CN102231899B - Guassian-declustered routing control method for wireless sensor network - Google Patents

Abstract

The invention relates to a Guassian-declustered routing control method for a wireless sensor network. According to the routing control method, a mean value is determined according to the optimal distance between cluster header nodes; standard difference is determined according to the network size and the cluster quantity; and a probability threshold function is determined according to the residual energy of nodes and the average energy of target nodes in a probability band; in the cluster selection process, the distance between the target nodes and known cluster headers is used as a parameter for generating a probability threshold; and in the stage of cluster generation, the nodes select closest cluster headers as final cluster headers. The Guassian-declustered routing control method for a wireless sensor network, disclosed by the invention, has the advantages of achieving more uniform cluster header distribution, reducing the cluster header energy consumption, realizing the uniform entire energy distribution of a network and slowing the network dead speed.

Description

A kind of Guassian-declustered routing control method for wireless sensor network

Technical field

The present invention relates to wireless sensor network, especially a kind of wireless sensor network route control method.

Background technology

Wireless sensor network is that one forms by being deployed in a large amount of cheap microsensor nodes in monitored area, by wireless Ad Hoc communication mode, the information of perceptive object in perception collaboratively, acquisition and processing network's coverage area, and send to user's intelligent network.The application of wireless sensor network is very extensive, can be applicable to the numerous areas such as medical monitoring, environmental monitoring, business logistics, traffic administration, precision agriculture, Military Application, space exploration, is a study hotspot of computer and the communications field.

At present, the research of radio sensing network is optimized mainly for the integral frame of wireless sensor network.It is wherein a focus of wireless sensor network research field to the research of Routing Protocol.Routing algorithm for wireless sensor is a kind of process contacting between node of setting up, and is also between node, to carry out information interaction, the synchronous important prerequisite of information, and between sensor node, only having could image data after the contact of foundation, otherwise will cause loss of data.For feature and the communication requirement of wireless sensor network, sensor node need to solve by utilizing local message to carry out decision-making and optimizing the problem of global behavior (be route generate and Route Selection).Because traditional wired network protocol does not need the problems such as energy, computing capability, storage capacity and the transmission range of worry about node, therefore the Routing Protocol algorithm complexity generally adopting in legacy network, amount of calculation is large, is not suitable for the application of wireless sensor network.Due to hierarchy type Clustering Routing have topological structure simple, be easy to safeguard, be applicable to the features such as large scale network to be the focus of wireless sensor network route research always.

Abroad start to walk relatively early for the research of wireless senser networking routing algorithm.The example hierarchy formula Routing Protocol that the people such as Heinzelman W B have taken the lead in proposing being adapted at applying in wireless sensor network is LEACH agreement.But the dump energy of not considering node causes node death too fast.

In recent years, domestic experts and scholars also study the routing algorithm of wireless sensor network one after another.The people such as the Hang Haicun of Tongji University combine LEACH algorithm and apply in complicated network environment with ant group algorithm, increased the Packet Generation amount in network, balance the energy consumption of network, the life span that has extended network.But its search time is longer, and network utilization is low.The people such as the Shao Xiaomeng of Shanghai Communications University intend, for the communication system under mine, using chamber, each hole as one bunch, making its requirement that more meets mine structure, but also not putting in real world applications and go after LEACH agreement is improved.Bunch algorithm based on node density that the people such as the Qiao Junfeng of Xian Electronics Science and Technology University propose has been considered the density of node in bunch head is chosen process, thus balanced scale of dividing each bunch, the offered load that balance is each bunch, the inequality of Energy distribution in reduction network.But it does not consider the dump energy of candidate cluster head node in bunch head is chosen process, thereby likely make the node choosing that dump energy is less work as for a bunch head, accelerate dead speed.

Summary of the invention

For overcome bunch head of existing wireless sensor network route control method distribute more inhomogeneous, a bunch energy consumption is large, the deficiency of network integral energy skewness, the dead speed of network, the invention provides a kind of bunch of head and distribute more evenly, reduce a bunch energy consumption, network integral energy and be evenly distributed, reduce the Guassian-declustered routing control method for wireless sensor network of the dead speed of network.

The technical solution adopted for the present invention to solve the technical problems is:

A kind of Guassian-declustered routing control method for wireless sensor network, described route control method comprises the following steps:

A1, by wireless sensor node random arrangement to monitored area;

A2, base station send routing command to each node in monitored area;

A3, survival node (being that dump energy is greater than zero) receive after order, determine the distance B between base station according to signal strength signal intensity _tosink, node identity marks is node, and by No. ID of node, energy information E _residualand D _tosinkbe that package_node sends to base station etc. information package;

The quantity N of survival node is determined in A4, base station according to the package_node of the each node receiving _all, energy E _residual, distance B _tosinketc. information, and calculate gross energy E _all:

4.1) if Nall > 0, bunch radius R is determined according to the scale of monitored area, survival number of nodes Nall, network gross energy Eall in base station, because node is uniformly distributed in network, network finally forms N bunch, network area is S, choosing in the process of bunch head, for arbitrary bunch of head, the radius at its place bunch is R, bunch area be π R ²,, the Gaussian Profile poor σ that settles the standard;

4.2) otherwise, finish

A5, base station send packets of information net_infor to node, comprise R value, σ value, number of nodes N _all, network gross energy E _alland start the information such as routing command;

A6, each node receive base station data bag net_infor and resolve obtain related data, determine R value, σ value start routing algorithm, and monitor the head_infor of leader cluster node transmission simultaneously;

If A7 node listens to head_infor, forward A8 to, otherwise according to probability threshold value formula

${\mathrm{CH}}_{\mathrm{prob}}=\frac{1}{\sqrt{2\mathrm{\π}}\mathrm{\σ}}{e}^{-\frac{{(\mathrm{dist}-\sqrt{3}R)}^2}{2{\mathrm{\σ}}^2}}\·\frac{E_{\mathrm{residual}}\×N_{\mathrm{all}}}{E_{\mathrm{all}}}$

Generate first leader cluster node, with for parameter generates the probability CH that is elected to bunch head _prob, node generates a random decimal random simultaneously,

7.1) if this probable value CH _probbe greater than random decimal random, node is elected as first leader cluster node, and identity marks is changed into header, and is covered by himself status indication, calculates the quantity N of its probability band interior nodes _sumand the gross energy E of the interior both candidate nodes of probability band _sum, be packet head_infor broadcast by information package;

7.2) otherwise, node continues first bunch of head of competition and monitors bunch header, repeated execution of steps A7 simultaneously;

A8, the node that is labeled as covered deposit information in local bunch header table in after receiving head_infor packets of information; Unmarked is that the node of covered is determined and the distance B of this bunch of head according to signal strength signal intensity _tohead, and resolve head_infor and obtain the relevant information of bunch head under packets of information,

8.1) if D _toheadbe less than or equal to R, vertex ticks oneself state is covered, and a bunch header head_infor is saved in local bunch header table, exits the competition of leader cluster node;

8.2) otherwise, node generates decimal random at random, and with D _toheadfor parameter, according to probability threshold value formula

${\mathrm{CH}}_{\mathrm{prob}}=\frac{1}{\sqrt{2\mathrm{\π}}\mathrm{\σ}}{e}^{-\frac{{(D_{\mathrm{tohead}}-\sqrt{3}R)}^2}{2{\mathrm{\σ}}^2}}\·\frac{E_{\mathrm{residual}}{\×N}_{\mathrm{sum}}}{E_{\mathrm{sum}}}$

Generate the probable value of elected bunch head,

8.3) if this probable value CH _probbe greater than random decimal random, node is elected as leader cluster node, and identity marks is changed into header, and oneself state is labeled as covered, calculates the quantity N of the new probability band interior nodes forming _sumand energy summation E _sumand broadcast data packet head_infor;

8.4) otherwise, node continue monitor a bunch signal;

A9, judge whether that all nodes are all labeled as covered, if so, a base station broadcast bunch head is chosen the finish command choose_over, forwards steps A 10 to; Otherwise repeated execution of steps A8;

A10, all nodes receive a bunch head and choose after the finish command choose_over, identity marks is that the node of node is its final bunch of head according to bunch leader cluster node header that header table selected distance is nearest of self, and transmit into a bunch request, identity marks is the leader cluster node setting-up time sheet of header, receive request into clusters simultaneously, after timeslice arrives, TDMA timetable in leader cluster node distributes bunch according to the request into clusters that receives, and take R as radius bunch in broadcast.

Further, described route control method is further comprising the steps of:

A11, node start acquisition testing data, send Monitoring Data according to the TDMA timetable receiving to a bunch hair;

A12, leader cluster node by receive bunch in data carry out data fusion, compress and send to base station.

Described route control method is further comprising the steps of: A13, repeating step A2～A12, until all nodes are all dead.

Further, described step 4.1) in, from the character of Gaussian Profile, the probability of the elected bunch head of other nodes outside probability band is far smaller than P _minlevel off to 0, will greatly improve the time complexity of algorithm if these nodes are also chosen to be to both candidate nodes, therefore by remaining S-π R ²region is divided N-1 bunch, chooses a bunch head in each candidate region, and average candidate's area of each bunch of head is (S-π R ²)/(N-1), and the area of candidate region as shown in Figure 2, therefore have:

$\mathrm{\π}[{(\sqrt{3}R+\mathrm{\Δd})}^2-{(\sqrt{3}R-\mathrm{\Δd})}^2]=\frac{S-{\mathrm{\πR}}^2}{N-1}$

$\mathrm{\Δd}=\frac{S-\mathrm{\π}{R}^2}{4\sqrt{3}\mathrm{\πR}(N-1)}$

Meanwhile, the probability occurring due to event is small probability event lower than 1%, in order to reduce the time complexity of algorithm, introduces minimum probability threshold value P in a bunch head is chosen process _min=0.01, probability is lower than P _minnode in epicycle no longer competition bunch head, set

make like this probability be greater than P _minnode fall within candidate's probability band scope within, so just the selection range of bunch head delimited width be in the annulus of 2 Δ d.

Due to

have according to the character of Gaussian Profile

can obtain by looking into the accurate Gaussian Profile numerical tabular of label taking:

$\frac{\mathrm{\Δd}}{\mathrm{\σ}}=P_{\mathrm{rob}}^{-1}\left(P_{\min }\right)$

?

$\mathrm{\σ}=\frac{S{-\mathrm{\πR}}^2}{4\sqrt{3}\mathrm{\πR}(N-1)\·P_{\mathrm{rob}}^{-1}\left(P_{\min }\right)}$

Determine thus the standard deviation of Gaussian Profile;

Technical conceive of the present invention is: two key factors in wireless sensor network route cluster process: distance and dump energy, propose a kind of clustering routing control method based on Gaussian Profile.This route control method is determined average according to the optimal distance between leader cluster node, settles the standard poor, and determine probability threshold value function in conjunction with the average energy of destination node in the dump energy of node and probability band according to the size of network and cluster quantity.Bunch head choose destination node in process using with the distance of the known bunch of head parameter as generating probability threshold value, be final bunch of head at the nearest bunch head of bunch formation stages node selection.

Choose the generation of heavy bunch phenomenon in process for fear of wireless sensor network cluster head, bunch head distributes and should meet the optimum overlay model of network, the elected bunch probability of known bunch of nearlyer node of head of distance is lower, only be elected to the just maximum of probability of bunch head at the node of suitable distance, just meet the character of Gaussian Profile.Algorithm according to the optimal distance between leader cluster node determine average, settle the standard according to network size and cluster quantity poor, the distance affects factor of generating probability threshold value formula, establishes final probability threshold value function in conjunction with the average energy of destination node in the dump energy of node and probability band.Bunch head choose in process destination node using with the distance of each known bunch of head as parameter generating probability threshold value, be final bunch of head at the nearest bunch head of bunch formation stages node selection, reduce the energy consumption that node is communicated by letter with bunch head, the communication of assurance network.Simultaneously leader cluster node to bunch in the disposable broadcast of member node bunch information changed man-to-man communication mode, saved the consumption of node energy.

Beneficial effect of the present invention is mainly manifested in: the invention enables the distribution of bunch head more even, and avoid man-to-man communication between leader cluster node and member node, effectively reduce the energy consumption of bunch head, the integral energy of equalizing network distributes.

Accompanying drawing explanation

Fig. 1 is that leader cluster node optimization covers schematic diagram; This patent is exactly the effect that will reach as shown in the figure, leader cluster node be uniformly distributed and each other distance be roughly

Fig. 2 is sensor node distribution map;

Fig. 3 is probability threshold value function curve diagram, determines the value of σ according to the minimum probability of the elected bunch head of probability band fringe node;

Fig. 4 is first node selection posterior probability band schematic diagram, and wherein, the bandwidth of the probability band of both candidate nodes composition is 2 Δ d, with optimum chosen area

annulus between distance be Δ d, exist with known bunch of head distance

node be the both candidate nodes of the second bunch of head, all participate in the competition that leader cluster node is chosen, distance is greater than

all nodes do not participate in choosing of bunch head;

Fig. 5 is second node selection posterior probability band schematic diagram;

Fig. 6 is all leader cluster nodes bunch distribution schematic diagrams while having chosen;

Fig. 7 is that the nearest leader cluster node of node selection enters bunch, and node is that final bunch of head is incorporated to bunch according to the nearest bunch head of local routing information table selected distance;

Embodiment

Below in conjunction with accompanying drawing, the invention will be further described.

With reference to Fig. 1～Fig. 7, a kind of Guassian-declustered routing control method for wireless sensor network, comprises the following steps:

A1, in monitored area random placement sensor node, as shown in Figure 2;

A2, base station send routing command to each node in monitored area;

A3, survival node receive after order, determine the distance B between base station _tosink, and be labeled as node, send package_node packets of information to base station;

The quantity N of survival node is determined in A4, base station _all, energy E _residual, distance B _tosinketc. information, and calculate gross energy E _all.Due to N _all> 0, the value of parameter σ is determined in base station according to the scale S of monitored area and bunch radius R;

Net_infor is broadcasted to the whole network in A5, base station;

A6, each node receive net_infor and resolve and determine R value, σ value and start routing algorithm, monitor and whether have head_infor packet broadcast simultaneously;

A7, now do not have node to be elected as header, each node is according to probability threshold value formula

${\mathrm{CH}}_{\mathrm{prob}}=\frac{1}{\sqrt{2\mathrm{\π}}\mathrm{\σ}}{e}^{-\frac{{(\mathrm{dist}-\sqrt{3}R)}^2}{2{\mathrm{\σ}}^2}}\·\frac{E_{\mathrm{residual}}\×N_{\mathrm{all}}}{E_{\mathrm{all}}}$

With

for elected first leader cluster node of parameter competition.If just the node of ID=8 is elected as first leader cluster node, be header by its identity marks, status indication is covered, calculates the quantity N of its probability band interior nodes _sumgross energy E with these nodes _sum, be packet head_infor broadcast by information package, the probability band figure that this stage forms is as shown in Figure 4;

A8, other nodes are and are labeled as covered, receive head_infor and determine and the distance B of this bunch of head _tohead, D _toheadthe vertex ticks that is less than or equal to R is covered, and a bunch header head_infor is saved in local bunch header table, exits the competition of leader cluster node, D _toheadthe node that is greater than R generates decimal random at random, and according to threshold value formula

${\mathrm{CH}}_{\mathrm{prob}}=\frac{1}{\sqrt{2\mathrm{\π}}\mathrm{\σ}}{e}^{-\frac{{(D_{\mathrm{tohead}}-\sqrt{3}R)}^2}{2{\mathrm{\σ}}^2}}\·\frac{E_{\mathrm{residual}}\×N_{\mathrm{sum}}}{E_{\mathrm{sum}}}$

Judge whether to be elected as a bunch head, as shown in Figure 5, the curve synoptic diagram of probability threshold value formula as shown in Figure 4 for the cluster figure after second leader cluster node chosen;

A9, repeated execution of steps A8, until all nodes are all capped, the final bunch head of whole network distributes as shown in Figure 6;

A10, the nearest leader cluster node of node selection are its final bunch of head, send request into clusters;

TDMA in A11, leader cluster node distribute bunch according to the request into clusters receiving, and broadcast bunch message, now the state of network is as shown in Figure 7;

A12, node send Monitoring Data according to TDMA timetable to a bunch hair;

A13, repeating step A2～A12, until all nodes are all dead.

Claims (3)

1. a Guassian-declustered routing control method for wireless sensor network, is characterized in that: described route control method comprises the following steps:

A1, by wireless sensor node random arrangement to monitored area;

A2, base station send routing command to each node in monitored area;

A3, survival node receive after order, determine the distance B between base station according to signal strength signal intensity _tosink, node identity marks is node, and by No. ID of node, energy information E _residualand D _tosinkinformation package is that package_node sends to base station;

The quantity N of survival node is determined in A4, base station according to the package_node of the each node receiving _all, energy E _residual, distance B _tosinkinformation, and calculate gross energy E _all:

4.1) if N _all>0, base station is according to the scale of monitored area, survival number of nodes N _all, network gross energy E _alldetermine bunch radius R, form N bunch, network area is S, and choosing in the process of bunch head, for arbitrary bunch of head, the radius at its place bunch is R, bunch area be π R ², the Gaussian Profile poor σ that settles the standard;

4.2) otherwise, finish;

A5, base station send packets of information net_infor to node, comprise R value, σ value, number of nodes N _all, network gross energy E _alland beginning routing command information;

A6, each node receive base station data bag net_infor and resolve obtain related data, determine R value, σ value start routing algorithm, and monitor the head_infor of leader cluster node transmission simultaneously;

If A7 node listens to head_infor, forward A8 to, otherwise according to probability threshold value formula

${\mathrm{CH}}_{\mathrm{prob}}=\frac{1}{\sqrt{2\mathrm{\π}}\mathrm{\σ}}{e}^{-\frac{{(\mathrm{dist}-\sqrt{3}R)}^2}{{2\mathrm{\σ}}^2}}\·\frac{E_{\mathrm{residual}}\×N_{\mathrm{all}}}{E_{\mathrm{all}}}$

Generate first leader cluster node, with

for parameter generates the probability CH that is elected to bunch head _prob, node generates a random decimal random simultaneously,

7.1) if this probable value CH _probbe greater than random decimal random, node is elected as first leader cluster node, and identity marks is changed into header, and is covered by himself status indication, calculates the quantity N of its probability band interior nodes _sumand the gross energy E of the interior both candidate nodes of probability band _sum, be packet head_infor broadcast by information package;

7.2) otherwise, node continues first bunch of head of competition and monitors bunch header, repeated execution of steps A7 simultaneously;

A8, the node that is labeled as covered deposit information in local bunch header table in after receiving head_infor packets of information; Unmarked is that the node of covered is determined and the distance B of this bunch of head according to signal strength signal intensity _tohead, and resolve head_infor and obtain the relevant information of bunch head under packets of information,

8.1) if D _toheadbe less than or equal to R, vertex ticks oneself state is covered, and a bunch header head_infor is saved in local bunch header table, exits the competition of leader cluster node;

8.2) otherwise, node generates decimal random at random, and with D _toheadfor parameter, according to probability threshold value formula

${\mathrm{CH}}_{\mathrm{prob}}=\frac{1}{\sqrt{2\mathrm{\π}}\mathrm{\σ}}{e}^{-\frac{{(D_{\mathrm{tohead}}-\sqrt{3}R)}^2}{{2\mathrm{\σ}}^2}}\·\frac{E_{\mathrm{residual}}\×N_{\mathrm{sum}}}{E_{\mathrm{sum}}}$

Generate the probable value CH of elected bunch head _prob;

8.3) if probable value CH _probbe greater than random decimal random, node is elected as leader cluster node, and identity marks is changed into header, and oneself state is labeled as covered, calculates the quantity N of the new probability band interior nodes forming _sumand energy summation E _sumand broadcast data packet head_infor;

8.4) otherwise, node continue monitor a bunch signal;

A9, judge whether that all nodes are all labeled as covered, if so, a base station broadcast bunch head is chosen the finish command choose_over, forwards steps A 10 to; Otherwise repeated execution of steps A8;

A10, all nodes receive a bunch head and choose after the finish command choose_over, identity marks is that the node of node is its final bunch of head according to bunch leader cluster node header that header table selected distance is nearest of self, and transmit into a bunch request, identity marks is the leader cluster node setting-up time sheet of header, receive request into clusters simultaneously, after timeslice arrives, TDMA timetable in leader cluster node distributes bunch according to the request into clusters that receives, and take R as radius bunch in broadcast.

2. a kind of Guassian-declustered routing control method for wireless sensor network as claimed in claim 1, is characterized in that: described route control method is further comprising the steps of:

A11, node start acquisition testing data, send Monitoring Data according to the TDMA timetable receiving to a bunch hair;

A12, leader cluster node by receive bunch in data carry out data fusion, compress and send to base station.

3. a kind of Guassian-declustered routing control method for wireless sensor network as claimed in claim 1 or 2, is characterized in that: described step 4.1) in, by remaining S-π R ²region is divided N-1 bunch, chooses a bunch head in each candidate region, and average candidate's area of each bunch of head is (S-π R ²)/(N-1), has:

$\mathrm{\π}[{(\sqrt{3}R+\mathrm{\Δd})}^2-{(\sqrt{3}R-\mathrm{\Δd})}^2]=\frac{S-\mathrm{\π}{R}^2}{N-1}$

$\mathrm{\Δd}=\frac{S-\mathrm{\π}{R}^2}{4\sqrt{3}\mathrm{\πR}(N-1)}$

Meanwhile, in choosing process, introduces on a bunch head minimum probability threshold value P _min=0.01, probability is lower than P _minnode in epicycle no longer competition bunch head, set make probability be greater than P _minnode fall within candidate's probability band scope within, by the selection range of bunch head delimit width be in the annulus of 2 Δ d;

Due to

have according to the character of Gaussian Profile

obtain by looking into the accurate Gaussian Profile numerical tabular of label taking:

$\frac{\mathrm{\Δd}}{\mathrm{\σ}}=P_{\mathrm{rob}}^{-1}\left(P_{\min }\right)$

?

$\mathrm{\σ}=\frac{S-\mathrm{\π}{R}^2}{4\sqrt{3}\mathrm{\πR}(N-1)\·P_{\mathrm{rob}}^{-1}\left(P_{\min }\right)}$

Determine thus the standard deviation sigma of Gaussian Profile.

Images