q to the graph partition G i The road network distance of all boundary points is equal to the distance of the last query plus the distance interval between the two query points, then the graph partition G i Among all the data points above, the data points that were not skyline points of the road network safety approximate continuous skyline query in the last query are still not skyline points. (4) Based on the public safety approximate dominance, the filtered data set is subjected to a continuous skyline query to quickly obtain an approximate continuous skyline result that satisfies the protection of user location privacy.

2. According to claim 1, a method for processing location privacy approximate continuous skyline queries based on public places in a road network is characterized in that: The public safety approximate area and the safety end point specifically include: given a road network graph G, a public place point set P r , query point q, distance threshold d ε and the common point threshold λ, starting from the query point q, the end point q of each path in each direction along the road network z Meet one of the following two conditions: 1) End point q z ∈P r , end point q z The road network distance to the query point q is less than d ε ; and query point q to end point q z There are at least λ public places on the shortest path; 2) End point q z The road network distance to the query point q is equal to d ε ; The closed area surrounded by all endpoints is the public safety approximate area of ​​the query point q. z It is the safe end point of the public safety approximate area. The result set of the public safety approximate dominated skyline in the public safety approximate area remains unchanged, which is the same as the result set of the query point q.

3. According to claim 1, a method for processing location privacy approximate continuous skyline queries based on public places in a road network is characterized in that: The public safety approximate control specifically includes: given a road network distance threshold d for constructing a public safety area ε , given a query point q and data points p1 and p2 in a road network environment, d(q,p1) is used to represent the road network distance between the query point q and the data point p1, if one of the two conditions is met: 1) p2 does not spatially dominate p1, that is And satisfy d(q,p1)-2d ε ≤d(q,p2); 2) Satisfaction And d(q,p1)-2d ε <d(q,p2); Then the public safety of p2 is said to approximately dominate p1, denoted by 4. According to claim 1, a method for processing location privacy approximate continuous skyline queries based on public places in a road network is characterized in that: The weak data points specifically include: there is a data point p1∈P, and p1∈G i , G i is a subgraph of graph G, if for each graph G i The boundary point is used as the query point, and there is another data point p2, p2∈G i ,satisfy Then the data point p1 is called graph G i Weak data points.

Images