CN109862536B - Accessibility method among multiple communities of large-scale Internet of vehicles - Google Patents
Accessibility method among multiple communities of large-scale Internet of vehicles Download PDFInfo
- Publication number
- CN109862536B CN109862536B CN201910171832.2A CN201910171832A CN109862536B CN 109862536 B CN109862536 B CN 109862536B CN 201910171832 A CN201910171832 A CN 201910171832A CN 109862536 B CN109862536 B CN 109862536B
- Authority
- CN
- China
- Prior art keywords
- community
- node
- nodes
- communication
- forwarding
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Abstract
In order to detect the communication between the large-scale Internet of vehicles and multiple social areas and keep the communication stable, the invention provides the accessibility method between the large-scale Internet of vehicles and the multiple social areas, applies the learning automata theory to the communication scheme between the large-scale Internet of vehicles and the multiple social areas, and adaptively adjusts the forwarding probability of different routes through information exchange and competition among the learning automata deployed in the community nodes, thereby achieving the purpose of optimizing network communication on the whole and improving the accessibility of the large-scale Internet of vehicles.
Description
The invention is further research and development of a patent document filed by inventor of chengdu et al on 2019, 3, 1, 3 and 2019101555842 on dynamic evolution method of internet of vehicles (applicant: college of congressions, patent application number: 2019101555842)) in the prior art, and the prior patent document can be considered as a part of the description of the invention.
Technical Field
The invention relates to the field of Internet of vehicles, in particular to a accessibility method among multiple communities of a large-scale Internet of vehicles.
Background
Accessibility is one of the most important characteristics of the network to realize interconnection and interworking, and mainly comprises connectivity and stability in the network. Connectivity mainly solves whether point-to-point routing in the network is reachable; the key point of stability is to optimize the network structure and routing strategy, and avoid the network efficiency problem caused by information congestion and transmission delay. The following description focuses on connectivity and stability in the car networking accessibility method.
(1) Connectivity
The research on connectivity of the internet of vehicles is divided into qualitative analysis and quantitative measurement analysis. Qualitative analysis generally refers to the influence of the distribution condition of vehicles on roads or the inherent characteristics of the internet of vehicles on connectivity, and quantitative measurement is to study the specific advantages and disadvantages of different connectivity strategies by comparing indexes such as average data delay or packet loss rate. Jin et al consider the distribution of vehicles on a road as a poisson distribution, and combined with the characteristics of vehicle road constraints, study the effect of vehicle density and communication range on connectivity. In the literature, a two-dimensional random graph model is further adopted for modeling on the basis that vehicle nodes conform to Poisson distribution, the correlation degree between the vehicle density and the minimum wireless transmission distance is qualitatively and quantitatively analyzed, and guidance is provided for the position deployment of key nodes in the internet of vehicles according to the correlation degree. In addition, the MCEGR method mentioned in the first chapter is a compromise for connectivity within and among the car networking communities, but there is a problem that MCEGR is a two-hop routing method, the community size is not very large, and the application range is limited.
(2) Stability of
The stability is one of key indexes for ensuring the continuous communication of the self-organizing network with the topology structure rapidly changing in the Internet of vehicles, and is an important component of the accessibility of the Internet of vehicles. Regarding accessibility schemes based on community evolution, stability of communities is a major concern. Morales et al propose a vehicle-mounted network self-adaptive community clustering algorithm, which predicts the position of a vehicle node at the next moment according to the motion track, the current speed and position, the road condition and the like of the vehicle node, and performs community division by combining the current network topology and the predicted network topology, thereby considering the trend of future changes of the network and improving the duration and stability of the community.
In summary, the methods have problems in connectivity and stability, when the network size is large and the topology is highly dynamic, due to the lack of adaptive routing, the network accessibility may not be sufficient because part of the connection is lost and cannot be recovered quickly. Aiming at the problems, the invention provides a method for adaptively adjusting forwarding probabilities of different routes based on a learning automaton for large-scale Internet of vehicles and multiple communities on the basis that an Internet of vehicles dynamic community evolution mechanism is researched by an Internet of vehicles dynamic community evolution method (applicant: university of unity, patent application number: 2019101555842) applied in 2019, 3, 1, and the Internet of vehicles dynamic community structure is obtained in real time.
Disclosure of Invention
The purpose of the invention is as follows:
the research method of the invention aims at the accessibility problem caused by the phenomena of complicated road network interleaving, frequent topology change, various communication protocols and the like objectively existing in a large-scale car networking network, sets corresponding excitation functions and penalty functions by utilizing the theory of learning automata and through information exchange and competition deployed among community nodes, adaptively adjusts the forwarding probability of different routes, and achieves a Nash equilibrium state, thereby realizing the purposes of optimizing data transmission in the network as a whole and improving the accessibility of the large-scale car networking network.
The existing problems related to connectivity and stability of the car networking community are that when the network scale is large and the topological structure is highly dynamic, due to lack of adaptive routing, rapid recovery cannot be achieved due to partial connection loss, so that the network connectivity is insufficient, or network congestion is caused due to data aggregation, so that partial nodes exceed the upper limit of communication load and cannot work normally, so that the network stability is poor. The final conclusion is then: the large-scale internet of vehicles as a dynamic self-organizing network can improve the accessibility in the community of the internet of vehicles only by establishing different routing forwarding probabilities which are adaptively adjusted and achieving a Nash equilibrium state.
Therefore, the invention specifically provides the following technical scheme: the community accessibility method of the Internet of vehicles specifically comprises the following steps:
Step2, accessibility method among multiple communities of large-scale Internet of vehicles
Step 2.1 initialization and update of node information table
Step 2.2 Community head node and gateway node screening
Step 2.3 method for adjusting forwarding behavior probability on forwarding LA between communities
Step 2.4 Access routing Algorithm between multiple communities
Advantageous effects
The invention aims to provide a method capable of improving multi-community communication and stabilizing accessibility of a vehicle networking under the condition of considering high dynamic complex conditions of the large-scale vehicle networking.
On the basis of a dynamic evolution method of the car networking community, (the inventor of Chengdu et al applies a dynamic evolution method of the car networking community (applicant: university of the same society, patent application number 2019101555842)) in 3, 1 and 2019, the invention provides a accessibility method of a large-scale car networking multi-community interval.
Description of the attached tables
TABLE 1 node information Table field
TABLE 2 node vchInter-community forwarding behavior probability vector table
Drawings
FIG. 1 is a schematic diagram of a gateway node and a head node in a vehicle networking community
FIG. 2 shows a flow chart of the community head node and gateway node screening (i.e., Algorithm 1 flow chart)
FIG. 3 is a schematic diagram of a communication process between the Internet of vehicles society
FIG. 4 is a flowchart of an inter-community forwarding LA forwarding behavior probability adjustment algorithm (i.e., Algorithm 2 flowchart)
FIG. 5 is a flow chart of an inter-community reachability routing algorithm (i.e., Algorithm 3 flow chart)
FIG. 6 LA operating diagram between communities in the Internet of vehicles
FIG. 7 TAPASCITONE data set road topology map
FIG. 8 PDR vs. vehicle Density variation
FIG. 9 PDR vs. packet transmission rate
FIG. 10E 2ED comparison with vehicle Density variation
FIG. 11E 2ED vs. packet transmission rate
FIG. 12 ROR vs. vehicle Density variation
FIG. 13 ROR vs. packet transmission rate
FIG. 14 is a flow chart of the method of the present invention
Detailed Description
The specific implementation process of the invention is shown in fig. 14, and includes the following 6 aspects:
① definition of related Properties
② node information table initialization and update
③ Community head node and gateway node screening
④ method for adjusting probability of forwarding behavior on community-based forwarding LA
⑤ multi-community accessibility routing algorithm
⑥ simulation experiment and result analysis
①
Correlation property definition
For nodes in the vehicle networking community, the accessibility scheme adopted by the invention endows the nodes with different roles, namely a community head node, a community gateway node and a community common node, and the definition is as follows:
define 1 Community head node set (CHSet): community CiThe head node (CH) of (1) is a set of nodes with large community centripetal force in the community, if in the community CiThe existence node u satisfies the mathematical expression (1):
wherein η is head node selection factor, generally η is (0.75, 1)]One value of (1), community CiIn which a node u satisfying the above condition is added to CiCHSet of (1). The node in the CHSet is the current community CiThe communication quality of the nodes is better, and the nodes of the CHSet are usually selected as relay nodes for communication in the community.
Define 2 Community gateway node set (GWSet): community CiThe gateway node of (2) is the node with the greatest attraction between the community and the adjacent communities. Namely: if community CiAnd CjAdjacent then CiRelative to CjThe gateway node u satisfies the mathematical expression (2):
zeta is a gateway node selection factor, and is generally taken as (0.9, 1)]One value of (1), community CiThe node u meeting the above condition is added to CiRelative to CjGWSet of (1). If community CiThere are multiple contiguous communities, then CiThere must be multiple gateway nodes, all of which are added to CiGWSet of (1). Each gateway node point will be used to communicate with its corresponding contiguous community. The community head node and the community gateway node are shown in fig. 1.
As can be seen from fig. 1, the head node of the community 1 is most closely connected to the nodes in the community, and is generally located at a more central position of the community. Further, there is more than one gateway node for community 1, with respect to community 2 and community 3, respectively, and there may be more than one.
Define 3 Community common node (CM): all nodes within a community except the head node may be referred to as community common nodes.
Generally, since the car networking is a highly dynamic network, the role of each node changes with the change of the movement and topology of the node and the communication situation, and the roles of the common nodes and the head nodes may be interchanged according to the requirement of network connectivity.
Defining 4 Node Connectivity Probability (NCP) refers to the credibility of Node Connectivity in the internet of vehicles.
If node u is adjacent to node v and within the respective wireless communication range, their Direct Node Connectivity Probability (DNCP) is the mathematical expression (3):
where dist (u, v) represents the distance between nodes u and v, and TR represents the maximum communication radius of the node. When the distance between the nodes is larger than the maximum communication radius of the nodes, the communication probability between the nodes is 0; otherwise, the probability of connectivity between nodes increases as the distance between nodes decreases.
If two nodes communicate indirectly, i.e. the nodes u and v can form a node communication path through other nodes, if this path is denoted as NodePathi＝{e1,e2,…,enIn which e1＝u,en＝v,n>2, n represents the number of nodes on the path, then nodes u and v are in NodePathiThe upper Node Connectivity Probability (PNCP) is:
i.e. the multiplication of the direct node connectivity probabilities of the u and v communication paths. Since there may be multiple Node communication paths between u and v, defining the Indirect Node Communication Probability (INCP) between nodes u and v as the maximum value of the communication probability on all Node communication paths:
INCP(u,v)＝max(PNCP(NodePathi)) (5)
in summary, the node connectivity probability between nodes is defined as the maximum value of DNCP and INCP:
NCP(u,v)＝max(DNCP(u,v),INCP(u,v)) (6)
a 5-Community Connectivity Probability (CCP) is defined to mean the credibility of the communication between two car networking communities.
If two car networking community CiAnd CjAdjacent and all having gateway nodes capable of maintaining communication with the opposite Community, their Direct Connectivity Probability (DCCP) satisfies the mathematical expression (7):
wherein u and v are communities C respectivelyiAnd CjGateway node of, two adjacent communities CiAnd CjThe direct connectivity probability equals the maximum of the node connectivity probabilities of their gateway nodes。
If two communities are communicating indirectly, i.e. community CiAnd CjCommunicypath with community communication pathi＝{C1,C2,…,CmIn which C is1＝Ci,C2＝Cj,m>2, m represents the number of communities on the path, then community CiAnd CjIn CommunnypathiThe above Community Connectivity Probability (PCCP) satisfies the mathematical expression (8):
i.e. CiAnd CjAnd (4) accumulating the direct connection probability of the community communication paths. Likewise, due to CiAnd CjThere may be multiple communication paths between communities, defining community CiAnd CjThe Indirect communication Probability (ICCP) between the two is the maximum value of the communication Probability on all the community communication paths, and satisfies the mathematical expression (9):
ICCP(Ci,Cj)＝max(PCCP(NodePathi)) (9)
in summary, the present invention defines the connection probability between communities as the maximum value of the direct connection probability and the maximum indirect connection probability satisfies the mathematical expression (10):
CCP(Ci,Cj)＝max(DCCP(Ci,Cj),ICCP(Ci,Cj)) (10)
②
node information table initialization and update
In a network communication layer of the Internet of vehicles, each node is provided with a node information table, and fields contained in the table comprise the ID of the node, the current time, the speed, the acceleration, the longitude and latitude of the position, the community attribution, the role of the node, the ID of the head node of the community of the node, the ID of a gateway node and the like. Specifically, the results are shown in Table 1.
In Table 1, the node ID is the unique identification of the node in the Internet of vehicles, and whether it is RSUThe type of the node is determined, the timestamp represents the current time, and the basic information of the node such as speed, acceleration, longitude and latitude can be obtained through the sensor. For node community attribution, the calculation process is as follows: at the initial moment, each node needs to broadcast a Neighbor Node Detection Message (NNDM) to nodes within the wireless signal propagation range, and a node receiving the NNDM needs to reply a confirmation message. After that, a dynamic evolution method of the vehicle networking community is adopted, (this part of inventions is referred to as the dynamic evolution method of the vehicle networking community (applicant: university of the same society, patent application number: 2019101555842), which is applied by inventor of Chengdu et al on 3/1 in 2019), to determine the Node community attributioniThe number of nodes is m, and its NAL is:
wherein, neip,qWhen 0 denotes community CiNode v inpAnd vqThere is no edge in between, otherwise, v is representedpAnd vqDirect connectivity probability of. In the community merging process based on the node similarity and the evolution process based on the increment, NAL is exchanged among the nodes, so that each node knows the node adjacency information of the community to which the node belongs.
After the Community structure at the current time is determined, a Community Adjacency List (CAL) of each Community, that is, adjacency Community information of the Community, may be obtained in a broadcast manner similar to the node adjacency list.
③
Community head node and gateway node screening
After the Community structure at the current time is determined, a Community Adjacency List (CAL) of each Community, that is, adjacency Community information of the Community, may be obtained in a broadcast manner similar to the node adjacency list. The screening of the community head nodes can be implemented by referring to the content of definition 1, the specific steps are shown as algorithm 1, and the specific flow chart is shown as fig. 2.
Through the steps in the algorithm 1, the head node and the gateway node of the community are screened out, and the content of the CHSet is stored in the node information table of each node, so that information support is provided for the establishment of the subsequent reachability routing.
④
Method for adjusting forwarding behavior probability on social interval forwarding LA
For inter-social communication of the Internet of vehicles, i.e. source node voriAnd a target node vdesThe communication process can be roughly divided into three steps:
(1)vorigateway node GW for forwarding information to community in which source node is located via intra-community communicationori；
(2) Gateway node GWoriGateway node GW for forwarding information to community where target node is located through inter-community communicationdes；
(3)GWdesForwarding information to target node v via intra-community communicationdes。
Wherein steps (1) and (3) belong to intra-community communication. In the step (1), if the community gateway node GW is in the communication range of the common node v, the common node directly forwards the information to the gateway node; if the community head node CH cannot directly communicate with the GWSet, the common node firstly sends the information to the head node CH and then forwards the information to a proper gateway node in the GWSet through the CH because the community head node CH has the routes of all nodes in the community. Step (3) is the reverse process of step (1), and is only different in communities, which is not described herein again. And (2) the communication among communities belongs, and the gateway node is a bearing unit for the communication among the communities and plays a role in contacting adjacent communities. A schematic diagram of the above process is shown in fig. 3.
In FIG. 3, source community CoriWith target community CdesWhen establishing communication, CoriThe head node in (2) needs to determine which community to forward the message next, and forward the message to the gateway node corresponding to the community. To this end, for example, the current community is CiAt its head node vchA set of social interval forwarding behavior vector tables is maintained, as shown in table 2.
The inter-community forwarding behavior probability vectors in table 2 do not refer to gateway nodes because at the next hop CnextThe expression already includes the gateway node corresponding thereto. CPFrIs represented as C in the target communitypThe next hop community is CnextqThe forwarding probability of (2). Likewise, the value of PF changes as the communication progresses, and the changing process is deployed at the community head node v by the environment pairchThe feedback mechanism of upper LA. In order to quantify the feedback of the environment during the inter-social communication, a new-scale Community Opportunity to Forward Evaluation (COFE) is defined to satisfy the mathematical expression (12):
where CCP is the community connectivity probability, RER and Delay represent the community residual energy and community latency, η ″, respectively,
And ψ "are the adjustment coefficients for NCP, RER, and Delay, respectively.
Likewise, deployed at vchThe forwarding behavior vector number of the forwarding LA between communities at the initial time is set as l, and the LA forwarding probability is initialized as follows:
i.e. t is 0, incEach forwarding probability on the inter-zone forwarding LA is the same. Selecting COFF of ith forwarding action at subsequent timeiAnd the average chance forwarding evaluation factor COFE on the LAavgBy comparison, if COFEi≥COFEavgThen LA takes an incentive action on this behavior:
if COFEi≤COFEavgThen LA makes a penalty action for this behavior:
wherein ρ is an excitation parameter and ρ' is a penalty parameter.
In summary, the specific steps of adjusting the forwarding behavior probability on the inter-community forwarding LA are shown in algorithm 2, and the specific flowchart is shown in fig. 4.
The algorithm 2 is to adjust the forwarding probability of the community-based forwarding LAs deployed on each community head node, and when the community-based message forwarding occurs, the LAs excite or punish the forwarding behavior probability vector according to the COFE value at the previous moment.
⑤
Multi-community accessibility routing algorithm
Based on the algorithm 2, the inter-community accessibility routing method is obtained, the detailed steps are shown as an algorithm 3, and the specific flow chart is shown as fig. 5.
In Algorithm 3, between communitiesThe adjacency list CAL is similar to NAL and indicates whether or not communities are adjacent. If community CiAnd CjIs not adjacent to each other
If adjacent to each other
Is community CiAnd CjDirect connectivity probability of. In addition, in order to prevent the message from trapping in a dead loop or losing in the inter-community transfer, the maximum community forwarding hop value is defined as CHOPmaxIf the jump value K exceeds CHOPmaxThe source community will retransmit.
The forwarding LAs deployed among communities in each Internet of vehicles community exchange data and compete with each other in network communication, the network accessibility of the highly dynamically-changed Internet of vehicles can be continuously optimized through the operation of the LAs, and the operation condition of the LAs among the Internet of vehicles communities is shown in FIG. 6.
⑥
Simulation experiment and result analysis
(1) Simulation experiment data and method
1) Experimental data
In order to verify the correctness of the method provided by the invention, a data set with a wider range, more road categories and larger magnitude order, namely a tapascrone data set, is selected by the experimental data of the invention, the data set collects road information and vehicle movement information (a road topological graph is shown in fig. 7 and covers a central urban road section, a suburban road section and a rural road section) in the range of four hundred square kilometers in the city of germany, and generates movement track data of all vehicles in the area within 24 hours.
According to the experiment, vehicle track data in a time period of 6: 00-8: 00(am) in TAPASColone is used as the basis of the evolution simulation experiment of the Internet of vehicles community, and in the time period, the number of vehicle nodes moving on a road is at most 8000, so that the experiment meets the requirement of the experiment on the large-scale Internet of vehicles community evolution experiment.
2) Experimental methods
The simulation experiment is based on the tapascrologne data set, wherein the SUMO simulator is still a vehicle simulation tool, and the OMNET + + software and the Veins framework are still used as network simulation tools. The invention provides a learning automaton-based dynamic evolution method of an internet of vehicles (CAVN-LA) on the basis of a dynamic evolution method of the internet of vehicles (applicant: university of unity, patent application number 2019101555842) applied by inventor of Chengdu et al on 3/1 in 2019, and the accuracy of the dynamic evolution mechanism of the internet of vehicles (CAVN-LA) is verified by comparing the advantages of the CAVN-LA relative to other algorithms (MCEGR and AMACAD) through simulation experiments.
(a) Average Packet Delivery Rate (PDR): the ratio average value of the number of the destination node data packets and the total number of the data packets sent to the destination node by the source node is successfully achieved;
(b) average End-to-End delay (End-to-End delay, E2 ED): the mean value of the time required for a data packet sent by a source node to reach a destination node;
(c) average Routing Overhead Radio (ROR): the data packets for discovering and exchanging routes occupy the average of the ratio of all the data packets from the beginning of the route probing and building to the completion of the communication.
(2) Analysis of results
The simulation experiment result of the invention is obtained by comparing three network indexes of average packet delivery rate, average end-to-end delay, average route overhead rate and the like of different algorithms under two experimental control conditions of node density and data packet sending rate. In order to control the mutual influence of the variables as much as possible, when the relation between the network index and the vehicle density is tested, the invention takes another variable, namely the sending rate of the data packet is 0.3 p/s; when the relationship between the network index and the packet transmission rate is tested, another variable, that is, the vehicle density is 0.05 vehicle/meter, which will be described in detail below.
(1) Average packet delivery rate PDR
FIG. 8 is a PDR value as a function of vehicle density. Obviously, the PDR values of the three algorithms are positively correlated with the vehicle density, the CAVN-LA and AMACAD algorithms converge faster, and the MCRGR converges slower. The CAVN-LA and AMACAD algorithms are stable in performance after the vehicle density is larger than 0.03 vehicle/meter, and due to the fact that self-adaptive and self-adjusting mechanisms are adopted in the two algorithms, the success rate and stability of data packet delivery are high. The MCRGR algorithm has larger variation amplitude because the algorithm limits the community range to two hops at most, has smaller community scale, performs better under the condition of larger vehicle density, and performs generally on node sparse scenes, but the method has the data packet delivery rate close to the CAVN-LA algorithm proposed by the text after the vehicle density is more than 0.05 vehicle/meter, because the evolution competition mechanism in the MCRGR has similarity with the CAVN-LA algorithm, and performs better under the node dense scenes. In general, the learning automata mechanism in the CAVN-LA algorithm enables the self-adaptive self-adjusting capability to be strong, and the PDR of the CAVN-LA algorithm has good performance in both node sparse and dense scenes.
Fig. 9 is a graph of PDR value as a function of packet transmission rate. It can be seen that the PDR values of the three algorithms all decrease as the packet transmission rate increases, because as the packet transmission rate continues to increase, the data in the car networking network becomes denser, the probability of data collision increases, and the rate of packet delivery success decreases. The PDR value change of CAVN-LA and AMACAD algorithms is still stable, but the highest PDR of CAVN-LA is higher than AMACAD by more than 10%. The highest PDR of MCRGR is similar to CAVN-LA, close to 90%, but after the packet sending rate is greater than 0.8, the performance of MCRGR decreases sharply, and CAVN-LA still performs well, because the number of each community head node and gateway node in the algorithm may be more than one, and the mechanism of learning automata plays a role in load balancing.
(2) Average end-to-end delay E2ED
Fig. 10 is a graph of E2ED versus vehicle density change. As can be seen from the figure, the E2ED values of the three algorithms decrease as the vehicle density increases, because as the vehicle density increases, the network connectivity between the internet of vehicles nodes increases, the number of times of carrying forwarding decreases, and thus the average delay decreases. Where the average delay of CAVN-LA is lowest, but when the vehicle density is greater than 0.055 vehicles/meter, the average delay of the MCRGR algorithm is reduced to be comparable to the CAVN-LA algorithm. In addition, the E2ED value change of the AMACAD algorithm is also relatively stable, but its average delay is slightly higher than that of the other two algorithms in the case of dense nodes due to insufficient adaptive capability of routing adjustment.
Fig. 11 is a graph of E2ED versus packet transmission rate change. The values of E2ED of all three accessibility routing algorithms become larger as the sending rate of the data packet increases, and the trend of the larger values becomes larger, because as the sending rate of the data packet increases, the load of the network becomes larger, so that the subsequent data packet enters the waiting sequence of the forwarding node, and the average delay of the network becomes larger. At a packet transmission rate of 0.005p/s, the E2ED values for the three methods are quite different, all being about 0.2s, but at a packet transmission rate of 0.06p/s, the performance of the MCRGR algorithm has degraded dramatically to 6s, exceeding the CAVN-LA algorithm by nearly 2.3 s. Thus, overall, the CAVN-LA algorithm has a slightly lower average delay than the other two algorithms, which is advantageous.
(3) Average route overhead rate ROR
FIG. 12 is a graph of ROR as a function of vehicle density. It can be seen that the ROR of the three algorithms increases with the increase of the vehicle density, wherein the ROR value of the AMACAD algorithm is the highest, because the algorithm costs a large amount of overhead for the packet switching mechanism for establishing the internet-of-vehicles community, and the routing overhead increases continuously with the increase of the vehicle density. The ROR value of the CAVN-LA algorithm is ranked second, because a competitive evolution mechanism in the algorithm requires a certain number of control messages for dynamic update of the head node and the gateway node set and punishment of exchange of the excitation table. The MCEGR algorithm is not very sensitive to the change of the vehicle density, the algorithm is established on the basis of node position prediction, and the two-hop community determines that the proportion of control messages is low and keeps about 5 percent.
Fig. 13 is a graph of ROR as a function of packet transmission rate. It can be seen in fig. 12 that the RORs of the CAVN-LA algorithm and the MCEGR algorithm are the same when the packet transmission rate is 0.3p/s and the vehicle density is 0.015 vehicle/meter, which is also demonstrated in fig. 13. In addition, the ROR value of the AMACAD algorithm is still the highest of the three, and as the packet transmission rate increases, the disadvantage of the algorithm in terms of routing overhead becomes more and more obvious. Overall, the ROR of the CAVN-LA algorithm is slightly higher than MCEGR and lower than AMACAD by about 2%, and by combining the performance of the algorithm in terms of PDR and E2ED, the indication of accessibility performance obtained by replacing the small cost of ROR is worth.
Innovation point
The innovation points are as follows: based on a dynamic evolution method of the vehicle networking community, (the dynamic evolution method of the vehicle networking community (applicant: university of Tongji, patent application number: 2019101555842) applied by inventor of Chengdu et al in 2019, 3 and 1), by utilizing a learning automata technology, through information exchange and competition deployed among community nodes, corresponding excitation functions and penalty functions are set, forwarding probabilities of different routes are adjusted in a self-adaptive mode, and accessibility among multiple communities of the large-scale vehicle networking is achieved.
The large-scale Internet of vehicles network has characteristics of complicated road network staggering, frequent topology change, various communication protocols and the like objectively, and great challenges are brought to the communication and stability of the large-scale Internet of vehicles network to a great extent. Aiming at the problems, the learning automata theory is applied to a communication scheme of a large-scale Internet of vehicles and multiple social intervals, and the forwarding probabilities of different routes are adaptively adjusted through information exchange and competition among the learning automata deployed in community nodes, so that the aim of optimizing network communication on the whole is fulfilled, and the accessibility of the large-scale Internet of vehicles is improved.
Attached table of the specification
TABLE 1
TABLE 2
Claims (1)
1. A large-scale Internet of vehicles multi-community accessibility method specifically comprises the following steps:
step1 definition step:
define 1 Community head node set (CHSet): community CiThe head node (CH) of (C) is the set of nodes in the community with larger community attraction, if in the community CiThe existence node u satisfies the mathematical expression (1):
wherein η is head node selection factor, η is (0.75, 1)]One value of (1), community CiIn which a node u satisfying the above condition is added to CiCHSet of (1); the node in the CHSet is the current community CiThe communication in the community selects a CHSet node as a relay node;
define 2 Community gateway node set (GWSet): community CiThe gateway node of (a) means the node with the greatest attraction between the community and its adjacent communities respectively, namely: if community CiAnd CjAdjacent then CiRelative to CjThe gateway node u satisfies the mathematical expression (2):
zeta is a gateway node selection factor, and is generally taken as (0.9, 1)]One value of (1), community CiThe node u meeting the above condition is added to CiRelative to CjGWSet of (1); if community CiThere are multiple contiguous communities, then CiThere must be multiple gateway nodes, all of which are added to CiGWSet of (1); each gateway node is to be used for communication with its corresponding adjacent community;
define 3 Community common node (CM): except the head node in one community, the nodes can be called community common nodes;
the role of each node can change along with the change of the movement, the topology and the communication condition of the node, and the roles of the common nodes and the head nodes can be interchanged according to the requirement of network communication;
defining 4 Node Connectivity Probability (NCP) to be the credibility of Node Connectivity in the Internet of vehicles;
if node u is adjacent to node v and within the respective wireless communication range, then their Direct Node Connectivity Probability (DNCP) is the mathematical expression (3):
where dist (u, v) represents the distance between nodes u and v, and TR represents the maximum communication radius of the node; when the distance between the nodes is larger than the maximum communication radius of the nodes, the communication probability between the nodes is 0; otherwise, the connection probability between the nodes is increased along with the reduction of the distance between the nodes;
defining 5 Community Connectivity Probability (CCP) which is the credibility of communication between two internet of vehicles communities;
if two car networking community CiAnd CjAdjacent and all having gateway nodes capable of maintaining communication with the opposite Community, then their Direct Connectivity Probability (DCCP) mathematical expression (4):
wherein u and v are communities C respectivelyiAnd CjGateway node of, two adjacent communities CiAnd CjThe direct connectivity probability is equal to the maximum of the node connectivity probabilities of their gateway nodes;
step2, accessibility method among multiple communities of large-scale Internet of vehicles
Step 2.1 node information table initialization and update:
in a network communication layer of the Internet of vehicles, each node is provided with a node information table, and fields contained in the table comprise the ID of the node, the current time, the speed, the acceleration, the longitude and latitude of the position, the community attribution, the role of the node, the ID of the head node of the community and the ID of a gateway node;
the node ID is the unique identification of the node in the Internet of vehicles;
the timestamp represents the current time;
the speed, the acceleration and the position longitude and latitude are obtained through sensors;
for node community attribution, the calculation process is as follows: at the initial moment, each node needs to broadcast NNDM (Neighbor node detection message) to nodes in a wireless signal propagation range, the nodes receiving the NNDM need to reply a confirmation message, whether edges exist between the nodes and Neighbor nodes is judged through the process, so that Neighbor information and a network topological structure are determined, and then the known dynamic evolution method of the Internet of vehicles community is adopted to determine the attribution of the node community;
the node roles comprise a common node, a head node and a gateway node;
in the NNDM message, not only the basic information of the Node itself, but also a Node Adjacency List (NAL) of the community where the Node is located, wherein the NAL is a two-dimensional array; if community CiThe number of nodes is m, and its NAL is:
wherein, neip,qWhen 0 denotes community CiNode v inpAnd vqThere is no edge in between, otherwise, v is representedpAnd vqIn the community merging process based on the similarity of the nodes and the evolution process based on the increment, the NAL is exchanged among the nodes, so that each node knows the node adjacent information of the community to which the node belongs;
after the Community structure at the current moment is determined, a Community Adjacency List (CAL) of each Community, that is, Adjacency Community information of the Community, can be obtained in a broadcast manner similar to the node Adjacency List;
step 2.2 Community head node and gateway node screening
After the Community structure at the current moment is determined, a Community Adjacency List (CAL) of each Community, that is, Adjacency Community information of the Community, can be obtained in a broadcast manner similar to the node Adjacency List; the screening of the community head nodes and the community gateway nodes can be realized by referring to the contents of the definition 1 and the definition 2 in the step 1;
step 2.3 method for adjusting forwarding behavior probability on forwarding LA between communities
For inter-social communication of the Internet of vehicles, i.e. source node voriAnd a target node vdesThe communication process can be roughly divided into three steps:
(1)vorigateway node GW for forwarding information to community in which source node is located via intra-community communicationori；
(2) Gateway node GWoriGateway node GW for forwarding information to community where target node is located through inter-community communicationdes；
(3)GWdesForwarding information to target node v via intra-community communicationdes；
Wherein, the steps (1) and (3) belong to intra-community communication; in the step (1), if the community gateway node GW is in the communication range of the common node v, the common node directly forwards the information to the gateway node; if the direct communication cannot be carried out, the community head node CH has the routes of all nodes in the community, so that the common node firstly sends the information to the head node CH and then forwards the information to a proper gateway node in the GWSet through the CH; the step (3) is the reverse process of the step (1), and the two processes are in different communities; step (2) belongs to inter-community communication, and the gateway node is a bearing unit for inter-community communication and plays a role in contacting adjacent communities;
source community CoriWith target community CdesWhen establishing communication, CoriThe head node in the network node needs to judge which community to forward the message next, and forward the message to the gateway node corresponding to the community;
the process of adjusting the probability of forwarding behavior on the social interval forwarding LA is algorithm 2:
inputting: community set CS
And (3) outputting: air conditioner
step1, judging whether a certain community Ci in a community set CS has a forwarding task, if so, step2, otherwise, accessing other communities in the community set;
step2, calculating a community opportunity forwarding judgment factor of each community Ci about each community to be issued to the target community;
step3, if the community opportunity forwarding judgment factor is larger than the average community opportunity forwarding judgment factor, if the community opportunity forwarding judgment factor is satisfied, the step4 is reached, otherwise, the step5 is reached;
step4, adjusting the forwarding probability CPF by using a reward function;
step5, adjusting the forwarding probability CPF by a penalty function;
step6, judging whether other communities which are not visited exist, if so, reaching step1, otherwise, finishing the algorithm;
step 2.4 Access routing Algorithm between multiple communities
An inter-community accessibility routing algorithm, algorithm 3:
the algorithm 3 describes:
inputting: a community set CS; a primary communication request between communities comprises a source community Cori and a target community Cdes; a social interval adjacency table CAL;
and (3) outputting: air conditioner
step1, if Cori and Cdes are adjacent communities, directly communicating, otherwise to step 2;
step2, assigning Cori to a temporary community Ctemp, and assigning the value of the jumping times K to 0;
step3, if Ctemp is not the target community Cdes and K is less than the maximum jump number CHOPmax, go to step4, otherwise go to step 8;
step4, selecting the CPF with the maximum forwarding probability to forward the message;
step5, updating each forwarding row probability CPF of the community Ctemp by using an algorithm 2;
step6, adding one to the value of K and assigning Cnext hop of the forward message Cnext to Ctemp;
step7, jumping to step 3;
step8 if the value of K is greater than CHOPmax, jump to step2, otherwise the message is forwarded to Cdes and the algorithm ends.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910171832.2A CN109862536B (en) | 2019-03-07 | 2019-03-07 | Accessibility method among multiple communities of large-scale Internet of vehicles |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910171832.2A CN109862536B (en) | 2019-03-07 | 2019-03-07 | Accessibility method among multiple communities of large-scale Internet of vehicles |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109862536A CN109862536A (en) | 2019-06-07 |
| CN109862536B true CN109862536B (en) | 2020-06-16 |
Family
ID=66900157
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910171832.2A Active CN109862536B (en) | 2019-03-07 | 2019-03-07 | Accessibility method among multiple communities of large-scale Internet of vehicles |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109862536B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110247826B (en) * | 2019-07-10 | 2022-03-25 | 上海理工大学 | Point-to-point network connectivity test method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102917372A (en) * | 2012-10-18 | 2013-02-06 | 北京理工大学 | Internet of vehicles hierarchical model design method orienting to multi-scene data exchange |
| CN103079270A (en) * | 2013-02-05 | 2013-05-01 | 中国电子科技集团公司电子科学研究院 | Method for sensing aggregation of mobile nodes on basis of wireless positioning |
| CN108712287A (en) * | 2018-05-22 | 2018-10-26 | 同济大学 | VANET community discovery methods based on node similarity |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8874477B2 (en) * | 2005-10-04 | 2014-10-28 | Steven Mark Hoffberg | Multifactorial optimization system and method |
| US8570930B2 (en) * | 2009-12-08 | 2013-10-29 | Electronics And Telecommunications Research Institute | Location based vehicle multihop protocol unicast apparatus and routing method using the apparatus |
| CN101808382B (en) * | 2010-01-20 | 2012-09-05 | 西安电子科技大学 | Data communication method of wireless sensor network |
| CN103781148B (en) * | 2014-02-25 | 2016-12-21 | 重庆邮电大学 | Stable cluster method for routing based on link-aware in vehicular ad hoc network |
| KR101700588B1 (en) * | 2015-09-15 | 2017-02-13 | 가톨릭관동대학교산학협력단 | VANET system using a Mutual Authentication Scheme and Black Hole detection method using thereof |
| CN105792311B (en) * | 2016-02-29 | 2019-04-09 | 重庆邮电大学 | A kind of car networking method for routing based on User Activity regional model |
| CN107509229B (en) * | 2017-08-03 | 2019-10-18 | 华南理工大学 | A kind of car networking chance method for routing calculated based on vehicle node centrality |
| CN108881016B (en) * | 2018-08-30 | 2020-12-08 | 同济大学 | Method for constructing self-organizing network community of Internet of vehicles and balancing communication load of community nodes |
| CN109302696B (en) * | 2018-08-30 | 2021-04-30 | 同济大学 | Multi-role classification community clustering method for self-organizing network of Internet of vehicles |
-
2019
- 2019-03-07 CN CN201910171832.2A patent/CN109862536B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102917372A (en) * | 2012-10-18 | 2013-02-06 | 北京理工大学 | Internet of vehicles hierarchical model design method orienting to multi-scene data exchange |
| CN103079270A (en) * | 2013-02-05 | 2013-05-01 | 中国电子科技集团公司电子科学研究院 | Method for sensing aggregation of mobile nodes on basis of wireless positioning |
| CN108712287A (en) * | 2018-05-22 | 2018-10-26 | 同济大学 | VANET community discovery methods based on node similarity |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109862536A (en) | 2019-06-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Distributed cooperative reinforcement learning-based traffic signal control that integrates V2X networks’ dynamic clustering | |
| CN105208616A (en) | Road topology based adaptive multi-copy routing method in vehicular ad hoc network | |
| Smiri et al. | Geographic and topology based routing protocols in vehicular ad-hoc networks: Performance evaluation and QoS analysis | |
| CN111510956B (en) | Hybrid routing method based on clustering and reinforcement learning and ocean communication system | |
| Houssaini et al. | Improvement of GPSR protocol by using future position estimation of participating nodes in vehicular ad-hoc Networks | |
| Deshmukh et al. | Analysis of cluster based routing protocol (CBRP) for vehicular adhoc network (VANet) in real geographic scenario | |
| Vafaei et al. | A new QoS adaptive multi-path routing for video streaming in urban VANETs integrating ant colony optimization algorithm and fuzzy logic | |
| CN107148064A (en) | A kind of In-vehicle networking routed path optimization method based on population | |
| CN108684063A (en) | A kind of on-demand routing protocol improved method based on network topology change | |
| Yang et al. | V2V routing in VANET based on heuristic Q-learning | |
| CN109862537B (en) | Accessibility method in Internet of vehicles community | |
| Inam et al. | A novel protocol for vehicle cluster formation and vehicle head selection in vehicular ad-hoc networks | |
| Shen et al. | AODV-PNT: An improved version of AODV routing protocol with predicting node trend in VANET | |
| Bengag et al. | A novel greedy forwarding mechanism based on density, speed and direction parameters for vanets | |
| CN109862536B (en) | Accessibility method among multiple communities of large-scale Internet of vehicles | |
| Zhao et al. | A vehicle density and load aware routing protocol for VANETs in city scenarios | |
| CN105228215A (en) | Based on many copies method for routing of decision tree mechanism in vehicular ad hoc network | |
| Meraihi et al. | QoS performance evaluation of AODV and DSR routing protocols in city VANET scenarios | |
| Sharma et al. | A contemporary proportional exploration of numerous routing protocol in VANET | |
| Ghanbarzadeh et al. | Reducing message overhead of AODV routing protocol in urban area by using link availability prediction | |
| Hamedani et al. | A new two level cluster-based routing protocol for vehicular ad hoc network (VANET) | |
| CN108881016A (en) | The polygonal colour sorting community load-balancing algorithm of car networking self-organizing network | |
| Manel et al. | SODV speed based ad hoc on demand vector link routing protocol: A routing protocol for VANET networks | |
| Omar et al. | Design and development of GreedLea routing protocol for Internet of Vehicle (IoV) | |
| He et al. | A link state aware hierarchical road routing protocol for 3D scenario in VANETs |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |