CN106604222A - Mobile social network routing method based on spatial-temporal relation - Google Patents

Abstract

At present, a large number of routing algorithms conduct routing according to history records, and data is transmitted to a node with a maximal number of historical communication times with a destination node. The mobile social network routing method based on a spatial-temporal relation adopts an individual node moving rule as a starting point, and introduces a trip history model according to moving characteristics of an individual node. The spatial and temporal characteristics of human movement can be analyzed according to human activity traces stored in the trip history model, thus the next movement of a node can be predicted by adopting machine learning methods such as a decision-making tree. The mobile social network routing method guides routing based on the trip history model, and provides a routing algorithm STC for spatial-temporal relation in human society. A next-hop node is selected according to mobile destination prediction information of node (human) in a current time period and combined with historical access record of the node in a recent time period, and the data is always transmitted to a node which is most possibly communicated with the destination node. Experiments show that the algorithm effectively improves transmission performance of the network.

Description

A kind of mobile community network method for routing based on time-space relationship

Technical field

The invention belongs to wireless sensor network technology field, specifically a kind of route side for moving community network Method.

Background technology

As communication and the progress of science and technology, the function of handheld device are increasingly perfect, this allows for portable handheld device（Such as Mobile phone, PDA, notebook computer etc.）In generally there is a large amount of still undeveloped available resources and information, including wireless bandwidth, deposit Storage capacity, CPU and multi-medium data etc..Can be in communication with each other and be mutually unified into by these mobile communication equipments that people carries with For the mobile sensor network of a self-organizing, to reach the shared of data or information.From the net of this kind of mobile sensor network From the point of view of network characteristic, this kind of network is the one kind in delay-tolerant network, as in network, mobile sensor node has portability The characteristics of, it is referred to as pocket exchange network (Pocket Switched Network, PSN) in early days.And the movement of network node Feature is consistent with the mechanicses of people, generally with certain social characteristic, therefore the mobile community network that is otherwise known as（Mobile Social Network, MSN）.For MSN networks, network node is by people's carrying, interpersonal social relations and social row For regularity, define a metastable annexation, also some Social Characteristics with people of the node in network.

MSN networks, as a kind of new DTN networks, are the combination products of delay-tolerant network and interpersonal community network, Characteristic with both networks, the characteristics of the discontinuity of existing delay-tolerant network connects, has met the social behavior of people again Law characteristic.After network has the characteristic of intermittent communication, the route of data is no longer the via node continuously through multi-hop It is forwarded to destination node.As the connectedness of DTN nodes frequently changes with the movement of node, node for data forwarding Can only occur in when meet with other nodes.Therefore DTN networks are generally using storage forwarding（store-and - forward）Mode transmit data, i.e., when node periphery does not have neighbor node, node save the data in caching in move； After two nodes meet, both sides set up wireless session, complete route and the storage of data, subsequently forward the data to the next one The node for running into, until being sent to destination node.Additionally, the information exchange between user always might not be entered by the Internet OK.Found by the questionnaire survey to 70 people of certain laboratory in colleges and universities, about 50% email exchange is met daily at them Carry out in the middle of people, and the information that the Internet is provided also is not always able to meet the demand of user.Therefore it is contemplated that not Long future, numerous man-portable's equipment can dynamically connect into network, by the network, can pass through between men Handheld device mutually passes various available informations, mutually enjoys the available resources in equipment.

There is currently tens people for holding portable hand-held communications equipment, it is envisaged that using these with taking People with handheld communication devices will constitute mobile community network maximum in the world, make the social life and physical world of people It is real to blend together, if effectively utilizes, produce huge economic benefit.Meanwhile, with the popularization of smart mobile phone, more Mobile intelligent terminal equipment will be held come more people, according to the prediction of Mobile TeleSystems, set using intelligent terminal Standby number will obtain explosive growth.Mobile community network interior joint has the regularity of mankind's activity, by people's The research of mechanicses, reasonably using the mobility feature of node, designs the route technology for meeting application demand, so as to carry High router efficiency, with important theory and realistic meaning.

The content of the invention

The present invention adopts the following technical scheme that realization, a kind of method for routing based on space-time community, including following Step：

Set up navigation history model

The movement of people can be described with a series of shift motions including when and where information.Under normal conditions, in the mankind In motor process, destination and time are the objects most paid close attention to.Therefore, the GPS location of process is constantly recorded by handheld device, The shift motion of people can be expressed as departure time, departure place, time of stopping over, stop place, changes movement by us The record of the data such as direction time, moving direction, the time of advent and place of arrival, for the collection that someone shift motion is recorded Conjunction also just constitutes the navigation history of his activity.

Node motion is predicted

After the navigation history record of itself is established, the movement of node can be predicted using decision tree.Deposited by above Storage navigation history data in the handheld device, we can construct decision tree, as shown in Figure 1.For each leaf node, Select the probability of next one DestinationpIt is designated as：

(1)

In formula (1)It is in leaf nodePlaceThe quantity of class stroke,It is the quantity of all strokes of the node.As people sets out During motion, it is known that current time and the position being located, decision tree can be found by the value of Day, Time and Source to be had The leaf node of maximum of probability, and select the Destination to be initial predicted destination.

In the motor process of people, handheld device will at set intervals（Such as 5 minutes）The current position residing for itself of detection Put and whether be consistent with Midpoint or Destination towards prediction.If be not inconsistent, prediction destination can be recalculated general Rate：

(2)

In formula (2)It is to remove remaining stroke recording sum after all unreachable destinations.After this, handheld device will be selected The Destination or Midpoint represented by maximum of probability leaf node is selected as new prediction destination.

Statistics node meets table

We assume that each nodeV _i A table is preserved all, for memory nodeV _i Number of times and meet with meeting for other nodes The current destination information of forecasting of node, as shown in table 1.

The table that meets of 1 node of table

If people has found that in moving process mobile destination is not inconsistent with current predictive destination, or someone needs beginning next Individual stroke（Such as come home from work）, now destination need re-start prediction.If nodeV _i Prediction destination change, Then nodeV _i A particular message can be sent to all nodes in its neighbor scope, this message should include nodeiNumbering and new Destination's information of forecasting.Meet node and the message is broadcasted in a network in epidemic modes again, own in network Till node all receives the particular message.The node for receiving this particular message will perform following operation：

（1）If it was found that included and node in the table that meetsV _i Information of meeting, then adjustV _i Current destination；

（2）Otherwise directly increase in the table that meets of this nodeV _i Item, and both number of times that meets are calculated as 0.

Calculate intensity transmission

Make P_vioRepresent certain moment t nodes V_iTo point of destination V_oIntensity transmission, wherein intensity transmission represents the node and purpose The probability of point communication.Then

Wherein θ represents node V_iCurrent predictive point of destination to node V_oThe communication range of current predictive point of destination two cut The angle of line, as shown in Fig. 2 wherein blue straight line represents street.And m is the ratio of two variables, f therein_allRepresent the moment (such as 2 hours), node V in the special time period of t places_iAfter moving to current predictive point of destination, next step accesses other places Historical statistics total degree；And f_voRepresent node V_iAfter moving to current predictive point of destination, next step accesses V_oPrediction point of destination Historical statistics number of times.Therefore m represents V_iAfter moving to prediction point of destination, next step accesses point of destination V_oNumber of times account for total degree Percentage ratio.It is apparent from, m is bigger, node V_iEasier and V_oCommunication.

In addition, can be obtained according to the property and triangle formula of right angled triangle

θ=2arcsin(R/d) (5)

WhereinRThe position collection of point of destination is represented,dFor nodeV _i Prediction point of destination and nodeV _o Prediction point of destination between away from From.

It can be seen that, the intensity transmission of node is not changeless, and its size is over time with node present position not Change together.

Method for routing：

It is assumed that a certain nodeV _o Current prediction destination is close to convergent point, and can have and be in communication with each other with convergent point.We Or consider nodeV _i , it is assumed thattMoment nodeV _i There is data-message to need to be sent to destination nodeV _o , and node at presentV _i Neighbour Occupyneighbor(V _i )ComprisingZ’Individual other mobile nodes.NodeV _i Understood by handshake information respectively firstZ’Individual node it is pre- Point of destination is surveyed, if finding there is node and destination nodeV _o Co-located collection, then be directly sent to these sections data-message Point.These nodes are eventually messaging to point of destination.

If this node is can not find in neighbor scope, thenV _i Want to send messages to nodeV _o It may first have to find nodeV _o , that is, understand nodeV _o Situation of movement.In order to obtain nowV _o Mobile destination, nodeV _i First look at meeting for oneself List item, if the information of needs is can not find in the table that locally meets,V _i Then need to send messages to neighbor node, neighbor node is received Whether have in the table that meets that oneself is checked after corresponding messageV _o Mobile destination record.Because when arbitrary node destination predicts When changing, in network, all nodes can all receive the particular message that the node prediction destination changes, so passing through Check the table that meets of oneself and neighbours, nodeV _i Point of destination can be obtained with very big probabilityV _o Mobile destination.If still can not ObtainV _o Mobile destination information, then expand find scope, till finding the information of needs.

Obtain point of destination V_oPrediction destination after, V_iCan be V_oPrediction destination notify in neighbor scope own Node, these nodes will be according to V_oPrediction destination calculate oneself to V_oIntensity transmission.Then V_iBy handshake information Data-message is transmitted to neighbor (V after the intensity transmission of solution neighbor node_i) in arrive V_oTransmission probability compare P_vioBig institute There is node.

Description of the drawings

Fig. 1：Navigation history decision tree structure.Built by storage navigation history data in the handheld device.

Fig. 2：The calculating of transmission probability：Wherein R represents the position collection of point of destination, and d is node V_iPrediction point of destination and node V_o The distance between prediction point of destination.

Claims (8)

1. a kind of mobile community network method for routing based on time-space relationship, it is characterised in that：Comprise the following steps：

Active strokes history is constituted by the set that shift motion is recorded, and is temporally divided into two discrete data sets：Day and Time of Time, the Day set expression in units of week, time of the Time set expressions in units of day.

2. by navigation history data, we can construct decision tree, for each leaf node, calculate next target The probability on ground, for predicting the movement of node.

3. according to the table that meets stored in node, including meet number of times and the node that meets of certain node and other nodes is current Destination's information of forecasting, such as prediction destination change, then node sends a particular message to all sections in its neighbor scope Point, is met node and the message is broadcasted in a network in epidemic modes again, and in network, all of node all receives this Till particular message.

4. the node of this particular message is received according to information of forecasting, calculate the possibility that intensity transmission, i.e. node are communicated with point of destination Property.

5. a certain node is assumedV _o Current prediction destination is close to convergent point, and can have and be in communication with each other with convergent point, it is assumed thattMoment nodeV _i There is data-message to need to be sent to destination nodeV _o , and node at presentV _i Neighboursneighbor(V _i )ComprisingZ’ Individual other mobile nodes, nodeV _i Understood by handshake information respectively firstZ’The prediction point of destination of individual node, if finding there is section Point and destination nodeV _o Co-located collection, then be directly sent to these nodes data-message.

If this node is 6. can not find in neighbor scope, thenV _i Want to send messages to nodeV _o It may first have to find nodeV _o , that is, understand nodeV _o Situation of movement, nodeV _i The list item that meets of oneself is first looked at, if can not find in the table that locally meets The information of needs,V _i Then need to send messages to neighbor node, neighbor node checks the table that meets of oneself after receiving corresponding message In whether haveV _o Mobile destination record.

7., when the prediction of arbitrary node destination changes, in network, all nodes can all receive the node prediction destination and occur The particular message of change, so the table that meets by checking oneself and neighbours, nodeV _i Purpose can be obtained with very big probability PointV _o Mobile destination.

8. point of destination is obtainedV _o Prediction destination after,V _i Can handleV _o Prediction destination notify to all sections in neighbor scope Point, these nodes will basesV _o Prediction destination calculate oneself and arriveV _o Intensity transmission, thenV _i Understood by handshake information Data-message is transmitted to after the intensity transmission of neighbor nodeneighbor(V _i )In arriveV _o Transmission probability ratioP _vio Big is all Node.