CN109511123A - A kind of software definition vehicle network adaptive routing method based on temporal information - Google Patents
A kind of software definition vehicle network adaptive routing method based on temporal information Download PDFInfo
- Publication number
- CN109511123A CN109511123A CN201811609539.1A CN201811609539A CN109511123A CN 109511123 A CN109511123 A CN 109511123A CN 201811609539 A CN201811609539 A CN 201811609539A CN 109511123 A CN109511123 A CN 109511123A
- Authority
- CN
- China
- Prior art keywords
- time
- routing
- node
- vehicle
- prediction
- 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.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/22—Traffic simulation tools or models
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W40/00—Communication routing or communication path finding
- H04W40/02—Communication route or path selection, e.g. power-based or shortest path routing
- H04W40/20—Communication route or path selection, e.g. power-based or shortest path routing based on geographic position or location
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W40/00—Communication routing or communication path finding
- H04W40/02—Communication route or path selection, e.g. power-based or shortest path routing
- H04W40/22—Communication route or path selection, e.g. power-based or shortest path routing using selective relaying for reaching a BTS [Base Transceiver Station] or an access point
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/40—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P]
- H04W4/44—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P] for communication between vehicles and infrastructures, e.g. vehicle-to-cloud [V2C] or vehicle-to-home [V2H]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/70—Services for machine-to-machine communication [M2M] or machine type communication [MTC]
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention belongs to wireless communication technology fields, are related to a kind of software definition vehicle network adaptive routing method based on temporal information.The present invention routes data and current distance information, Markov model corresponding to current vehicle network struction according to real-time history.Following routing is predicted using the Markov model of building, optimal routing plan is rapidly finally calculated by a time diagram optimal path algorithm with these data.The concept of time diagram has been brought into router-level topology by the method, and vehicle network is regarded as one group of static map compared with the past, and time diagram considers more temporal informations, so that strategy is more met reality scene, while using more efficient time nomography；Using software defined network framework, the network architecture is divided into control plane and data plane.The present invention can effectively promote communication efficiency.
Description
Technical field
The invention belongs to wireless communication technology fields, are applied to car networking, and it is fixed to be related to a kind of software based on temporal information
Adopted vehicle network adaptive routing method is based on software defined network framework, controls plane based on the number collected on data plane
According to decision is carried out, optimal routing plan is calculated.
Background technique
Although distributed management mode has been widely adopted, vehicle self-organizing network (VANET) still can not optimize pipe
Manage and meet the needs of intelligent transportation system (ITS).However as a kind of emerging network system, software definition In-vehicle networking
(SDVN) limitation caused by current vehicle-carrying communication framework has been broken.Software definition vehicle network (SDVN) is a kind of novel net
Network system, it is intended to network be configured with reinforcing the management and programmable to vehicle communication, SDVN main thought is will to count
According to plane and control planar separation, so that being interacted between vehicle even closer.SDVN uses logical set Chinese style network controller, claims
It for controller, can mainly manage, reconcile and reinforce the communication between network element, and obtain global network situation from data plane.
In this case, underlay network device only needs to be responsible for forwarding data packet.At the same time, controller collecting network information, and
Calculate global Optimization route path.The framework provides greater flexibility for network management.However, the framework is being used for vehicle
There are still many challenging problems with management aspect for the center control of network.One of outstanding question be as
What effectively calculates most suitable path for each car in real time, and adjusts calculative strategy according to current network condition.For example, vehicle
Between, or exchange with the mass data of controller and also to bring unconfined delay during being grouped transmission, packet loss and
Network congestion.
In SDVN, In-vehicle networking is considered as a series of static maps by most of existing researchs.Each static state chart represents spy
Node and the side fixed time on a little.Then, by being based on single static map application static routing algorithm (usually as Dijkstra
Or the static shortest path first of Bellman-Ford) calculate flow table.However, true vehicle network is the actual time
Figure.Node be it is highly dynamic, the presence of link may only last for the very short time.Within such networks, node is specific
It is communicated in period with another node, and each side of vehicle time figure has its specific temporal information.Therefore, in no reality
When the network information in the case where routing is calculated based on static map may cause packet loss and cause in the forwarding stage later tight
Postpone again.
For example, Fig. 1 (a) shows the time diagram G of vehicle network, and Fig. 1 (b) shows its corresponding static map.As above
Described, each edge has its temporal information, is indicated by (u, v, t, d), indicates since the edge of u to v is in time t, and it
Continue d timestamp, until grouping successfully arrives at v.In fact, there may be multiple summits between u and v, show they
Relationship on different time stamp.For the sake of simplicity, it will be assumed that the duration of each edge is 1 in Fig. 1 (a), and in each edge
Number be its time started.We can calculate the shortest path from A to J now.In Fig. 1 (b), shortest path first is that
<A, D, G, J>, the distance of this paths is obviously 3.However, in Fig. 1 (a), it is impossible to find any access from A to J
Diameter.This is because we can only reach J from G, but the departure time the latest from G to J is when reaching earliest with arrival G the time 9
Between be the time 10.If we are using A as starting point, the difference of the temporal information between two figures arrive J can not in time diagram
It reaches.Fig. 1 explains that traditional static figure can generate error message in dynamic topology.When we calculate subsequent time in SDVN
When routed path (flow table), this mistake can also occur.First, it will be assumed that each node is vehicle or infrastructure, side generation
Vehicle is just being sent packets to another vehicle by table.Before data packet arrives at the destination (recipient), side (or link, in mistake
Go to be effective) it may be destroyed.In case of above situation, then need to generate retransmission in the controller of SDVN new
Routing, this will lead to packet loss, routing cost and delay.Therefore, it is necessary to which temporal information is stored in SDVN.
However, existing SDVN controller research is concentrated mainly on the flow table based on static topological and calculates, and vehicle network
Actually time diagram.Based on nearest research, time diagram optimal path algorithm only needs the time complexity and O of O (m+n)
(n) space complexity.If we calculate optimal path in a manner of being similar to dijkstra's algorithm using greedy strategy, just
Minimum priority query is needed, then greedy strategy may be very inefficient compared with the optimal path algorithm of time diagram.O(mlogπ
+ mlogn) time complexity and O (M+n) space complexity.Therefore, it is necessary to consider the time letter in the routing of SDVN
Breath.Moreover, controller should next timestamp fall into a trap point counting group transmission routing, wherein we need to predict this calculating
Required temporal information.Therefore, we have proposed the routing algorithms of SDVN a kind of, i.e., based on the adaptive routing of temporal information
(TibAR)。
Summary of the invention
Most suitable routing, this hair are calculated in real time to solve the problems, such as how existing SDVN controller is efficiently each car
Bright technical problem to be solved is just to provide a kind of software definition vehicle network adaptive routing method based on temporal information,
Based on history routing and its temporal information, vehicle ground route requests are efficiently handled.
The invention is realized in this way a kind of software definition vehicle network adaptive routing method based on temporal information,
Include the following steps:
1) vehicle node periodically sends position and history routing iinformation to controller, controller according to these information architectures simultaneously
Update the Markov model of corresponding current vehicle network；
2) when controller receives the route requests of vehicle node, it is appropriate to be predicted according to the Markov model that step 1 constructs
The following routing iinformation of quantity, the input as optimal route computing method；
3) controller constructs following corresponding time diagram using the following routing iinformation predicted, and most by time diagram
Shortest path algorithm calculates optimal routing plan, and is sent to request vehicle.
Further, in step 1), the Turnover Index i between node pair each in In-vehicle networking is utilizedijTo construct Ma Er
It can husband's model prediction matrix:
Wherein, dijFor the distance between node i to node j, distance is longer, and Turnover Index is with regard to smaller, thFor time value, tp
For current time value, t0For the time value that the history routing being able to record earliest starts, longer apart from current time, then conversion refers to
Number is smaller, this history routing iinformation is more inessential,For in time thWhen, node i to a route to be valid between node j
Value, if it is in time thWhen, for node i to information transmission is occurring between node j, then the value is 1, otherwise is for 0, k and l
Weight coefficient, k is bigger, then range information is more important, otherwise l bigger, and routing iinformation is more important；
A Turnover Index is calculated when each time cycle starts, and is stored it in two-dimensional N N matrix, wherein
One-dimensional representation current state, the state at another one-dimensional representation lower a moment, so far the original state of Markov model, which constructs, completes.
Further, in step 2), the prediction of the following routing iinformation is carried out using the Markov model that step 1 constructs,
Including forward prediction and backward prediction:
Forward prediction:
A) an empty ordered list L is created；
B) M [s] is traversed, M [s] is Turnover Index of the starting point s to other states, if the Turnover Index between s and j is higher than
From the average value of the Turnover Index of s, then this routing is qualified, and L is inserted on this side；
C) forward prediction is recursively used, starting point is j, and initial time is ta+dsj, and the number of iterations adds 1；
D) when router-level topology is reached home or the number of iterations is equal to MaxTime, this branch will reach its end, all
After recurrence is completed, which terminates, and the routing of all forward predictions is stored in L by the sequence of its time started；
Backward prediction:
Difference with forward prediction step is: since terminal, each recurrence all using the starting point of previous step as terminal,
I.e. backward prediction routes from finish to start, and termination condition is to calculate to reach starting point or the number of iterations and be equal to MaxTime, and preceding
It is identical to predicting；
It is arranged using d as the termination time of the upper a line of terminal, then executes backward prediction to postorder, once encounter one
The starting point on side is the terminal for a line being stored in L, then involved edge in L will be adjusted at the beginning of this side
The minimum end time, together with the routing link for thus predicting forward prediction algorithm and backward prediction algorithm, it is appropriate to obtain
The following routing iinformation of quantity.
It further, include following step by the method that time diagram optimal path algorithm calculates optimal routing in step 3)
It is rapid:
First, in accordance with the side e (u, v, t, d) in the order traversal L of time started, the earliest arrival of each node is initialized
Time and upper hop node are indicated with array t [N] and prev [N], then carry out filter side, if this edge arriving earliest in u
U is left before up to the time, then this edge is not considered, if earliest arrival time of the end time on this side earlier than v, updates most
Early arrival time simultaneously rebuilds upper hop node, if the end time is equal to earliest arrival time, prev [N] is added in this edge；
Finally, terminating this algorithm after scanning through all sides, t [N] and prev [N] are Optimization route results.
Further, when controller obtains the distance between vehicle u and vehicle v information update, Turnover Index updates are as follows:
Wherein, iuvIndicate the target Turnover Index between u and v, i 'uvIt is the Turnover Index calculated according to formula (1), duv
With d 'uvIt is the current and history range information between u and v, transition index is adaptively adjusted according to the distance of current and past,
Similarly, when the grouping transmission from u to v occurs, Turnover Index is adjusted, as described in formula (3):
Wherein de refers to the decline coefficient in 0 and 1 range, in one for occurring to be grouped transmission every time between u and v
After the timestamp past, due to being not intended to same side to undertake multi information transformation task, so dropping the importance on the side of u to v
It is low, Turnover Index also corresponding decline, finally, the Turnover Index average value of u needs corresponding change:
In order to guarantee accuracy, at a fixed time be spaced after, restarting formula (2), formula (3), formula (4) into
Journey recalculates Markov model according to newest distance and routing.
Compared with the prior art, the advantages of the present invention are as follows: the present invention uses the software definition of short-range wireless communication technologies
Car networking is obtained the location information and history routing iinformation of node by controller, constructs the Markovian model of adaptive change
Type gives a forecast to Shape Of Things To Come network using model, after constituting corresponding time diagram, is counted by the optimal road of efficient time diagram
It calculates, is most there is routing plan.The concept of time diagram is brought into traditional vehicle network by the present invention, with to temporal information
The considerations of controller in the mapping more closer to reality of vehicle network.And due to the optimal road of time diagram calculate it is born excellent
Gesture substantially increases the efficiency of router-level topology, reduces the expense of central controller.
Detailed description of the invention
Fig. 1 (a) is the time diagram of background technique citing；
Fig. 1 (b) is the corresponding static map of Fig. 1 (a)；
Fig. 2 is the overall framework figure of system；
Fig. 3 is software definition car networking framework proposed by the present invention and application scenarios；
Fig. 4 is the schematic diagram of system detailed functions；
Fig. 5 is the experimental data comparison diagram that routing calculates the time；
Fig. 6 is the experimental data comparison diagram for routing average delay.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, right below with reference to embodiment and attached drawing
The present invention is further elaborated.It should be appreciated that specific embodiment described herein is used only for explaining the present invention, and
It is not used in the restriction present invention.
Fig. 2 illustrates the overall architecture of software definition car networking in the present invention, is divided into control plane and data forwarding is flat
Face.Controlling plane is controller, and data plane is vehicle, each vehicle is equipped with GPS device to obtain geographical location.At this
In framework, vehicle periodically sends beacon message to its next hop neighbor.Next hop neighbor is received beacon and is disappeared based on beacon
The GPS position information carried in breath calculates the distance to sender.Then range information is transmitted to SDVN controller by recipient.
When controller receives route requests from vehicle, it is proposed that optimal path algorithm then be applied to effectively calculate most
Excellent routed path (flow table).When the vehicle node on Optimization route path receives grouping, corresponding forwarding work is executed, and
It needs to send controller for current routing information.This method is all based on the distance and history obtained from data plane (vehicle)
Route data.
Application environment of the invention is as shown in figure 3, pass through wireless short range communication skill in the highway vehicle node sailed of surfing the net
Art is attached, and is formed vehicle self-organizing network, the transmission of data flow is carried out between vehicle node, usually vehicle is sent to controller
Range information makes controller adaptively update Markov model to controller, and when vehicle has route requests, controller is high
Effect ground calculates optimal routing plan, and transfers to associated vehicle, while vehicle sends historical information and makes it preferably to controller
More New Network Information.
Software definition vehicle network adaptive routing method provided by the invention based on temporal information, including following step
It is rapid:
1) vehicle node periodically sends position and history routing iinformation to controller, controller according to these information architectures simultaneously
Update the Markov model of corresponding current vehicle network；
2) when controller receives the route requests of vehicle node, it is appropriate to be predicted according to the Markov model that step 1 constructs
The following routing iinformation of quantity, the input as optimal route computing method；
3) controller constructs following corresponding time diagram using the following routing iinformation predicted, and most by time diagram
Shortest path algorithm calculates optimal routing plan, and is sent to request vehicle.
Detailed process of the invention is as shown in figure 4, the process of the figure has implemented the software definition based on temporal information
Vehicle network adaptive routing, in order to clearly demonstrate process, first the expression to routing iinformation in time diagram is given a definition:
Routing link information (REI): during the grouping transmission in SDVN, the grouping sent from vehicle i is reached best
Next vehicle j in routing.The temporal information of the routing link is referred to as routing link information (REI).REI includes sender's section
Point i, recipient node j, time started t and duration d.
Detailed process shown in Fig. 4 is as follows: step 101 represents the building of Markov model: going through to make full use of
History REI constructs the prediction matrix of Markov model by each vehicle for corresponding to each state in network.We will
A possibility that being converted to subsequent time state j (currently transmit and be grouped from vehicle i to vehicle j) from current state i is from i to j
Turnover Index, by iijIt indicates.Using all Turnover Indexes between node pair each in In-vehicle networking, Markovian model is constructed
Type.History REI and distance can determine once to have or the following every two node that may be in communication with each other between calculate it is respective
Turnover Index.Distance more closely means that the success rate of transmission grouping is higher, and data rate is higher, and expense is smaller.Meanwhile history
REI shows that communication is at merits and demerits between two vehicle of past.This link is now very possible to keep stable and reliable.Certainly, it is passed through
The time crossed is longer, and history REI is more inessential.Then the weight of this REI can be smaller.Therefore, the i between i and jijIt calculates
It is as follows:
Wherein dijFor the distance between node i to node j, distance is longer, and Turnover Index is with regard to smaller, thFor time value, tp
For current time value, t0For the time value that the history routing being able to record earliest starts, longer apart from current time, then conversion refers to
Number is smaller, this history routing iinformation is more inessential,For in time thWhen, node i to a route to be valid between node j
Value, if it is in time thWhen, node i is to occurring information transmission between node j, then the value is 1, otherwise is 0.K is with l
Weight coefficient, k is bigger, then range information is more important, otherwise l bigger, and routing iinformation is more important.
Using in the history REI (for example, REI of a upper hour) and step 201 obtained in step 202 from data plane
The current distance information of the vehicle network of acquisition, can initial phase calculate network in each pair of node Turnover Index.I.e.
It only needs to execute when each time cycle starts primary.These indexes are stored in two-dimensional N N matrix, wherein dimensional table
Show current state, the state at another one-dimensional representation lower a moment.So far, the original state of Markov model has constructed completion.
Step 102: doing REI prediction using Markov model.Prediction data is mainly needed to solve two mainly to ask
Topic.Firstly, the information that requesting party issues allows for arriving at the destination (recipient of data grouping) by these predictions REI,
Accessibility i.e. between source vehicle and reception vehicle.Needing to guarantee that at least one can be extracted in the REI of all predictions has
The link road of orthochronous sequence by.Secondly, the REI quantity of prediction must be in appropriate range.In an experiment, the number of node
Amount is 4000.It is 4000 × 4000 matrixes according to Markov model, if iteratively predicting REI for each interdependent node,
The quantity of REI may can not further calculate greatly very much.If the REI quantity of prediction is seldom, accessibility not can guarantee at all.
In order to solve both of these problems, a kind of forward-backward recutrnce prediction technique based on Markov model is proposed.It uses
Pseudocode in algorithm 1 (forward prediction) and algorithm 2 (backward prediction) shows process described below.Before the projection, it creates
An empty ordered list L is built to store using the REI time started as the prediction result of sequence.Then algorithm is formally started.Firstly, time
Go through M [s], the i.e. Turnover Index of starting point s to other states.Refer to if the Turnover Index between s and j is higher than from the conversion of s
Several average value (the 5th row), it is believed that the REI is qualified and L (the 6th row and the 7th row) is inserted on this side.Then, recursively make
With forward prediction, starting point is j, and initial time is ta+dsj..And the number of iterations adds 1.Finally, reach home when router-level topology or
When the number of iterations is equal to MaxTime, this branch will reach its end.After the completion of all recurrence, which terminates, and institute
There is the REI of prediction to be stored in L by the sequence of its time started.
An appropriate number of REI can be generated in forward prediction.But accessibility is not can guarantee.In this case, from terminal
To starting point backward prediction REI.In this way, if positive two parts REI with backward prediction can be suitable with the correct time
It is linked together in sequence, then certainly exists the routing of at least one connection by these REI.Therefore, algorithm for design 2 is (reversed pre-
Survey) solve this problem.
The main body of the backward prediction algorithm is similar to forward prediction algorithm.Compared with the latter, in order to increase by two nodes it
Between accessibility, there are two main differences.1) backward prediction is since terminal (destination node).Each recurrence is all by previous step
Starting point as terminal.Such as backward prediction REI from finish to start.Its termination condition is to calculate to reach starting point or iteration time
Number is equal to MaxTime, identical as forward prediction.2) being arranged using d is the termination time of the upper a line of terminal (at the beginning of side
+ its duration).Then this algorithm is executed to postorder.Once encountering a line, starting point is one be stored in L
While terminal (it is worth noting that, in L while be side that we predict in forward prediction), at the beginning of this side
It is adjusted to the minimum end time at involved edge in L.Due to being postorder ground recurrence this algorithm, prediction in front can also be from
Temporal information is obtained in subsequent prediction (closer to the prediction of s).In this approach, by forward prediction algorithm and backward prediction
The REI of algorithm prediction is linked together.By running both algorithms and adjust MaxTime value, guarantee accessibility and appropriate
REI amount.
Step 103: the optimal road algorithm of time diagram；After predicting following enough REI, using based on temporal information
Optimal path algorithm calculates the optimal routings of needs.As for the optimal path of time diagram, Wu, Huanhuan propose four kinds most
Short path algorithm.Herein, select the optimal road algorithm of first calculating earliest arrival time, it and route mesh having the same
Mark.The main thought of the algorithm is as described below.Firstly, according to the order traversal side of time started.It is real during prediction
This target is showed.L is the REI sequence to sort by its time started.It is every to record and manage that one list t [N] is then set
The earliest arrival time of a node.In this way, earliest arrival time can be obtained, and tracks entire routing.Finally, inspection
Look into each side e (u, v, t, d) in time diagram.If be later than at the beginning of e the current earliest arrival time of u and at the end of
Between the current earliest arrival time of (s+d) earlier than v, then update the earliest arrival time of v.It traverses, will not miss in chronological order
Any possible Optimization route.
The main problem of the algorithm is input, it is the edge sequence using the time started as sequence.It will during prediction
It is embodied as L.But other than earliest arrival time, it is also necessary to optimal path nodal information.It therefore, is every in algorithm
A node provides a list, to record the upper hop node on optimal path.Because there may be on multiple on optimal path
One hop node.There are multiple optimal paths of earliest arrival time having the same.Therefore, multiple upper hop nodes will be recorded
Rather than one, after algorithm, the earliest arrival time and its corresponding upper one of each node on optimal path will be obtained
Hop node.When scanning previous dive node listing, sequence node can be easily obtained on optimal path.Then in algorithm 3
The details of middle Representation algorithm.
1st row and the 2nd row are used to initialize the earliest arrival time and upper hop node of each node: using array t [N]
It is indicated with prev [N].Side is filtered for the first time in the 4th row.If this edge leaves u before the earliest arrival time of u, no
It can consider this edge.Then, if earliest arrival time of the end time at this edge earlier than v, when will update arrival earliest
Between and rebuild upper hop node.The reason of reconstruction is that this may be a completely new optimal routing (the 5th row to eighth row).If
End time is equal to earliest arrival time, only prev [N] need to be added in this edge.Finally, terminating this after scanning through all sides
Algorithm.T [N] and prev [N] is Optimization route result.
Finally propose the optimization to Markov model, the calculation amount in algorithm is concentrated mainly on the structure of Markov model
If making controller receives route requests, it will calculate specific Markov model for current whole network.Algorithm becomes too
It is complicated.Need to be adaptively adjusted original Markov model.Therefore, it is proposed to a kind of improved method.At the beginning, similar
In conventional model, Markov model is constructed；Prediction REI simultaneously calculates Optimization route.The difference is that in new method,
Once constructing Markov model, a completely new model need not be just constructed when up-to-date information occurs whithin a period of time.It can
To update the relevant information of Markov model, rather than deletes past model and create a new model.Once Ma Erke
Husband's model obtains the distance between vehicle u and vehicle v information update, and Turnover Index is as updating as defined in (2).
iuvTarget Turnover Index between the u and v of expression.i′uvIt is the Turnover Index calculated according to formula (1), it is
It is out-of-date.K and l is the weight parameter of above-mentioned distance and historical data.duvWith d 'uvIt is current between u and v and goes through
History range information.Transition index is adaptively adjusted according to the distance of current and past.Similarly, the grouping when generation from u to v
When transmission, Turnover Index is adjusted, as described in (3).
De refers to the decline coefficient in 0 and 1 range.It is unwise for allowing a link to take over excessive transmission work
, this may cause unbalanced load.Therefore, go in the timestamp that grouping transmission occurs between u and v every time
Afterwards, due to being not intended to same side to undertake multi information transformation task, so making the importance on the side of u to v reduces, Turnover Index
Also corresponding decline.Finally, the Turnover Index average value of u needs corresponding change.
In order to guarantee accuracy, after being spaced at a fixed time, restart a process, according to newest distance and
REI recalculates Markov model.These data are via control planar collecting.
Embodiment,
The details of experiment, including parameter and assessment will be introduced, Intel 2.4GHz CPU and 16 GB RAM is being housed
10 system of Windows PC machine on run all experiments.According to the variation of vehicle location, constructed with C++ programming language
SDVN emulation platform.Location information is generated by SUMO.In addition, history REI is generated under GPSR agreement by NS-2.In simulations,
Consider four kinds of different number of nodes.The major parameter of vehicle network is described in table 1.
| Parameter | Numerical value |
| Vehicle fleet size | 500/1000/2000/4000 |
| Car speed | 1-60km/h |
| Vehicle moving range | 1200km*850km |
| Mac-layer protocol | IEEE 802.11p |
| It is grouped transmission range | 250m |
| Test duration | 100s |
TibAR is compared with dijkstra's algorithm.Send the route requests of random vehicles to once per second
SDVN controller.It is main to consider two performance parameters to assess the performance of TibAR.Firstly, consider the average calculation times of routing,
Refer to that controller is completed to calculate the average time of the optimal routing of each request, including building Markov model and update
Time.Secondly, considering the average delay of routing, the average retardation of each routing of source node to other each nodes is referred to
(delay class herein is indicated using relative value).
As shown in figure 5, the difference in low-density traffic scene (500 nodes), between TibAR and the efficiency of Dijkstra
It is different to be not obvious.However, with the increase of node, TibAR is gradually better than Dijkstra.Under best-case, with Dijkstra phase
Than the efficiency of TibAR is its six times, this may be the increase with node, and Dijkstra number of edges amount to be treated is with O
(n2) speed growth.Meanwhile in TibAR, by applying proposed forward prediction and backward prediction algorithm, edge is scanned
Quantity is in always in appropriate range.As shown in fig. 6, dijkstra's algorithm calculates the absolute optimal road in vehicle network
Diameter, and Dijkstra is better than TibAR in the performance in terms of routing delay.However, with the increase of node, TibAR can be
More training datas are obtained in building process, to reduce time delay.It can be seen that TibAR is produced when number of nodes reaches 4000
Raw routing delay is very close to absolute optimal path.Observe computational efficiency and road of the TibAR in high density flow scene
Existing primary circuit routing scheme SDVN is better than by delay aspect.
Claims (5)
1. a kind of software definition vehicle network adaptive routing method based on temporal information, which is characterized in that including walking as follows
It is rapid:
1) vehicle node periodically sends position and history routing iinformation to controller, and controller is according to these information architectures and updates
The Markov model of corresponding current vehicle network；
2) when controller receives the route requests of vehicle node, right quantity is predicted according to the Markov model that step 1 constructs
The following routing iinformation, the input as optimal route computing method；
3) controller constructs following corresponding time diagram using the following routing iinformation predicted, and passes through the optimal road of time diagram
Diameter algorithm calculates optimal routing plan, and is sent to request vehicle.
2. the software definition vehicle network adaptive routing method based on temporal information, feature exist as described in claim 1
In utilizing the Turnover Index i between node pair each in In-vehicle networking in step 1)ijTo construct Markov model prediction square
Battle array:
Wherein, dijFor the distance between node i to node j, distance is longer, and Turnover Index is with regard to smaller, thFor time value, tpTo work as
Preceding time value, t0For the time value that the history routing being able to record earliest starts, longer apart from current time, then Turnover Index is got over
Small, this history routing iinformation is more inessential,For in time thWhen, node i is to a route to be valid value between node j, such as
Fruit is in time thWhen, node i is to occurring information transmission between node j, then the value is 1, and on the contrary is 0, k and l is weight
Coefficient, k is bigger, then range information is more important, otherwise l bigger, and routing iinformation is more important；
A Turnover Index is calculated when each time cycle starts, and is stored it in two-dimensional N N matrix, wherein one-dimensional
Indicate current state, the state at another one-dimensional representation lower a moment, so far the original state of Markov model, which constructs, completes.
3. the software definition vehicle network adaptive routing method based on temporal information, feature exist as described in claim 1
In, in step 2), the Markov model constructed using step 1 carries out the prediction of the following routing iinformation, including forward prediction and
Backward prediction:
Forward prediction:
A) an empty ordered list L is created；
B) M [s] is traversed, M [s] is Turnover Index of the starting point s to other states, if the Turnover Index between s and j is higher than from s
The average value of the Turnover Index to set out, then this routing is qualified, and L is inserted on this side；
C) forward prediction is recursively used, starting point is j, and initial time is ta+dsj, and the number of iterations adds 1:
D) when router-level topology is reached home or the number of iterations is equal to MaxTime, this branch will reach its end, in all recurrence
After completion, which terminates, and the routing of all forward predictions is stored in L by the sequence of its time started；
Backward prediction:
Difference with forward prediction step is: since terminal, each recurrence all using the starting point of previous step as terminal, i.e., from
Terminate to backward prediction routing is started, termination condition is to calculate arrival starting point or the number of iterations equal to MaxTime, pre- with forward direction
It surveys identical；
It is arranged using d as the termination time of the upper a line of terminal, then executes backward prediction to postorder, once encounter a line
Starting point is the terminal for a line being stored in L, then involved edge will be adjusted in L at the beginning of this side most
The small end time together with the routing link for thus predicting forward prediction algorithm and backward prediction algorithm, obtains right quantity
The following routing iinformation.
4. the software definition vehicle network adaptive routing method based on temporal information, feature exist as claimed in claim 3
In being included the following steps: in step 3) by the method that time diagram optimal path algorithm calculates optimal routing
First, in accordance with the side e (u, v, t, d) in the order traversal L of time started, the earliest arrival time of each node is initialized
It with upper hop node, is indicated with array t [N] and prev [N], then carries out filter side, if this edge is in the earliest arrival of u
Between before leave u, then do not consider this edge, if earliest arrival time of the end time on this side earlier than v, update arrive earliest
Up to the time and upper hop node is rebuild, if the end time is equal to earliest arrival time, prev [N] is added in this edge；Finally,
Terminate this algorithm after scanning through all sides, t [N] and prev [N] are Optimization route results.
5. the software definition vehicle network adaptive routing method based on temporal information, feature exist as claimed in claim 2
In when controller obtains the distance between vehicle u and vehicle v information update, Turnover Index updates are as follows:
Wherein, iuvIndicate the target Turnover Index between u and v, i 'uvIt is the Turnover Index calculated according to formula (1), duvWith d 'uv
It is the current and history range information between u and v, transition index is adaptively adjusted according to the distance of current and past, it is similar
Ground adjusts Turnover Index when the grouping transmission from u to v occurs, as described in formula (3):
Wherein de refers to the decline coefficient in 0 and 1 range, in the time that grouping transmission occurs between u and v every time
After the stamp past, due to being not intended to same side to undertake multi information transformation task, so making the importance on the side of u to v reduces, turn
Index also corresponding decline is changed, finally, the Turnover Index average value of u needs corresponding change:
In order to guarantee accuracy, after being spaced at a fixed time, formula (2), formula (3), formula (4) process are restarted,
Markov model is recalculated according to newest distance and routing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811609539.1A CN109511123B (en) | 2018-12-27 | 2018-12-27 | Software-defined vehicle network adaptive routing method based on time information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811609539.1A CN109511123B (en) | 2018-12-27 | 2018-12-27 | Software-defined vehicle network adaptive routing method based on time information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109511123A true CN109511123A (en) | 2019-03-22 |
| CN109511123B CN109511123B (en) | 2022-01-14 |
Family
ID=65755332
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811609539.1A Active CN109511123B (en) | 2018-12-27 | 2018-12-27 | Software-defined vehicle network adaptive routing method based on time information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109511123B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110049527A (en) * | 2019-04-19 | 2019-07-23 | 厦门大学 | Method and system based on the routing of software definition car networking |
| CN110442624A (en) * | 2019-08-09 | 2019-11-12 | 沈阳航空航天大学 | Service time limited polymerization nearest Neighbor in a kind of Time Dependent road network |
| CN110855507A (en) * | 2019-11-28 | 2020-02-28 | 湖南率为控制科技有限公司 | Unmanned data network interaction method based on software definition |
| CN111130853A (en) * | 2019-12-04 | 2020-05-08 | 沈阳航空航天大学 | Future route prediction method of software defined vehicle network based on time information |
| CN113207124A (en) * | 2021-03-17 | 2021-08-03 | 北京邮电大学 | Vehicle-mounted ad hoc network data packet transmission method and device |
| CN114866461A (en) * | 2022-04-28 | 2022-08-05 | 抖动科技(深圳)有限公司 | RTC (real time clock) streaming media self-adaptive transmission method, device, equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040192341A1 (en) * | 2002-12-26 | 2004-09-30 | Shu-Shaw Wang | Apparatus, and an associated method, for reserving resources in a mobile communication system through the use of historical indicia |
| CN103702387A (en) * | 2014-01-08 | 2014-04-02 | 重庆邮电大学 | Social network-based vehicle-mounted self-organization network routing method |
| CN103945486A (en) * | 2014-04-02 | 2014-07-23 | 北京理工大学 | Multi-objective optimization data transmission method based on mobile node locus |
| CN104640168A (en) * | 2014-12-04 | 2015-05-20 | 北京理工大学 | Q-learning based vehicular ad hoc network routing method |
| CN107147582A (en) * | 2017-04-26 | 2017-09-08 | 沈阳航空航天大学 | A kind of vehicle self-organizing network mixed type Generating Routing |
-
2018
- 2018-12-27 CN CN201811609539.1A patent/CN109511123B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040192341A1 (en) * | 2002-12-26 | 2004-09-30 | Shu-Shaw Wang | Apparatus, and an associated method, for reserving resources in a mobile communication system through the use of historical indicia |
| CN103702387A (en) * | 2014-01-08 | 2014-04-02 | 重庆邮电大学 | Social network-based vehicle-mounted self-organization network routing method |
| CN103945486A (en) * | 2014-04-02 | 2014-07-23 | 北京理工大学 | Multi-objective optimization data transmission method based on mobile node locus |
| CN104640168A (en) * | 2014-12-04 | 2015-05-20 | 北京理工大学 | Q-learning based vehicular ad hoc network routing method |
| CN107147582A (en) * | 2017-04-26 | 2017-09-08 | 沈阳航空航天大学 | A kind of vehicle self-organizing network mixed type Generating Routing |
Non-Patent Citations (2)
| Title |
|---|
| ISABEL V. MARTIN-FAUS，LUIS URQUIZA-AGUIAR，ETC.: "Transient Analysis of Idle Time in VANETs Using Markov-Reward Models", 《IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY ( VOLUME: 67, ISSUE: 4, APRIL 》 * |
| 朱东杰: "车载自组网路由协议及路径决策模型研究", 《中国博士学位论文全文数据库工程科技Ⅱ辑（月刊）》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110049527A (en) * | 2019-04-19 | 2019-07-23 | 厦门大学 | Method and system based on the routing of software definition car networking |
| CN110049527B (en) * | 2019-04-19 | 2021-09-03 | 厦门大学 | Method and system for defining Internet of vehicles routing based on software |
| CN110442624A (en) * | 2019-08-09 | 2019-11-12 | 沈阳航空航天大学 | Service time limited polymerization nearest Neighbor in a kind of Time Dependent road network |
| CN110442624B (en) * | 2019-08-09 | 2022-12-27 | 沈阳航空航天大学 | Aggregation neighbor query method with limited service time in time-dependent road network |
| CN110855507A (en) * | 2019-11-28 | 2020-02-28 | 湖南率为控制科技有限公司 | Unmanned data network interaction method based on software definition |
| CN110855507B (en) * | 2019-11-28 | 2022-04-22 | 湖南率为控制科技有限公司 | Unmanned data network interaction method based on software definition |
| CN111130853A (en) * | 2019-12-04 | 2020-05-08 | 沈阳航空航天大学 | Future route prediction method of software defined vehicle network based on time information |
| CN111130853B (en) * | 2019-12-04 | 2023-05-02 | 沈阳航空航天大学 | Future route prediction method of software defined vehicle network based on time information |
| CN113207124A (en) * | 2021-03-17 | 2021-08-03 | 北京邮电大学 | Vehicle-mounted ad hoc network data packet transmission method and device |
| CN113207124B (en) * | 2021-03-17 | 2022-11-01 | 北京邮电大学 | Vehicle-mounted ad hoc network data packet transmission method and device |
| CN114866461A (en) * | 2022-04-28 | 2022-08-05 | 抖动科技(深圳)有限公司 | RTC (real time clock) streaming media self-adaptive transmission method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109511123B (en) | 2022-01-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109511123A (en) | A kind of software definition vehicle network adaptive routing method based on temporal information | |
| CN103945486B (en) | Multiple-objection optimization data transferring method based on mobile node track | |
| CN104080056B (en) | The message distributing method of the vehicular ad hoc network perceived based on degree of communication probability | |
| Zhao et al. | A novel prediction-based temporal graph routing algorithm for software-defined vehicular networks | |
| CN104640168B (en) | Vehicular ad hoc network method for routing based on Q study | |
| CN111130853B (en) | Future route prediction method of software defined vehicle network based on time information | |
| Zhao et al. | A novel adaptive routing and switching scheme for software-defined vehicular networks | |
| CN105208616A (en) | Road topology based adaptive multi-copy routing method in vehicular ad hoc network | |
| CN109413707B (en) | Intelligent routing method based on deep reinforcement learning technology in wireless network environment | |
| CN103702387A (en) | Social network-based vehicle-mounted self-organization network routing method | |
| CN105245608A (en) | Telematics network node screening and accessibility routing construction method based on self-encoding network | |
| CN108519737B (en) | Unmanned equipment path planning method considering energy supply | |
| CN105101086B (en) | A kind of data transfer path system of selection based on traffic density distribution | |
| Zhao et al. | A temporal-information-based adaptive routing algorithm for software defined vehicular networks | |
| Zhao et al. | Learning based massive data offloading in the iov: Routing based on pre-rlga | |
| CN109275154B (en) | Dynamic self-adaptive routing path planning method based on double-layer topological routing algorithm | |
| CN106161237A (en) | Centralized method for routing, device and system | |
| CN106060888A (en) | VANET clustering routing method based on complex network centrality | |
| Li et al. | An efficient reinforcement learning based charging data delivery scheme in VANET-enhanced smart grid | |
| Du et al. | GAQ-EBkSP: a DRL-based urban traffic dynamic rerouting framework using fog-cloud architecture | |
| CN102065446B (en) | Topology control system and method orienting group mobile environment | |
| Pal et al. | Biogeographic-based temporal prediction of link stability in mobile ad hoc networks | |
| CN113382382A (en) | Vehicle ad hoc network routing method based on fuzzy logic optimization strategy | |
| CN102843301B (en) | Routing method for mobile delay tolerant network based on contact predication | |
| CN104835316B (en) | Traffic flow density-based solution to problem of VANET sparse connectivity |
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 |