CN105307232A - Routing optimization method for vehicular self-organized network based on connection probabilities - Google Patents

Abstract

The invention discloses a routing optimization method for a vehicular self-organized network based on connection probabilities. The method comprises the following steps that: a gateway, where a source node is, calculates connection probabilities on various lanes in a corresponding network at first; then, a path having the highest connection probability is selected through intercommunication between gateways; the id number of the gateway in the selected path is recorded and added into the head part of a data packet to be transmitted; practical transmission of the data packet is waited; and a practical data packet forwarding process is carried out after the path is selected. According to the invention, the next-hop selection mechanism based on the priority is provided; compared with the original greedy algorithm, the back-off rate is reduced; and the routing effectiveness is greatly increased.

Description

A kind of routing optimization method based on connected probability of vehicular ad hoc network

Technical field

The present invention relates to a kind of routing optimization method based on connected probability of vehicular ad hoc network, belong to communication technical field.

Background technology

Because the factors such as the high-speed mobile of vehicle and the uncertainty of vehicle fleet size can cause the frequent variations of network topology, therefore location-based Routing Protocol is more suitable for vehicular ad hoc network.Position Routing Protocol (GeographicalRoutingProtocol) is a kind of typical In-vehicle networking agreement.It carries out route by the geographical location information introducing node.Usually, the geographical location information of each node by using GPS or other positioners to obtain oneself, in network, each node only needs the geographical location information of the neighbors known in its communication radius, and Route establishment just can complete by means of only several single-hop topology information.Therefore when source node to the transfer of data of destination node only needs to know the geographical position of destination node and each data retransmission, the geographical position of next-hop node just can realize, and does not need other topology information.

The down hop of existing position Routing Protocol selects machine-processed algorithm cardinal principle to be, when forward node and destination node are in same signaling zone, but forward node does not have neighbor node in this signaling zone, GRP does not select any neighbors, is reselected by packet rollback to upstream node.Obvious this mechanism is existing defects.When there is the neighbors of present node the periphery adjacent area of destination node, and this neighbors can one jump to when reaching destination node, then select the neighbors of destination node adjacent area as down hop, otherwise do not consider any neighbors and carry out the operation of return data to upper hop node.When forward node and destination node be not in same signaling zone, and in the signaling zone of destination node, have the neighbor node of forward node, then Stochastic choice neighbor node carries out the forwarding of down hop.This obvious existing defects, it is also proposed the method for improvement equally.First to calculate in destination node adjacent area all neighborss to the distance of destination node, then sort and select from the nearest neighbors of destination node as next-hop node.The next-hop node that such sensor selection problem is more suitable for, decreases and forwards jumping figure and time delay, improve delivery ratio.

But the high-speed mobile due to node causes the information of routing table can accurately not reflect the positional information of neighbor node, and due to the finiteness of Radio Resource, the update cycle of neighbor table can not arrange too small.As shown in Figure 1, in straight road, the communication distance of sending node is 200m, the cycle arranging node transmission beacon frame in Routing Protocol is 1s, namely the positional information that the neighbor table in sending node maintains is that 1s upgrades once, the update time of last beacon frame is t, as moment t, sending node will send data, t '-t<1s, get t-T=0.5s, the speed of neighbor node is 20m/s, if the coordinate position distance sending node in neighbor table is 195m, now sending node is 205m apart from the distance of this node, not in the communication range of sending node, at this moment the node following GRP can according to this packet of record rollback of rollback table to upstream node, upstream node can reselect down hop, if failure can return back to again previous node and carry out reselecting of down hop, circulation like this, until return back to source node, then this Packet Generation failure.So, down hop selection mechanism only rely on greedy algorithm be the transmitting that can not ensure packet.And the present invention can solve problem above well.

Summary of the invention

The object of the invention is to solve above-mentioned existing technical problem, propose a kind of routing optimization method based on connected probability of vehicular ad hoc network, first the method is calculated the connected probability on each track in corresponding network by the gateway at source node place, then by the selection path that connected probability is the highest that intercoms mutually between gateway, and the gateway in selected path No. id and track number are recorded, add the stem of data to be transferred bag to, wait for that packet carries out actual transmissions.Path selection well just enters actual packet repeating process afterwards, and the down hop that the present invention proposes based on priority selects mechanism, and contrast and original greedy algorithm, this mechanism decreases rollback probability, substantially increases the validity of route.

The present invention solves the technical scheme that its technical problem takes: a kind of routing optimization method based on connected probability of vehicular ad hoc network, the method comprises the steps:

Step 1: suppose that the intersection of road is configured with gateway, each gateway maintains and is updated periodically the density of the vehicle node in corresponding road section on different track, travel speed etc., the half subsequently track being divided into communication distance is the section of unit, if each section there is a vehicle, then path is communicated with.And use the mathematical method of permutation and combination to extrapolate assistance connected probability between three different tracks.

Step 2: according to the assistance connected probability of the three lanes that above-mentioned steps 1 provides, communicated by wired mode between gateway, finally select the highest path of connected probability between a source node and a destination node, this path is specific to track, subsequently this route messages is turned back to the packet of source node, and the head gateway number in selected path and track number being stored in packet carries out actual forwarding.

Step 3: in the repeating process of reality, namely in the mechanism of the selection of packet down hop, source node forwards based on the priority of neighbor node in neighbor table.

Step 4: about the algorithm of priority, the present invention is that neighbor node is relative and the position of sending node, speed and direction take into account, and done different Algorithm Analysis for Through Lane and two kinds, crossroad road conditions, finally draw the priority of each neighbor node.

Step 5: sending node selects the high neighbor node of priority as next-hop node.

Beneficial effect:

1, the path selection mechanism of Routing Protocol of the present invention must consider the connectedness of car and car, owing to can carry out facilitating communications between track, so assistance connected probability of the present invention is more reliable.

2, the present invention is directed to the actual conditions of vehicular ad hoc network, improve position Routing Protocol GRP, routing strategy considers the connected probability between car and car, research scene is the dual three-lane carriageway highway tallied with the actual situation, and has write out the assistance connected probability expression formula in each track based on this scene.The traffic density emulating communication range and the track showing vehicle is higher, assists connected probability larger, and this is improve the reliability of assisting connected probability namely to communicate in track to provide strong foundation.

3, gateway of the present invention selects track that in three tracks, connected probability is maximum as routed path, this based on assisting the choosing lane of connected probability machine-processed refinement routing mechanism, reduce the link down probability caused due to the high-speed motion of vehicle, ensure that the reliability of communication.

4, present invention improves over the down hop forwarding strategy of geographic routing agreement, no longer greedy algorithm is relied on when selecting, but distribute priority according to the position of vehicle, speed messages to the neighbor node of sending node, the high neighbor node of priority is selected to forward, owing to there is uncertainty, so also will D-factor be considered in the distribution of priority near the traveling behavior of crossroad vehicle.

Accompanying drawing explanation

Fig. 1 is three lanes two-way road model schematic.

Fig. 2 is road model schematic diagram of the present invention.

Fig. 3 is the probability graph of the spacing of Adjacent vehicles of the present invention.

Fig. 4 (a) is the schematic diagram of moment Timestamp1, and Fig. 4 (b) is the schematic diagram of moment Timestamp2.

Fig. 5 is method flow diagram of the present invention.

Embodiment

Below in conjunction with Figure of description, the invention is described in further detail.

As shown in Figure 2, the present invention, in three lanes two-way road, supposes that the vehicle on each track carries out uniform motion, and car speed corresponding on track 1,2,3 is respectively V1, V2, V3.Each section is divided into the segment that length is communication distance half by the present invention, and gateway maintains car speed and the traffic density in path L in map network and each track.The node location in each segment is rule of thumb worth to be that obedience is equally distributed.Suppose that the node arriving amt on three tracks all obeys Poisson distribution, and density is respectively K1, K2, K3 (unit is/meter).Node on each segment then on track 1 is independent identically distributed, and is that to obey parameter be the Poisson distribution of (R/2) * K1, i.e. interstitial content N on the segment in three tracks _i(i=1,2,3) are independent and obey the Poisson distribution that parameter is (R/2) * Ki.So the interstitial content in each segment of three lanes is

$p(N_i=n_i)=\frac{{\left(\right(R/2)*Ki)}^{n_i}}{n_i!}{e}^{-(R/2)*Ki},i=1,2,3.$ Formula 1

Distance between Fig. 2 interior joint 1 and node 2 is greater than communication distance R, can not complete the communication between car and car, but can be communicated with the assistance of node 4 by node 3.Regard three tracks as an entirety, obviously because every little segment distance is R/2, as long as so ensure to have a node in each segment at least, just can ensure the connectedness (namely ignoring workshop lateral separation) of network.

In Fig. 2 due in track 2 without node, facilitating communications (segment namely in track 3 is independent identically distributed) can only be carried out, so the connected probability that can obtain node 1 and node 2 is by the node in track 3:

$P_C={(1-P(N_3=0\left)\right)}^2=e^{-R*K_3}$ Formula 2

Reasoning (only considers that track 1 completes situation about being communicated with by track 2 with the assistance in track 3 here) under common situation: if disconnected adjacent two workshops on track 1 are every two segments (i.e. n=2), the node on track 1 can be made by track 2 to be communicated with the assistance in track 3.

If n=2, for:

$P_C=C_2^2{\&lsqb;1-f(N_2=0)\&rsqb;}^2+C_2^1\&lsqb;1-f(N_2=0)\&rsqb;\&lsqb;1-f(N_3=0)\&rsqb;+C_2^0{\&lsqb;1-f(N_3=0)\&rsqb;}^2$ Formula 3

If n=3, for:

$P_C=C_3^3{\&lsqb;1-f(N_2=0)\&rsqb;}^3+C_3^2{\&lsqb;1-f(N_2=0)\&rsqb;}^2\&lsqb;1-f(N_3=0)\&rsqb;+C_3^1\&lsqb;1-f(N_2=0)\&rsqb;{\&lsqb;1-f(N_3=0)\&rsqb;}^2+C_3^0{\&lsqb;1-f(N_3=0)\&rsqb;}^3$ Formula 4

When reasoning show that n equals arbitrary value k,

$P_C=\underset{i=0}{\overset{k}{\&Sigma;}}{C}_k^i{\&lsqb;1-f(N_2=0)\&rsqb;}^i{\&lsqb;1-f(N_3=0)\&rsqb;}^{k-i}$ Formula 5

Note Xi is the distance on track 1 between adjacent two cars, and the segment number between Adjacent vehicles is n=Xi/ (R/2), for:

$P_C=\underset{i=0}{\overset{Xi/(R/2)}{\&Sigma;}}{C}_{X_i/R/2}^i{\&lsqb;1-f(N_2=0)\&rsqb;}^i{\&lsqb;1-f(N_3=0)\&rsqb;}^{Xi/(R/2)-i}$ Formula 6

Distance X between vehicle adjacent in track 1 ₁be obey parameter be the exponential distribution of traffic density K1, that is: $F_{X_1}\left(x\right)=1-e^{-K1*x}$ Formula 7

Suppose that the traffic density in track 1 is 0.008/meter, then the spacing of Adjacent vehicles is as Fig. 3:

The probability that distance as seen from Figure 3 between Adjacent vehicles is greater than 400m is about 1, gets 400m and 600m.When distance between adjacent two cars exceedes connection distance, two cars are not just communicated with.X ₁obeying parameter is the exponential distribution of K1, so the outage probability in track 1 is as follows:

P _d=P (X ₁> R)=e ^-R*K1formula 8

Consider and track 1 may have multilink to be disconnected, if disconnected link number is q, node has Q, then q ∈ 1,2 ... ..Q-1.Then track 1 has q bar link as follows by the probability assisting to be communicated with:

$P_{cq}=\underset{i=1}{\overset{q}{\&Pi;}}{P}_C$ Formula 9

If when not considering the facilitating communications between track, track 1 there is the disconnected probability of q bar link be $P_q=C_{Q-1}^q{P_d}^q{(1-P_d)}^{Q-1-q}$ Formula 10

Then the connected probability in track 1 is:

$P_{1C}=\underset{q=0}{\overset{Q-1}{\&Sigma;}}{P}_{cq}*P_q$ Formula 11

In like manner can calculate the connected probability in track 2 and track 3 between two gateways, be respectively P _2Cand P _3C.Gateway is selected to assist the highest track of connected probability to carry out the forwarding of data, i.e. P _s=Max{P _1Cp _2Cp _3C.Because the path from source node to destination node is made up of the track that n has been selected, then the connected probability of whole routed path is

$P=\underset{j=1}{\overset{n}{\&Pi;}}{P}_s$ Formula 12

Communicated by cable network between gateway with gateway, the gateway at source node place knows the connected probability on each track, and actual forwarding is carried out in the path that then selection track connected probability is the highest.

And in actual repeating process, due to the fast moving of node, only rely on greedy algorithm cannot meet connective requirement, so the present invention proposes a kind of optimized algorithm.Owing to there is no speed and directional information in neighbor table, only time and positional information.So by calculating distance, the neighbor node speed of each neighbor node distance sending node, and the information of these two fields must be increased in neighbor table, as table 1.Can calculate this three field values in neighbor table updated time, computational methods are as follows:

Neighbor table after table 1 improves

Type	Field	Explanation
			InetT_Address	Nbr_addr	IP address and type
double	Nbr_lat	Node longitude
			double	Nbr_long	Node latitude
double	timestamp	Update time
			double	timeout	Expired time
double	position	Node location
			double	velocity	Node speed

The longitude and latitude position that source node of the present invention obtains neighbor node from neighbor table is: (nbri_lat, nbri_long), the position that can be obtained self by GPS is: (self_lat, self_long), can calculate each neighbor node apart from the distance of self is:

$L_i=\sqrt{{(self\_lat-nbri\_lat)}^2+{(self\_long-nbri\_long)}^2}$ Formula 13

In order to reduce jumping figure, packet being transmitted to the node apart from self distance as far as possible, so give the priority that the peer distribution of distance is higher, obtaining the distance factor a of priority _ifor:

A _i=R/L _iformula 14

Because node high-speed mobile can increase offered load, and then may link failure be caused, so select slow-footed node to forward as far as possible, namely give the forwarding priority that slow-footed node is high.Source node can obtain the longitude and latitude (supposing uniform motion) of the neighbor node in timestamp1 and timestamp2 moment from its neighbor table, be respectively (nbri_lat1, and (nbri_lat2 nbri_long1), nbri_long2), the speed obtaining neighbor node i is as follows:

$V_i=\frac{\sqrt{{(nbri\_lat2-nbri\_lat1)}^2+{(nbri\_long2-nbri\_long1)}^2}}{timestamp2-timestamp1}$ Formula 15

Suppose V _maxbe the speed limit on selected track, obtain the velocity factor b of priority _ifor:

B _i=V _max/ V _iformula 16

So the priority assign of the neighbor node on straight road is:

S ₁=a _i* b _iformula 17

The present invention considers the uncertainty in vehicle future travel direction at the parting of the ways, so the selection of the next-hop node of crossroad vehicle is different from the selection on straight road.As shown in Fig. 4 (a), the scene in the Timestamp1 moment is represented in figure, if node 3 can be selected to forward according to greedy algorithm, but become as shown in Fig. 4 (b) in Timestamp2 moment scene, because node 1,2,3 certain moment respectively in [timestamp1, timestamp2] has carried out left-hand rotation, right-hand rotation, craspedodrome.When actual forwarding data, node 3 is not best selection like this, and node 1 is only best selection.So except considering distance, speed in down hop selection algorithm at the parting of the ways, also will consider the direction of motion of neighbor node and the source node angle to destination node, the neighbor node that angle is little can have high priority.

Claims (4)

1. the routing optimization method based on connected probability of vehicular ad hoc network, is characterized in that, described method comprises the steps:

Step 1: suppose that the intersection of road is configured with gateway, each gateway maintains and is updated periodically the density of the vehicle node in corresponding road section on different track, travel speed etc., the half subsequently track being divided into communication distance is the section of unit, if each section there is a vehicle, then path is communicated with; The mathematical method of permutation and combination is used to extrapolate assistance connected probability between three different tracks;

Step 2: according to the assistance connected probability of the three lanes that above-mentioned steps 1 provides, communicated by wired mode between gateway, finally select the highest path of connected probability between a source node and a destination node, this path is specific to track, subsequently this route messages is turned back to the packet of source node, and the head gateway number in selected path and track number being stored in packet carries out actual forwarding;

Step 3: in the repeating process of reality, namely in the mechanism of the selection of packet down hop, source node forwards based on the priority of neighbor node in neighbor table;

Step 4: sending node selects the high neighbor node of priority as next-hop node.

2. the routing optimization method based on connected probability of a kind of vehicular ad hoc network according to claim 1, it is characterized in that, described method comprises: the algorithm of priority, be by neighbor node relatively and the position of sending node, speed and direction take into account, and done different Algorithm Analysis for Through Lane and two kinds, crossroad road conditions, finally draw the priority of each neighbor node.

3. the routing optimization method based on connected probability of a kind of vehicular ad hoc network according to claim 1, is characterized in that, described connected probability is:

$P_C={(1-P(N_3=0\left)\right)}^2=e^{-R*K_3}.$

4. the routing optimization method based on connected probability of a kind of vehicular ad hoc network according to claim 1, it is characterized in that, if disconnected adjacent two workshops on the track of described method are every two segments (i.e. n=2), by the assistance in track and track, the node on track is communicated with;

If n=2, for:

$P_C=C_2^2{\&lsqb;1-f(N_2=0)\&rsqb;}^2+C_2^1\&lsqb;1-f(N_2=0)\&rsqb;\&lsqb;1-f(N_3=0)\&rsqb;+C_2^0{\&lsqb;1-f(N_3=0)\&rsqb;}^2$ Formula 3

If n=3, for:

$P_C=C_3^3{\&lsqb;1-f(N_2=0)\&rsqb;}^3+C_3^2{\&lsqb;1-f(N_2=0)\&rsqb;}^2\&lsqb;1-f(N_3=0)\&rsqb;+C_3^1\&lsqb;1-f(N_2=0)\&rsqb;{\&lsqb;1-f(N_3=0)\&rsqb;}^2+C_3^0{\&lsqb;1-f(N_3=0)\&rsqb;}^3$ Formula 4

When reasoning show that n equals arbitrary value k, for:

$P_C=\underset{i=0}{\overset{k}{\&Sigma;}}{C}_k^i{\&lsqb;1-f(N_2=0)\&rsqb;}^i{\&lsqb;1-f(N_3=0)\&rsqb;}^{k-i}$ Formula 5

Note Xi is the distance on track 1 between adjacent two cars, and the segment number between Adjacent vehicles is n=Xi/ (R/2), for:

$P_C=\underset{i=0}{\overset{Xi/(R/2)}{\&Sigma;}}{C}_{X_i/R/2}^i{\&lsqb;1-f(N_2=0)\&rsqb;}^i{\&lsqb;1-f(N_3=0)\&rsqb;}^{Xi/(R/2)-i}$ Formula 6

。