CN109597866B

CN109597866B - Traffic rule-based continuous nearest neighbor monitoring method for moving object in road network

Info

Publication number: CN109597866B
Application number: CN201811396521.8A
Authority: CN
Inventors: 李红军
Original assignee: Chengdu Univeristy of Technology
Current assignee: Chengdu Univeristy of Technology
Priority date: 2018-11-22
Filing date: 2018-11-22
Publication date: 2020-05-19
Anticipated expiration: 2038-11-22
Also published as: CN109597866A

Abstract

The invention discloses a traffic rule-based method for monitoring continuous nearest neighbor of a moving object in a road network, which comprises the following steps: s1, establishing a road network model; s2, establishing an expansion tree of the query object q according to the road network model and initializing the expansion tree to obtain a k neighbor set of the query object q; and S3, updating the road network model according to the current position of the query object q and returning to the step S2, thereby realizing the monitoring of the continuous nearest neighbor of the mobile object in the road network. The latest k neighbor of the query object q is stored in the q.result, and when the k neighbor of the query object q is provided, the current query time t and a query object set { q } are used as parameters to call the method, so that the latest k neighbor q.result of the query object q at the time t can be obtained. The invention can continuously update the road network and the expansion tree, so that the method is completely applied to the monitoring of the continuous nearest neighbor of the mobile object in the road network.

Description

Traffic rule-based continuous nearest neighbor monitoring method for moving object in road network

Technical Field

The invention relates to the field of space-time query of moving objects, in particular to a method for monitoring continuous nearest neighbor of a moving object in a road network based on traffic rules.

Background

With the rapid development of mobile computing, a number of mobile object-based applications have emerged that require efficient mobile object spatiotemporal queries. The k-Nearest Neighbor (k-NN) query is a basic problem in the research field of mobile object databases, and is widely concerned by researchers.

Early studies were primarily directed to static data spaces, looking for current nearest neighbors. With the application of the method, a plurality of k-NN-based extension researches are generated, and continuous nearest neighbor query (continuous k-NN, abbreviated as CkNN) is one of the most important extensions. It means that the moving object moves continuously in space, and the query can return the latest k neighbor of the query object in real time.

Some continuous k nearest neighbor query methods proposed earlier only aim at euclidean distance space, and cannot be used for continuous k nearest neighbor query of a road network. Another part of the study is designed to handle continuous nearest neighbor queries of static objects, i.e. query objects remain static. However, these methods have problems that the moving environment characteristics of the road network and the speed and direction of the moving object are not combined, and the practicability is poor.

Disclosure of Invention

Aiming at the defects in the prior art, the method for monitoring the continuous nearest neighbor of the moving object in the road network based on the traffic rules solves the problem that the prior k nearest neighbor query method is poor in practicability in the road network.

In order to achieve the purpose of the invention, the invention adopts the technical scheme that:

a method for monitoring the continuous nearest neighbor of a moving object in a road network based on traffic rules is provided, which comprises the following steps:

s1, obtaining the mobile object and the query object q, establishing a road network model containing traffic rules, and obtaining a road network topological structure;

s2, establishing an expansion tree of the query object q according to the road network model and initializing the expansion tree to obtain a k neighbor set of the query object q;

s3, updating the road network model according to the current positions of the mobile object and the query object q and the road network topological structure;

and S4, judging whether the nearest neighbor of the query object q needs to be monitored, if so, returning to the step S2, otherwise, finishing monitoring and finishing monitoring the continuous nearest neighbor of the mobile object in the road network.

Further, the road network model comprises a road model T_edgeRoad intersection model T_nodeMoving object model T_objAnd query object model T_query；

Road model T_edgeIs a 6-tuple (rid,<n_s,n_e>,<e_i.w_p,e_i.w_n>,<s_pmax,s_nmax>,S_objil), where rid is the number of the road;<n_s,n_e>respectively a starting point and an end point of the road;<e_i.w_p,e_i.w_n>the weights of the forward direction and the reverse direction of the road are respectively;<s_pmax,s_nmax>the road positive direction and the road negative direction maximum speed limit are respectively set; s_objIs the current moving object set on the road; IL is a query list affected by a road, which is a triplet<q,n,d>N is the starting point n of the road_sOr end point n_eD is a number which indicates that the query object q is influenced by the road, and the influence section is the range of the road starting from the node n by d, namely, all the objects in the range of the road starting from n by d belong to the k neighbor set of the query object q;

road intersection model T_nodeIs a 2-tuple (contid, rules), wherein the contid represents the connection id; rules represents the set of traffic rules for the connection<RoadID_from,END_from,RoadID_to,END_to>；RoadID_fromAnd RoadID_toIndicating respectively the road id, END on which the moving object is going to leave and enter_fromIs a Boolean type data, and represents that the mobile object leaves the roadID_fromWhether it is at the end point n of the road_eEND if at the END of the road_fromIs 1, otherwise END_fromIs 0; END_toRoadID indicating the entry of a moving object into a road_toWhether it is at the end point n of the road_eEND if at the END of the road_toIs 1, otherwise at the start of the road, END_toIs 0;

moving object model T_objIs 5-tuple (oid, e)_j,t_u,dist,v_i) Wherein oid represents the number of the moving object; e.g. of the type_jIndicating a road on which the moving object is located; t is t_uIndicating the latest update time of the mobile object; dist indicates that the moving object is at t_uThe time and the edge e_jStart node n of_sThe distance dist therebetween; v. of_iIndicates that the moving object is at t_uThe moving speed at the moment is positive and indicates that the moving object moves from the starting point to the end point, and otherwise indicates that the moving object moves from the end point to the starting point;

querying an object model as T_queryIncluding information of the mobile object, a current kNN set q.result of the query object q, a distance q.kNN _ dist from the query object q to the kth nearest neighbor, and an extended tree T of the query object q_q。

Further, step S2 specifically includes the following steps:

s2-1, expanding the road network along the moving direction of the query object q by adopting Dijkstra algorithm from the query object q, and storing the minimum pile H of road intersections encountered in the expansion process_nodeAnd for depositing a set O of mobile objects encountered during the expansion process_encountInitialized to null, expanded tree of query object qSetting a root node as a point q, initializing a query result q.result, and simultaneously setting the shortest distance from the point q to all nodes in the road network as an infinite distance;

s2-2, acquiring the road connection v of the moving object q in the road moving direction according to the moving speed of the moving object q, and recording the influence of a section from a point q to a point v on the road on the query object q;

s2-3, inserting the obtained road connection v into the expansion tree and the minimum heap H_nodeRecording the shortest network distance from the q point to the v point;

s2-4, adding the moving object on the road qv into k neighbor of the query object q, and simultaneously leading the query object q to reach any moving object O_iIs inserted into the moving object set O by key_encountPerforming the following steps; wherein O is_iE is O, and O is a set of all moving objects on the road qv;

s2-5, judging whether the current k neighbor set of the query object q contains k objects, if so, obtaining the k neighbor set of the query object q and entering the step S3, otherwise, entering the step S2-6;

s2-6, from the smallest pile H_nodeExtracting the node n with the minimum key to obtain the node n in the expansion tree T_qThe road from the father node to the node n is deleted;

s2-7, sequentially selecting a division node n in the expansion tree T_qAll adjacent nodes n except the parent node in_adjAnd obtaining the k neighbor set of the query object q until whether the current k neighbor set of the query object q contains k objects.

Further, step S3 specifically includes the following steps:

s3-1, updating the moving object model T_obj: according to moving object O_iCurrently on a new road e_newUpdate of the mobile object O by the position information on_iIn a moving object model T_objThe corresponding information in (1);

s3-2, updating the road model T_edge: according to the updated moving object O_iId and moving object model T_objFinding a moving object containing O_iRoad e_jAnd from the roade_jDelete moving object e in the moving object list of (2)_jTo move an object e_jJoining a new road e_newUpdating the weight of each road in the moving object list;

s3-3, updating the query model T_query: according to the node n and all adjacent nodes n_adjUpdating k neighbor set q.result of query object q, network distance q.kNN _ dist from query object q to k neighbor and moving object set O by the formed moving object information on the road_encount。

Further, the moving object set O is updated in step S3-3_encountThe specific method comprises the following steps:

s3-3-1, collecting O from original moving object_encountDeleting the mobile objects which do not belong to the roads on the subtree taking the position of the current query object q as the root node;

s3-3-2, judging that the object is reserved in the moving object set O_encountIf the number of the moving objects is larger than or equal to k, the step S3-3-4 is executed, otherwise, the step S3-3-3 is executed;

s3-3-3, using the position of the current query object q as the root node, adopting the same method as the steps S2-1 to S2-4 to the mobile object set O with part of the mobile objects deleted_encountSupplementing;

s3-3-4, subtracting the moving distance of the query object from the distance from the reserved moving object to the position above the query object q to obtain the query object q at the current moment and an updated moving object set O_encountOf each moving object.

The invention has the beneficial effects that: the latest k neighbor of a query object q is stored in q.result, when the k neighbor of the query object q is provided, the current query time t and a query object set { q } are used as parameters to call the method, the latest k neighbor q.result of the query object q at the moment t is obtained, and the latest k neighbor q.result is directly returned. The invention updates the road network and the expansion tree at each query time point, so that the method is completely applied to the monitoring of the continuous nearest neighbor of the moving object in the road network, and is beneficial to a driver to select a passenger receiving sequence or a passenger receiving route.

Drawings

FIG. 1 is a schematic flow diagram of the present invention;

FIG. 2 is a schematic diagram of a road network including traffic rules and a 4NN query according to an embodiment;

FIG. 3 is a q-point based tree derived from FIG. 2;

FIG. 4 is a diagram illustrating an updated tree for 3 mobile objects in the embodiment;

FIG. 5 is a diagram of the expanded tree of FIG. 4 after the partial mark has been moved.

Detailed Description

The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined in the appended claims, and all matters produced by the invention using the inventive concept are protected.

As shown in fig. 1, the method for monitoring the continuous nearest neighbor of the moving object in the road network based on the traffic rules comprises the following steps:

The road network model comprises a road model T_edgeRoad intersection model T_nodeMoving object model T_objAnd query object model T_query；

moving object model T_objIs 5-tuple (oid, e)_j,t_u,dist,v_i) Wherein oid represents the number of the moving object; e.g. of the type_jIndicating a road on which the moving object is located;t_uindicating the latest update time of the mobile object; dist indicates that the moving object is at t_uThe time and the edge e_jStart node n of_sThe distance dist therebetween; v. of_iIndicates that the moving object is at t_uThe moving speed at the moment is positive and indicates that the moving object moves from the starting point to the end point, and otherwise indicates that the moving object moves from the end point to the starting point;

Step S2 specifically includes the following steps:

s2-1, expanding the road network along the moving direction of the query object q by adopting Dijkstra algorithm from the query object q, and storing the minimum pile H of road intersections encountered in the expansion process_nodeAnd for depositing a set O of mobile objects encountered during the expansion process_encountInitializing to be empty, setting an extended tree root node of a query object q as a q point, initializing a query result q.result, and simultaneously setting the shortest distance from the q point to all nodes in a road network as an infinite;

Step S3 specifically includes the following steps:

s3-2, updating the road model T_edge: according to the updated moving object O_iId and moving object model T_objFinding a moving object containing O_iRoad e_jAnd from road e_jDelete moving object e in the moving object list of (2)_jTo move an object e_jJoining a new road e_newUpdating the weight of each road in the moving object list;

Updating the moving object set O in step S3-3_encountThe specific method comprises the following steps:

In an embodiment of the present invention, fig. 2 is a schematic diagram of a road network and 4NN query, and fig. 3 is a schematic diagram of an extended tree based on fig. 2, where q.knn _ dist is 50, that is, a mobile object q arrives at a mobile object p₄The network distance of (a). Wherein p1-p6 are all moving objects, n₁-n₁₃All are nodes, and the node q is a leaf node, a dotted line and a node (n) issued by the dotted line of the expanded tree₃、n₅、n₆、n₈And n₁₂) Not belonging to the tree, and mobile objects p1, p2, p3, p4 points and p6 other than the node q do not belong to the tree, and the number of parentheses beside the node indicates the shortest network distance from the mobile object q to the node, and for n₁₁Lower leaf node representing road n₁₁n₁₂Upper distance n₁₁Division position of length 10. At the next query time of the mobile object q, it is assumed that the locations of 3 mobile objects p4, p5, and p6 are updated, and the update type includes that the mobile object is still in the extended tree T_qInternal, still expanding tree T_qOuter, outside-in and inside-out, the first two update types do not affect the result change, the latter two cause the k neighbor result of the moving object q to change, and fig. 4 shows the update (moving object p6 is moved to n)₁₂The direction is moved a certain distance but still outside the extended tree, and the moving object p4 moves to n₅The direction is moved by a certain distance and moves out of the range of the extended tree, and the moving object p5 moves from the road n₁₃n₁Enters the road n₁n₂) The subsequent moving object position. When the number of moving objects moving out of the expansion tree does not exceed the number of moving objects entering the expansion tree, the graphThe partial labels in 4 will move a certain distance towards the root node, resulting in the extended tree shown in fig. 5.

New location q from query object q_newIt can be divided into 2 types of updates:

type 1 (query object exit expansion tree) at this time q_newLocated outside the tree.

Type 2 (query object cruises in the extended tree) q at this time_newStill within the spanning tree.

In the mobile environment of fig. 2, the query object q has moved to the road n₅n₄The above. As can be seen from the expanded tree diagram 3 corresponding to FIG. 2, the road n₅n₄Is not in the extension tree range, so that the extension tree and O cannot be utilized_encountThe k neighbors of the current q are calculated by the middle information, and at this time, the following steps need to be executed: 1. deleting q from e.IL of all roads in the expanded tree; 2. reinitializing the k neighbors.

In the mobile environment of fig. 2, the query object q has moved to the road n₁n₇The above. At this time q_newAt branch n₁n₇In q_newAs a subtree of the root node to the current q_newThe k-neighbor computation of (a) is still useful and therefore can utilize this information. After the query object is updated, the road condition and O in the data model are required_encountAnd k neighbor results are maintained, and the main steps are as follows:

1. modifying a road model T_edge: after the query object is updated, the roads that originally affected the query may no longer affect the query, so the object q needs to be deleted from the e.il of these roads. Here, it is only necessary to find the nodes not belonging to q in the extended tree_newThose roads being subtrees of the root node, e.g. n in FIG. 3₁n₂、n₁n₁₁And so on, deleting q from e.il of these roads.

2. Maintenance of O_encount: here, it is required that: (1) from O_encountDeletion in not belonging to q_newIs the object of the road on the subtree of the root node. (2) Updating O_encountDistance of the remaining objectFrom, only the moving distance of q needs to be subtracted from the original distance.

3. Maintaining a query model T_query: as can be seen from the spanning tree, this time, q_newAll objects on the subtree which are root nodes still belong to q_newK is close to O through the pair_encountAfter maintenance, the updated information of these objects is stored in O_encountIn and O_encountThe target of k neighbor which originally does not belong to q is likely to become q_newK is close to and better than non-O_encountAnd (4) the target. The following steps are therefore performed at this point:

(1) mixing O with_encountAdding the first k objects into a new k neighbor;

(2) if there are not enough k, then pair is q_newFor extending the subtree of the root node, the information of the subtree needs to be updated first, i.e. q is modified_newDistances to nodes in a subtree, and then initializing H with leaf nodes of the subtree_nodeAnd expanding the tree to obtain an updated expanded tree, and continuously maintaining the O in the expanding process_encountAnd (4) object information.

When the weight of the road in the expanded tree changes, 2 cases are divided:

(1) and (5) processing the reduction of the road weight. For example, the road n in FIG. 2₁n₁₁The weight is decreased.

This update may cause a mobile object not belonging to the tree to enter the tree, e.g. road n₁₁n₁₂Of moving object p₆It is possible to enter the expansion tree to replace other k neighbors while T is in the expansion tree_qOther part of (i.e. divide by n)₁n₁₁And with n₁₁Subtree of root node), the distance and shortest path to q from them may change, but only T will change_qIn a subtree with node n as root, n must satisfy dist (q, n)>dist(q,n₁₁) Otherwise no change occurs. Thus, when the road n is₁n₁₁When the weight is reduced, the following 3 steps are completed:

a) from T_qDeleting a subtree having node n as root node, n satisfying dist (q, n)>dist(q,n₁₁) And deleting q from IL of the edge contained in the subtree and simultaneously deleting q from O_encountDelete the affected object;

b) update with n₁₁Keys of all objects and nodes on subtrees of the root node, and modifying O_encountCorresponding object information;

c) initializing H with leaf nodes of the current expansion tree_nodeAnd expanding the expansion tree.

(2) And (5) processing of increasing the road weight. For example, the road n in FIG. 2₁n₇The weight increases.

At this time, q reaches n₁n₇The distance of the moving object o above should be modified to the current distance plus the weight change value. In the extended tree at n₇The shortest path and distance of moving objects on all roads as the root node will have an effect (since the objects may arrive from q without passing n)₁And arrives through other nodes), so it is necessary:

a) delete n from the expanded tree₇A subtree which is a root node;

b) expanding the current expansion tree (initializing H by leaf nodes in the expansion tree)_nodeSimultaneously to O_encountMaintenance is performed).

In summary, the latest k neighbor of the query object q is stored in the q.result, and when the k neighbor of the query object q is provided, the current query time t and the query object set { q } are used as parameters to call the method, so as to obtain the latest k neighbor q.result of the query object q at the time t and directly return the latest k neighbor q.result. The method updates the road network and the expansion tree at each query time point, so that the method is completely applied to monitoring the continuous nearest neighbor of the mobile object in the road network.

Claims

1. A method for monitoring the continuous nearest neighbor of a moving object in a road network based on traffic rules is characterized in that: the method comprises the following steps:

s4, judging whether the nearest neighbor of the query object q needs to be monitored, if so, returning to the step S2, otherwise, finishing monitoring and finishing monitoring the continuous nearest neighbor of the mobile object in the road network;

the step S2 specifically includes the following steps:

S2-6. from the smallest pile H_nodeExtracting the node n with the minimum key to obtain the node n in the expansion tree T_qThe road from the father node to the node n is deleted;

2. The method according to claim 1, wherein the method comprises the following steps: the road network model comprises a road model T_edgeRoad intersection model T_nodeMoving object model T_objAnd query object model T_query；

The road model T_edgeIs a 6-tuple (rid,<n_s,n_e>,<e_i.w_p,e_i.w_n>,<s_pmax,s_nmax>,S_objil), where rid is the number of the road;<n_s,n_e>respectively a starting point and an end point of the road;<e_i.w_p,e_i.w_n>the weights of the forward direction and the reverse direction of the road are respectively;<s_pmax,s_nmax>the road positive direction and the road negative direction maximum speed limit are respectively set; s_objIs the current moving object set on the road; IL is a query list affected by roads, being a triplet<q,n,d>N is the starting point n of the road_sOr end point n_eD is a number which indicates that the query object q is influenced by the road, and the influence section is the range of the road starting from the node n by d, namely, all the objects in the range of the road starting from n by d belong to the k neighbor set of the query object q;

the road intersection model T_nodeIs a 2-tuple (contid, rules), wherein the contid represents the connection id; rules represents the set of traffic rules for the connection<RoadID_from,END_from,RoadID_to,END_to>；RoadID_fromAnd RoadID_toIndicating respectively the road id, END on which the moving object is going to leave and enter_fromIs a Boolean type data, and represents that the mobile object leaves the roadID_fromWhether it is at the end point n of the road_eEND if at the END of the road_fromIs 1, otherwise END_fromIs 0; END_toRoadID indicating the entry of a moving object into a road_toWhether it is at the end point n of the road_eEND if at the END of the road_toIs 1, otherwise at the start of the road, END_toIs 0;

the moving object model T_objIs 5-tuple (oid, e)_j,t_u,dist,v_i) Wherein oid represents the number of the moving object; e.g. of the type_jIndicating a road on which the moving object is located; t is t_uIndicating the latest update time of the mobile object; dist indicates that the moving object is at t_uThe time and the edge e_jStart node n of_sThe distance dist therebetween; v. of_iIndicates that the moving object is at t_uThe moving speed at the moment is positive and indicates that the moving object moves from the starting point to the end point, and otherwise indicates that the moving object moves from the end point to the starting point;

the query object model is T_queryIncluding information of the mobile object, a current kNN set q.result of the query object q, a distance q.kNN _ dist from the query object q to the kth nearest neighbor, and an extended tree T of the query object q_q。

3. The method for monitoring the moving object nearest neighbor in the road network based on the traffic rules as claimed in claim 1, wherein said step S3 specifically comprises the steps of:

4. The method for monitoring the nearest neighbor of the moving object in the road network based on traffic rules as claimed in claim 3, wherein said step S3-3 is to update the set of moving objects O_encountThe specific method comprises the following steps: