CN110891253B

CN110891253B - Community popularity-based vehicle-mounted delay tolerant network routing method

Info

Publication number: CN110891253B
Application number: CN201910973338.8A
Authority: CN
Inventors: 李致远; 彭二帅; 毕俊蕾; 张威威; 宋跃
Original assignee: Jiangsu University
Current assignee: Jiangsu University
Priority date: 2019-10-14
Filing date: 2019-10-14
Publication date: 2022-09-16
Anticipated expiration: 2039-10-14
Also published as: CN110891253A

Abstract

The invention discloses a vehicle-mounted delay tolerant network routing method based on the popularity of communities in communities and scenes to which nodes belong, which comprises four stages. Stage 1: the phase 1 is executed every T days to count the movement information of the nodes in the scene in the social region, and then the community popularity of each community in the scene is quantified by using the Google Pagerank ranking technology. In the invention, T is 30. And (2) stage: the source node forwards the message to the encountered relay node with probability P and updates the value of P phase 3: and the message carrier delivers the message to the target community. And (4) stage: messages are propagated within the target community. The ThrowBox of the target community forwards the stored messages to the nodes entering the community, and the messages brought into the community are transmitted in the community through a classical routing method spread & Wait in a Mobile Social Network (MSNs) under the control of the routing protocol in the invention.

Description

Community popularity-based vehicle-mounted delay tolerant network routing method

Technical Field

The invention belongs to the field of vehicle-mounted delay tolerant network application, and particularly relates to a community popularity-based vehicle-mounted delay tolerant network routing method.

Background

The internet of vehicles is a first item of falling to the internet of things industry, and is increasingly concerned by academic circles and industrial circles. Currently, the major application services of the internet of vehicles include vehicle active security services and user entertainment experience services. The vehicle active safety service means that after an accident occurs on a road section in front, the accident situation is collected by a vehicle in front through an audio and video acquisition device, and then data forwarding is performed on rear vehicle nodes through a communication module of the internet of vehicles, so that an automobile driver is reminded of abnormal situations in front in an active mode, and further, a larger traffic accident is avoided. The user entertainment experience service is an emerging vehicle networking service, which means that a mobile ad hoc network is formed between vehicles in an ad hoc mode, and then short audio and short video streams are shared between the vehicles, so that fun is added to users in the process of traveling. However, the two core services in the car networking have high requirements on data transmission capability, and therefore, it is necessary to research a data transmission method in the car networking environment.

According to the distribution and road conditions of Vehicle nodes, the Vehicle networking can be divided into a Vehicular Ad Hoc network (VANETs) with all-connected nodes and a Vehicular Delay Tolerant Network (VDTNs) with intermittent communication among nodes, so that the current Vehicle networking research focuses on efficient data forwarding in two environments of the VANETs and the VDTNs. Because the connection between nodes in the VDTNs is intermittent and opportunistic, the data forwarding conditions of the VDTNs are stricter and more challenging compared with those of VANETs. Summarizing, existing internet of vehicles routing and data forwarding algorithms can be classified into 3 types: a connection awareness based routing algorithm, a geographical area based routing algorithm Geocast and a map navigation based routing algorithm. The routing algorithm based on connection perception mainly calculates the connection probability between nodes according to the distance between a source node and a destination node, the propagation power, the transmission ratio and the vehicle density, estimates the probability of successful reception of a data packet according to the connection probability, and then selects a node with high connection rate as a relay node. The method is simple to implement and high in efficiency. The disadvantage is that they are all for the fully connected covered VANETs network, and the routing mode adopted is store-forward, which is not suitable for the opportunistic network of VDTNs. The geographical area based routing algorithm Geocast is a multicast data forwarding protocol for transmitting data to all nodes in a specific area, and is mainly suitable for geographical area broadcast applications. However, the Geocast protocol is not suitable for the environment in which nodes in a Geocast group are sparse, and therefore, the existing research on the Geocast routing is still in the VANETs environment. The map navigation-based routing algorithm is characterized in that a GPS system is started when a vehicle runs for a long distance, and the vehicles running to the same or similar target areas are combined into a stable connected dominating set according to the characteristic of electronic map routing, so that the success rate of data forwarding is improved. The method has the defects that the method is excessively dependent on a GPS navigation system, so that the service cannot be obtained for vehicles which do not use the GPS navigation system or do not use the GPS navigation system of the same manufacturer, the sparsity of vehicle nodes in the vehicle networking environment is further reduced, the node connection intermittency and the opportunity degree are enhanced, and the unpredictability is increased.

Disclosure of Invention

At present, most of routing and data forwarding methods of the internet of vehicles are based on a VANETs environment with high communication coverage, and the existing methods are applied to a VDTNs network with intermittent and opportunistic node connection, so that the data forwarding efficiency is reduced; data transmission delay, delay jitter and packet loss rate are increased; the stability of data transmission becomes weak.

In order to solve the technical problems, the invention provides a vehicle-mounted delay tolerant network routing method based on the popularity of communities in the communities and scenes to which nodes belong, which adopts a storage-carrying-forwarding routing strategy based on the popularity of the communities in the communities and scenes to which the nodes belong according to the characteristics of VDTNs environment, and the specific technical scheme is as follows:

a vehicle-mounted delay tolerant network routing method based on the community to which a node belongs and the popularity of the community in a scene comprises the following stages and steps:

stage 1: route data preparation phase. Phase 1 is the data basis on which the

subsequent phases

2, 3, 4 can successfully execute. The phase 1 is executed every T days to count the movement information of the nodes in the scene in the social region, and then the community popularity of each community in the scene is quantified by using the Google Pagerank ranking technology. In the invention, T is 30.

And (2) stage: the source node forwards the message to the encountered relay node with the probability P and updates the value of P. The phase 2 is performed as follows:

step 1: if the encountered relay node is the destination node, the source node directly sends the message to the destination node, and the routing algorithm is finished.

And if the relay node encountered in the step 1 is not the destination node, executing the step 2.

Step 2: and if the destination node is the node newly added into the scene, executing the step 2-1, otherwise executing the steps 2-2 and 2-3.

Step 2-1: obtaining the latest GPS information of the destination node through the SAU, and recording the latest GPS information as GPS _last . According to GPS _last And the community heat of the communities nearby the node, and the target community of the predicted node, which is marked as CMU _expect And querying for incoming community CMU _expect And the road section of (1) and GPS _last Adjacent road sections, denoted RD _expect . And step 2 is finished, and step 3 is executed.

Step 2-2: inquiring the community to which the destination node belongs, and recording as CMU _home And querying for incoming community CMU _home Main segment of (D), denoted as RD _{home_main} Choose to reach the RD from the current road segment _{home_main} Path of highest probability, denoted as PH _{max_prob} The corresponding probability is noted as

Step 2-3: according to step 2-2

And updating the probability P of the source node of the message for forwarding the message next time, and finishing the phase 2.

And 3, step 3: selecting to reach RD from current road segment _expect The shortest time-consuming path of (1), denoted as PH _fast ；

And 4, step 4: calculating the specific path PH calculated in step 3 _fast To RD _last Probability of (2), is recorded as

And 5: according to step 4

And (3) stage: the message carrier (relay node) delivers the message to the target community. The relay node carrying the message will pass through the target community mentioned in the stage 2 (CMU in the stage 2, step 2-1) with a certain probability _expect Or CMU in step 2-2 _home ). The message is delivered to the ThrowBox of the target community, which is a type of ancillary message storage facility placed at the roadside.

And (4) stage: messages are propagated within the target community. The ThrowBox of the target community forwards the stored messages to the nodes entering the community, and the messages brought into the community are transmitted in the community through a classical routing method spread & Wait in a Mobile Social Network (MSNs) under the control of the routing protocol in the invention. These messages will be destroyed after their life cycle has arrived, and a successful route formation will occur if the destination node returns to the destination community and receives the message before it is destroyed.

Further, the method for quantifying the community popularity of each community in the scene by using the Pagerank technology of Google through long-term statistics on the movement information of the nodes in the scene among the communities in the phase 1 (preparation period) comprises the following steps:

step 1: recording a large amount of historical activity information of all nodes in the scene.

Step 2: obtaining a direct transition probability matrix among communities in the scene from the preparation step 1, and recording the direct transition probability matrix as M _{trans_prob} 。

And step 3: setting an initial value of community popularity of each community in a scene to 1, i.e.

And 4, step 4: utilizing the related iterative formula of Pagerank of Google in combination with the direct transition probability matrix M in the preliminary step 1 _{trans_prob} And obtaining the community heat value after one iteration.

And 5: and repeating the preparation step 4 until the community heat value of each community in the scene is converged.

Then, the converged community popularity value obtained in step 5 is the final value of the community popularity of each community in the scene, and the preparation period is ended.

Further, step 2-1 of phase 2 is based on GPS _last Community CMU of destination of community heat prediction node of its nearby community _expect The method comprises the following steps:

step 2-1-1: searching for a location GPS _last Nearby communities and forming a community set U by the communities;

step 2-1-2: defining functions

Where i belongs to U, i is an arbitrary community, Dist (i) is community i and GPS _last Euclidean distance of.

Step 2-1-3: then the target community

Wherein hot _i Is the heat of community i.

Further, in stage 2, the selection of step 3 reaches RD from the current road section _expect The shortest time path PH _fast The method comprises the following steps:

step 3-1: loading the average direct transition time matrix M between segments in the scenario obtained in routing Algorithm stage 1 _{avg_time} ；

Step 3-2: in matrix M _{avg_time} The upper running single source shortest Path algorithm SPFA (short Path fast Algorithm) will get a Path from the current Path section to RD _expect The single-source shortest path of (2), then, the single-source shortest path is from the current road section to the RD _expect The shortest time path PH _fast ；

Further, calculating PH as described in step 3 of stage 2 _fast To RD _last Probability of (2)

The method comprises the following steps:

suppose a path PH made up of links _fast ＝(RD ₁ ,...,RD _k ,...,RD _m )，PH _fast To a first order matrix, then the probability of the path is

Wherein M is _{trans_prob} Is a direct transition probability matrix between the segments in the scenario obtained in the routing algorithm stage 1;

further, the selection of phase 2, step 2-2, reaches RD from the current road segment _{home_main} Path PH with the highest probability _{max_prob} The method comprises the following steps:

step 2-2-1: loading direct transition probability matrix M between road sections in scene obtained based on mass statistics _{trans_prob} 。

Step 2-2-2: defining slave links RD _x Arrival section RD _y A path PH of length m _u ＝(RD ₁ ,...,RD _k ,...,RD _m ) Has a path probability of

Wherein RD ₁ ＝RD _x ， RD _m ＝RD _y 。

Step 2-2-3: due to the matrix M _{trans_prob} The elements of (b) represent direct transition probabilities between the segments, so all element values are in the interval [0, 1). Due to the particularity of the values of the elements, in the matrix M _{trans_prob} The idea of utilizing single-source shortest path algorithm SPFA can obtain a path from the current road section to the RD _{home_main} The path with the maximum probability is the path from the current road section to the RD _expect Path PH with the highest probability _{max_prob} 。

Further, step 5 in stage 2 according to

The method for updating the probability P of the next message forwarding of the source node of the message comprises the following steps

Further, steps 2-3 in stage 2 according to

The invention has the beneficial effects that:

(1) the delivery rate of the data packet is higher

The routing method of the vehicle-mounted delay tolerant network is based on communities to which nodes belong and the heat degree of the communities in a scene, benefits from a large amount of statistical information about the scene obtained in an algorithm preparation period, and comprises the community heat degree of the communities in the scene, an average direct transition time matrix among road segments and a direct transition probability matrix among road segments, and can always deliver messages to the communities where the nodes are most likely to appear. Meanwhile, due to the excellent performance of the classical routing method Spray & Wait in MSNs in the community, the method can always maintain a high successful message delivery rate.

(2) Reasonably controls the number of message copies

When other conditions are the same, the successful delivery rate of the message is improved along with the increase of the number of the message copies, and correspondingly, the increase of the number of the message copies also increases the network overhead. In order to reasonably control the number of copies and further reduce the network overhead under the condition of ensuring higher successful delivery rate of the message, the routing method adopts a mode of forwarding the message to the relay node according to probability. The initial value of the probability is 1, the probability is continuously reduced in the process of forwarding the message, and the reduction amplitude of the probability is not blindly specified but is obtained according to the data of the long-term statistics of a plurality of information of the scene in the preparation period of the algorithm. Therefore, the algorithm of the invention comprehensively considers the message delivery success rate and the network overhead and reasonably controls the number of the message copies.

(3) The algorithm has low real-time calculation overhead

The main node in the on-board delay tolerant network environment is a vehicle, which has the characteristic of moving at high speed. Thus, the time for each communication between nodes in the on-board delay tolerant network may be very short. This requires that the real-time computation overhead of the corresponding routing algorithm should be as low as possible and the speed of the correlation computation should be as fast as possible.

In the invention, a large amount of calculation is completed in the preparation period of the routing algorithm. During the run-time of the routing algorithm, the main computational overhead is concentrated at step 2-1 of stage 2 (for newly added nodes) or at step 5 of stage 2 (for nodes known to the community to which it belongs). Step 2-1 and step 5 of stage 2 are both operated on the direct transition matrix based on road section-road section by using the single-source shortest path SPFA algorithm. Suppose the number of links in a scene is N, and the number of links that a link can normally directly transit to reach can be considered to be 6 or less (in the case of assuming that both ends of the link are connected at a 4-way intersection). Therefore, the graph theory model G ═ (V, E) based on the link-link direct transition matrix is inevitably a sparse graph, the number of vertices of the graph is | V | ═ N, and the number of edges | E | ≦ 6N can be considered. The time complexity of the SPFA algorithm is O (k. E. I.) O (k.6N) ═ O (k.N), where k is usually 2. Therefore, the routing algorithm of the present invention can be considered as a linear time-complex algorithm in operation, and in reality, the number of links N in a scene (usually a city) can be considered to be less than 10 ⁴ For the computational performance of present mobile nodes, the execution of the algorithm can be done in the microsecond range. Therefore, the routing algorithm in the present invention has very low real-time computation overhead.

(4) High robustness of algorithm

The statistical information related in the routing algorithm in the invention comprises community heat in a scene, an average direct transition time matrix among road sections and a direct transition probability matrix among road sections, which are all based on the long-term statistical result of the mobile information of all nodes in the scene. Therefore, the statistical result has multiple information sources and large information amount, the statistical result cannot fluctuate due to the abnormal behavior of the individual, and the statistical result is stable and reliable. This provides a good data base for the running of the routing algorithm after the preparation period. Since the movement of the nodes in the vehicle-mounted delay tolerant network is essentially human production activity at the present stage, the behavior of the nodes in the newly added scene can be considered to be in accordance with the relevant statistics of the algorithm preparation period, so that the problem of routing algorithm failure caused by lack of historical information of the nodes in the newly added scene is solved.

Drawings

FIG. 1 is a schematic diagram of a successful route;

FIG. 2 is a schematic diagram of probability transitions between road segments in a scene;

FIG. 3 is a flow chart of a calculation for calculating community popularity using the Google Pagerank concept;

fig. 4 is a schematic flow chart of the phase 2 algorithm.

Detailed Description

The invention will be further explained with reference to the drawings.

For convenience of description, table 1 is an explanation of terms of related nouns to which the invention relates.

TABLE 1

The routing method provides that only the source node of the message can forward the message, and the relay node of the message can only carry the message. Secondly, when the source node of a message encounters a relay node, the message is not necessarily forwarded, but is forwarded with a certain probability P. In the routing method, four stages are included.

Stage 1: route data preparation phase. Phase 1 is the data basis on which the

subsequent phases

And (2) stage: the algorithm flow of stage 2 as shown in fig. 4 includes the following:

the source node forwards the message to the encountered relay node with probability P and updates the value of P. The execution steps of this stage are:

step 1: if the encountered relay node is the destination node, the source node directly sends the message to the destination node, and the routing algorithm is finished. And if the encountered relay node is not the destination node, executing the step 2.

Step 2: if the target node is a node newly added to the scene, executing the step 2-1; otherwise, step 2-2 and step 2-3 are executed.

Step 2-1: obtaining the latest GPS information of the destination node through the SAU, and recording the latest GPS information as GPS _last . According to GPS _last And the community heat of the adjacent communities, the target community of the prediction node, marked as CMU _expect And querying for incoming community CMU _expect And the road section of (1) and GPS _last Adjacent road sections, denoted RD _expect . And step 2 is finished, and step 3 is executed.

Step 2-2: inquiring the community to which the destination node belongs, and recording as CMU _home And querying for incoming community CMU _home Main segment of (D), denoted as RD _{home_main} Selection of a route from a current road segment to a RD according to the rules of the claims section of the present application _{home_main} Path of highest probability, denoted as PH _{max_prob} The corresponding probability is noted as

Step 2-3: according to step 2-2

The probability P of the next forwarding of the message by the source node of the message is updated according to the rules described in the section of the claims of the present invention, and phase 2 ends.

And step 3: the selection of the arrival of the RD from the current road segment is made according to the detailed rules described in the claims section of the present application _expect The shortest time-consuming path of (1), denoted as PH _fast ；

And 5: according to step 4

And (3) stage: and the message carrier delivers the message to the target community. The nodes carrying the message will pass through the target community mentioned in stage 2 (CMU in stage 2, step 2-1) with a certain probability _expect Or CMU in step 2-2 _home ). If so, the message is delivered to the ThrowBox (a device placed on the roadside that assists in storing messages) of the target community.

Fig. 1 is a schematic diagram of a successful route, and the present invention has been described in detail above with respect to the stages and steps involved in a routing process.

FIG. 2 is a schematic diagram of direct transition probabilities between road segments in a scene. Due to the presence of the intersections, the roads in the scene are separated into individual road sections by the corresponding intersections, such as R in fig. 2 ₁ ,R ₂ ,...,R ₁₃ . Each route section being transferable directly, i.e. only once, and the number of route sections reached being limited, e.g. R ₁ Capable of directly transferring to the road sectionOnly R ₂ ,R ₁₀ ，R ₅ The number of road sections that can be directly reached by the transfer is large, e.g. R ₄ ,R ₁₀ ,R ₁₁ ,R ₆ ,R ₁₂ ,R ₁₃ (ii) a For example, in FIG. 2, the road segments that can be directly reached by the R7 are R6, R8, R10 and R11, F ₇ 6,8,10, 11; the road sections which can be directly transferred to the road section R8 are R6, R7, R11, R13, then T ₈ ＝{6,7,11,13}。

Suppose that for road segment R ₁ Long-term statistics of directly diverted arriving road segments results from R ₁ To R ₂ ,R ₁₀ The number of transfers is N _1,2 ,N _1,10 Then, in FIG. 2

More generally, direct transition of elements in a probability matrix

Where V is the set of road segments to which a road segment i can be directly transferred, N _i,k Indicating the number of times the link i is directly transferred to link k during long-term statistics.

FIG. 3 is a flow chart of a calculation for calculating community popularity using the Google Pagerank concept. HOT representing community popularity _i The initial values of (1 ≦ i ≦ M) are all 1, where M is the number of communities in the scene. By iterative calculation, HOT _i (1 ≦ i ≦ M) eventually converging to a set of values, and this set of values is the community heat in the desired scene.

The routing method reasonably controls the number of message copies on the premise of ensuring high data packet delivery rate, and has low algorithm real-time calculation cost and strong algorithm robustness.

The above-listed detailed description is only a specific description of a possible embodiment of the present invention, and they are not intended to limit the scope of the present invention, and equivalent embodiments or modifications made without departing from the technical spirit of the present invention should be included in the scope of the present invention.

Claims

1. A routing method of a vehicle-mounted delay tolerant network based on community popularity is characterized by comprising the following four stages:

stage 1: a routing data preparation stage; quantifying the community popularity of each community in the scene by using a ranking technology; suppose there are M communities, HOT, in total throughout the scene _i Representing the community heat degree with the number of i, wherein i belongs to M; the HOT needs to be determined at this stage _i So that the numerical value can accurately describe the real community heat, specifically:

step 1: counting direct transfer conditions among road sections in a scene in a counting period T days; the statistical information includes

a) Road segment set F that road segment p can directly reach by transferring _p Wherein p ∈ [1, N ]]N is the total number of road sections;

b) road segment set T capable of directly transferring to road segment q _q ；

c) The total times of directly transferring any road section i to other road sections is strans _i ；

d) The number of times of direct transfer of any link i to link j is trans _ij Where j ∈ F _j ；

And 2, step: utilizing the statistic in step 1 to obtain direct probability transition matrix M between road sections _{trans_prob} And average direct transfer time matrix M _{avg_time} ：

Wherein the content of the first and second substances,

wherein the content of the first and second substances,

and step 3: order to represent community heatHOT of _i All the initial values of (1);

and 4, step 4: the HOT is updated using the following formula _i The value of (a) is,

where d is a constant set to ensure iteration convergence, and d is 0.85; HOT in the above formula _i Is rewritten as

Will HOT _j Is rewritten as

Preparing for step 5;

and 5: judging whether the vector HOT converges or not, wherein the judgment standard is

Whether the error value is smaller than a preset error value 1 e-3; if the convergence rate is less than the preset convergence rate, considering the convergence, otherwise, considering the convergence rate as not-converged; if not, turning to the step 4, otherwise, executing the step 6;

and 6: converged HOT obtained in step 5 _i (i is more than or equal to 1 and less than or equal to M) is the heat of the community i;

and (2) stage: the source node forwards the message to the encountered relay node according to the probability P and updates the value of P; the specific steps at this stage are:

step 1: judging whether the encountered relay node is a destination node:

if the relay node is the destination node, the source node directly sends the message to the destination node, and the routing algorithm is finished; otherwise, executing step 2;

step 2: judging whether the destination node is a node newly added to the scene;

if the target node is a node newly added to the scene, executing the step 2-1; otherwise, executing the step 2-2 and the step 2-3;

step 2-1: obtaining eyes by SAUThe latest GPS information of the nodes is recorded as GPS _last And the road section where the node is located is RD _last (ii) a According to GPS _last And the community heat of the communities nearby the node, and the target community of the predicted node, which is marked as CMU _expect And querying for incoming community CMU _expect In road section with GPS _last Adjacent road sections, denoted RD _expect And step 2 is finished, and step 3 is entered;

step 2-2: inquiring the community to which the destination node belongs, and recording as CMU _home And querying for incoming community CMU _home Main section of (D), noted RD _{home_main} Choose to reach RD from the current road segment _{home_main} Path of highest probability, denoted as PH _{max_prob} The corresponding probability is noted as

Step 2-3: according to step 2-2

Updating the probability P of the next message forwarding of the source node of the message, and ending the stage 2;

and step 3: selecting to reach RD from current road segment _expect The shortest time-consuming path of (1), denoted as PH _fast ；

And 4, step 4: calculating the specific path PH calculated in step 3 _fast To RD _last Probability of is noted as

And 5: according to step 3

and (3) stage: the message carrier delivers the message to the target community:

the relay node carrying the message has a certain probability to pass through the procedure mentioned in stage 2And target community, i.e. CMU in step 2-1 of phase 2 _expect Or CMU in step 2-2 _home (ii) a If so, delivering the message to the ThrowBox of the target community;

and (4) stage: messages travel within the target community:

the ThrowBox of the target community forwards the stored messages to the nodes entering the community, the messages brought into the community are transmitted in the community through a routing method spread & Wait of the mobile social network under the control of a routing protocol, the messages are destroyed after the life cycle of the messages arrives, and if the destination nodes return to the target community and receive the messages before the messages are destroyed, a successful route is formed.

2. The community popularity-based vehicular delay tolerant network routing method according to claim 1, wherein the GPS-based routing method is adopted in step 2-1 of the phase 2 _last Community CMU of destination of community heat prediction node of its nearby community _expect The method comprises the following steps:

step 2-1-2: defining functions

Indicating communities i and GPS _last Euclidean distance between;

step 2-1-3: target community

3. The community heat based routing method for the vehicle-mounted delay tolerant network, according to claim 1, wherein the selection of step 3 in the phase 2 is from the current road section to the RD _expect The shortest time path PH _fast The method comprises the following steps:

step 3-1: loading into the scenario obtained in routing Algorithm stage 1Mean direct transition time matrix M between segments _{avg_time} ；

Step 3-2: in matrix M _{avg_time} The SPFA running on the single-source shortest path algorithm obtains a path from the current road section to the RD _expect The single-source shortest path of (2), then, the single-source shortest path is from the current road section to the RD _expect The shortest time path PH _fast 。

4. The community heat-based vehicle-mounted delay tolerant network routing method according to claim 1, wherein the specific calculated path PH obtained in the step 4 in the stage 2 _fast To RD _last Probability of (2)

The method comprises the following steps:

suppose a path PH made up of links _fast ＝(RD ₁ ,...,RD _k ,...,RD _m ) Then the probability of the path is

Wherein M is _{trans_prob} Is a direct transition probability matrix between segments in the scenario obtained at routing algorithm stage 1.

5. The community heat based routing method for the vehicle-mounted delay tolerant network, according to claim 1, wherein the selection of step 2-2 in the phase 2 is from the current road section to the RD _{home_main} Path PH with the highest probability _{max_prob} The method comprises the following steps:

step 2-2-1: loading direct transition probability matrix M between road sections in scene obtained based on mass statistics _{trans_prob} ；

Wherein RD ₁ ＝RD _x ，RD _m ＝RD _y ；

Step 2-2-3: matrix M _{trans_prob} All element values of which are in the interval [0,1), in the matrix M _{trans_prob} The method of utilizing the single-source shortest path algorithm SPFA obtains a route from the current road section to the RD _{home_main} The path with the maximum probability is then the path with the maximum probability from the current road section to the RD _expect Path PH with the highest probability _{max_prob} 。

6. The community popularity-based vehicular delay tolerant network routing method according to claim 1, wherein the community popularity-based vehicular delay tolerant network routing method in the step 5 of the phase 2

Updating the probability P (noted as P) of the source node of the message forwarding the message next time _{new_fast} ) The method comprises

7. The community popularity-based vehicular delay tolerant network routing method according to claim 1, wherein the steps 2-3 in the phase 2 are according to

Updating the probability P (marked as P) of the source node of the message forwarding the message next time _{new_PH} ) The method comprises

8. The community popularity-based vehicular delay tolerant network routing method according to claim 1, wherein the period T is set to 30 days.