CN104936230A - Wireless sensor network energy balance route optimization method based on cluster head expectation - Google Patents

Abstract

The invention discloses a wireless sensor network energy balance route optimization method based on cluster head expectation. An algorithm considers two optimal objects which are cluster load distribution condition and node residual energy while electing cluster heads, and ensures the number of cluster heads in each round to be in an expected range by improving a threshold value while electing cluster heads so as to solve the problem that the energy consumption of the cluster heads is unbalanced. Simultaneously, the algorithm considers distance weight and residual energy weight by controlling coverage areas of different cluster heads in different positions so that the member node distribution of the cluster heads could be relative even to improve node energy efficiency. The algorithm presented by the invention has relative high energy efficiency, well balances the energy consumption of nodes in the network, and enables the cluster head distribution and number to be more stable. The optimization method of the invention could improve data transmission amount, prolongs network lifetime and preferably meet the requirement of the wireless sensor network to the network lifetime in a periodic monitoring environment.

Description

A kind of balancing energy of wireless sensor network routing optimization method expected based on bunch head

Technical field

The present invention relates to a kind of balancing energy of wireless sensor network routing optimization method, particularly relate to a kind of balancing energy of wireless sensor network routing optimization method expected based on bunch head.

Background technology

Wireless sensor network WSN (Wireless Sensor Networks) is as an important branch of Internet of Things, its effect is perception image data in institute's monitoring of environmental, adjacent node is perception same target information in real time, then merged, compress after send to base station by wireless from group multi-hop agreement.WSN clustering routing has clear layer, autgmentability and by force, easily realizes the advantages such as adjacent data fusion, is very applicable to the applications such as industry monitoring.

Clustering Routing comprises the election of bunch head, the communication bunch between head and base station, can adopt single-hop and multi-hop two kinds of communication modes.LEACH (the Low-energy Adaptive Clustering Hierarchy) algorithm that the people such as W.Heinzelman propose is a kind of self-organizing, adaptive complete distributed clustering routing algorithm.In order to extend the out-of-service time of first sensor node, LEACH algorithm utilizes randomness Cycle-switching Cluster-head, thus reaches the object of balancing network interior joint energy load.The people such as established law are for hot-zone problem, multihop routing algorithm based on Uneven Cluster is proposed, near base station bunch scale be less than away from base station bunch, bunch head therefore near base station can for bunch between data retransmission reserve energy, reach the object of balanced bunch head energy ezpenditure.There is the situation of very big bunch and minimal variety for LEACH algorithm in the people such as Lv Tao, the method proposed by controlling bunch number of members and merging minimal variety makes each sub-clustering energy balance in WSN; The people such as Jiang Changjiang for the problem of hot-zone in the Cluster Networks of multi-hop mode, propose a kind of efficient balancing energy, Uneven Cluster and bunch between multihop routing organically combine distributed routing algorithm; The people such as HuiLin, for integrated topological structure and clustering routing problem, propose a mixed-integer linear programming model, to determine best bunch head position; The hot-zone problem that Sun Yan waits clearly people uneven for node load and formed, proposes a kind of Clustering Routing based on dynamic partition load balancing; The people such as Su Jinshu for bunch between the lack of uniformity of load, propose the fault-tolerant cluster algorithm of wireless sensor network of a load balancing perception; Wasp pine waits people for the too much problem of multiple bunches of heads in LEACH and base station telecommunication energy ezpenditure, the energy of consideration node and positional factor, with the structure to optimize bunch; The people such as Aimin Wang introduce energy information in the election of cluster head threshold value of LEACH, adopt sliding window mechanism, can according to dynamic node number adjustment bunch head number.

Due to bunch head should to manage bunch in communicate to carry out again bunch between communication, so the energy consumption of bunch head will than bunch in member node many, thus cause node energy unbalanced.If bunch head premature failure, this bunch will be caused to lose efficacy in epicycle, and form route cavity, and then shorten network lifecycle.Although above Clustering Routing is energy efficient to a certain extent, but the unbalanced problem of node energy not yet solves, although Clustering Routing can the transmission quantity of optimization data, reduce network energy consumption, the lack of uniformity of bunch head load can have a strong impact on the performance of routing algorithm.Therefore, how to consider during election bunch head that the dump energy of node and the load of balance each bunch of head are vital.

Summary of the invention

The object of invention is the deficiency overcoming prior art, a kind of balancing energy of wireless sensor network routing optimization method (CHEEB) expected based on bunch head is provided, this algorithm considers sub-clustering load Distribution situation and residue energy of node two optimization aim when electing bunch head simultaneously, ensure that each takes turns bunch head number at expected range, to alleviate bunch unbalanced problem of head energy ezpenditure by threshold value when improving election bunch head; Simultaneously by controlling the coverage of diverse location bunch head, considering distance weights and dump energy weights, making the member node distribution of bunch head comparatively even, to improve node energy efficiency.

The technical solution adopted for the present invention to solve the technical problems is: provide a kind of balancing energy of wireless sensor network routing optimization method expected based on bunch head, it is characterized in that, comprise: by number of clusters order, residue energy of node and bunch the factor such as load balancing join in the election of bunch head, using the load Distribution situation of sub-clustering and residue energy of node two leading indicators as election bunch head; The operating time unit of the election of described bunch of head is wheel, and each is taken turns and is divided into bunch establishment stage and data transfer phase two parts, ensures that each takes turns bunch head number at expected range, solve bunch unbalanced problem of head energy ezpenditure by adjustment threshold value; By controlling the coverage of diverse location bunch head, calculating distance weights and dump energy weights, making node member's distribution of bunch head comparatively even, to improve node energy efficiency.

Preferably, Ci is whether node i became the indicator function of bunch head at current period, and the step of described bunch of establishment stage is:

A1, travel through each inefficacy and the node of Ci=1, the real number j between node i stochastic generation [0,1];

The threshold value Pi of A2, computing node i, judges the magnitude relationship of Pi and j; If j<Pi, then enter steps A 3, otherwise forward steps A 4 to;

A3, node i are elected as epicycle bunch head, and Ci is set to 0; The elected bunch header of bunch head broadcast; Enter steps A 5;

The not elected epicycle bunch head of A4, node i, Ci is set to 1; Receive the elected bunch header that all bunches of hairs come; Enter steps A 5;

The signal strength signal intensity of the elected bunch header that A5, the non-leader cluster node of not being elected as bunch head come according to each bunch of hair received, select the maximum bunch head of signal strength signal intensity as epicycle to add bunch; Connectivity request message is fed back to selected bunch head by non-leader cluster node;

A6, bunch head receive the connectivity request message of non-leader cluster node, and according to the quantity of bunch interior nodes, bunch head creates the timetable when an advised nodes can transmit data, and the node in this timetable being broadcast to bunch;

A7, bunch in node time of reception table and enter data transfer phase.

Preferably, the step of described data transfer phase is:

Whether B1, decision node i are bunch heads; If not then entering step B2, if yes then enter step B3;

The radio of B2, node in each bunch is closed until the transmission time distributing to this node arrives; The node of Ci=1 sends the packet of dump energy information to a bunch head in last time slot of oneself; The node of Ci=0 send in last time slot of oneself not containing the packet of dump energy information to a bunch head, forward step B3 to;

B3, bunch head open the packet that receiver reception bunch interior nodes is sent; Advanced row data fusion after bunch head receives the packet of all bunches of interior nodes, more incidentally dump energy information will send to base station;

B4, base station receive the next information of bunch hair and calculate the average energy of Ci=1 node, then being broadcast to the whole network;

B5, node receive the necessary average energy of calculated threshold;

B6, a new round start, and judge whether it is the new cycle, if the Ci of all nodes is then reset to 1 by the new cycle, proceed to steps A 1; Then proceed to step B1 if the judgment is No.

Preferably, in steps A 3 the elected bunch header of bunch head broadcast comprise elected leader cluster node ID and for distinguishing the stem whether this message be notice information.

Preferably, the computing formula of the threshold value Pi in described steps A 2 is:

$P_i=\frac{k*C_i}{n-k*\left(r\mathrm{mod}\frac{n}{k}\right)}*\frac{E_i}{E_{C_i=1}}---\left(1\right)$

Because the node only having current period not become bunch head can participate in election of cluster head, so the denominator of the energy proportion factor is not the average energy of all nodes in (1) formula, but be eligible for the node average energy of election; (1) formula can make to expect that bunch head number keeps k constant simultaneously; Bunch head expects that formula is:

$E[\#\mathrm{CH}]=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{P}_i---\left(2\right)$

(1) formula is substituted into (2) formula obtain:

$\begin{array}{c}E[\#\mathrm{CH}]\\ =\frac{k}{n-k*\left(r\mathrm{mod}\frac{n}{k}\right)}*\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}(C_i*E_i)}{\stackrel{\&OverBar;}{E_{C_i=1}}}\\ =\frac{k}{n-k\left(r\mathrm{mod}\frac{n}{k}\right)}*\frac{\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}(C_i*E_i)}{\frac{\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}(C_i*E_i)}{\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{C}_i}}\\ =\frac{k*\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{C}_i}{n-k\left(r\mathrm{mod}\frac{n}{k}\right)}\end{array}---\left(3\right)$

The document " An application-specific protocol architecture forwireless microsensor networks " studied from people such as W.Heinzelman:

$E\left[\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{C}_i\right]=n-k*\left(r\mathrm{mod}\frac{n}{k}\right)---\left(4\right)$

By (4) formula and (3) Shi Ke get:

$E[\#\mathrm{CH}]=\frac{k*[n-k*\left(r\mathrm{mod}\frac{n}{k}\right)]}{n-k*\left(r\mathrm{mod}\frac{n}{k}\right)}=k---\left(5\right)$

In order to node energy is balanced, high-energy node be allowed to be elected to bunch head, the address location of self, dump energy, node and the information broadcast such as the distance of bunch head and the distance of bunch head and base station are given other nodes by each candidate cluster head more; The primary power of node i is made to be E _i0, take turns the dump energy weights F (E before sub-clustering at r _ir) be

$F\left(E_{i^r}\right)=\left\{\begin{array}{c}1,r=1\\ \frac{E_{i^{r-1}}}{\frac{1}{N(m,r-1)}\underset{j=1}{\overset{N(m,r-1)}{\mathrm{\&Sigma;}}}{E}_{j^{r-1}}}\&CenterDot;\frac{E_{i^{r-1}}}{E_{i^0}},r>1\end{array}\right.---\left(6\right)$

F (E _ir) larger, illustrate that this node dump energy is at this moment larger;

Node i is to a bunch head CH _icommunication range less, the energy consumption bunch between head and node is less; If same node i is less to the communication range of base station BS, the energy consumption of transfer of data is also less.According to Free propagation energy model, then comprehensive distance weights can be expressed as

$D\left(v_i\right)=1-\frac{d{(i,{\mathrm{CH}}_i)}^2}{\underset{j=1}{\overset{N(m,r-1)}{\mathrm{\&Sigma;}}}d{(j,{\mathrm{CH}}_i)}^2}-\frac{d{(i,\mathrm{BS})}^2}{\underset{j=1}{\overset{N(m,r-1)}{\mathrm{\&Sigma;}}}d{(j,\mathrm{BS})}^2}---\left(7\right)$

In conjunction with formula (5), add comprehensive distance weights simultaneously and make bunch head as far as possible near base station, reduce the energy that transfer of data consumes; The probability that node i becomes bunch head is calculated by formula (8):

$P_{i-\mathrm{ch}}=\mathrm{\&alpha;F}\left(E_{i^r}\right)+\mathrm{\&beta;D}\left(v_i\right)---\left(8\right)$

If α, β probability ratio that to be adjustment node dump energy weights and comprehensive distance weights shared when bunch head competition, and alpha+beta=1.

The invention has the beneficial effects as follows: algorithm of the present invention considers sub-clustering load Distribution situation and residue energy of node two optimization aim when electing bunch head simultaneously, ensure that each takes turns bunch head number at expected range, to alleviate bunch unbalanced problem of head energy ezpenditure by threshold value when improving election bunch head; Simultaneously by controlling the coverage of diverse location bunch head, considering distance weights and dump energy weights, making the member node distribution of bunch head comparatively even, to improve node energy efficiency.The algorithm that the present invention proposes has higher capacity usage ratio, the balanced well energy ezpenditure of nodes, the distribution of bunch head and quantity can be more stable, data biography amount can be improved, extend network lifecycle, meet wireless sensor network requirement to network lifecycle in periodicity monitoring of environmental better.

Below in conjunction with drawings and Examples, the present invention is described in further detail; But a kind of balancing energy of wireless sensor network routing optimization method expected based on bunch head of the present invention is not limited to embodiment.

Accompanying drawing explanation

Fig. 1 is of the present invention based on a bunch sub-clustering phase flow figure for the balancing energy Routing Optimization Algorithm of head expectation;

Fig. 2 is of the present invention based on a bunch data transfer phase flow chart for the balancing energy Routing Optimization Algorithm of head expectation;

Fig. 3 is that the total energy consumption that three kinds of algorithms of the present invention are often taken turns compares;

Fig. 4 is that the network lifecycle of three kinds of algorithms of the present invention compares;

Fig. 5 is that the volume of transmitted data of three kinds of algorithms of the present invention compares;

Fig. 6 is that a bunch head distributed number for three kinds of algorithms of the present invention compares;

Fig. 7 to be of the present invention bunch of head number be 8 CHEEB sub-clustering effect.

Embodiment

Embodiment 1

Shown in Fig. 1 and Fig. 2, a kind of balancing energy of wireless sensor network routing optimization method expected based on bunch head of the present invention, it is characterized in that, comprise: by number of clusters order, residue energy of node and bunch the factor such as load balancing join in the election of bunch head, using the load Distribution situation of sub-clustering and residue energy of node two leading indicators as election bunch head; The operating time unit of the election of described bunch of head is wheel, and each is taken turns and is divided into bunch establishment stage and data transfer phase two parts, ensures that each takes turns bunch head number at expected range, solve bunch unbalanced problem of head energy ezpenditure by adjustment threshold value; By controlling the coverage of diverse location bunch head, calculating distance weights and dump energy weights, making node member's distribution of bunch head comparatively even, to improve node energy efficiency.

Further, Ci is whether node i became the indicator function of bunch head at current period, and the step of described bunch of establishment stage is:

A1, travel through each inefficacy and the node of Ci=1, the real number j between node i stochastic generation [0,1];

The threshold value Pi of A2, computing node i, judges the magnitude relationship of Pi and j; If j<Pi, then enter steps A 3, otherwise forward steps A 4 to;

A3, node i are elected as epicycle bunch head, and Ci is set to 0; The elected bunch header of bunch head broadcast; Enter steps A 5;

The not elected epicycle bunch head of A4, node i, Ci is set to 1; Receive the elected bunch header that all bunches of hairs come; Enter steps A 5;

The signal strength signal intensity of the elected bunch header that A5, the non-leader cluster node of not being elected as bunch head come according to each bunch of hair received, select the maximum bunch head of signal strength signal intensity as epicycle to add bunch; Connectivity request message is fed back to selected bunch head by non-leader cluster node;

A6, bunch head receive the connectivity request message of non-leader cluster node, and according to the quantity of bunch interior nodes, bunch head creates the timetable when an advised nodes can transmit data, and the node in this timetable being broadcast to bunch;

A7, bunch in node time of reception table and enter data transfer phase.

Further, the step of described data transfer phase is:

Whether B1, decision node i are bunch heads; If not then entering step B2, if yes then enter step B3;

The radio of B2, node in each bunch is closed until the transmission time distributing to this node arrives; The node of Ci=1 sends the packet of dump energy information to a bunch head in last time slot of oneself; The node of Ci=0 send in last time slot of oneself not containing the packet of dump energy information to a bunch head, forward step B3 to;

B3, bunch head open the packet that receiver reception bunch interior nodes is sent; Advanced row data fusion after bunch head receives the packet of all bunches of interior nodes, more incidentally dump energy information will send to base station;

B4, base station receive the next information of bunch hair and calculate the average energy of Ci=1 node, then being broadcast to the whole network;

B5, node receive the necessary average energy of calculated threshold;

B6, a new round start, and judge whether it is the new cycle, if the Ci of all nodes is then reset to 1 by the new cycle, proceed to steps A 1; Then proceed to step B1 if the judgment is No.

Preferably, in steps A 3 the elected bunch header of bunch head broadcast comprise elected leader cluster node ID and for distinguishing the stem whether this message be notice information.

Embodiment 2

LEACH algorithm have employed Cycle-switching Cluster-head method, and the operating time, unit was wheel, and each is taken turns and is divided into bunch establishment stage and data transfer phase two parts.Bunch establishment stage, the positive number that node i stochastic generation one is less than 1, if it is less than threshold value P _i, so node i is elected as epicycle bunch head.Threshold value P _ifor:

$P_i=\left\{\begin{array}{cc}\frac{k}{n-k*\left(r\mathrm{mod}\frac{n}{k}\right)},& C_i=1\\ 0,& C_i=0\end{array}\right.---\left(1\right)$

Wherein n is node total number, and bunch head number expected is k (k is custom parameter, as k=5%*n), and r is for working as front-wheel number, and definition n/k wheel is one-period.C _iwhether node i became the indicator function of bunch head at current period, if namely node i did not also become a bunch head at current period, so C _i=1, otherwise C _i=0.For the ease of computing and proof, (1) formula is rewritten as (2) formula:

$P_i=\frac{k*C_i}{n-k*\left(r\mathrm{mod}\frac{n}{k}\right)}---\left(2\right)$

Have many ALEACH algorithm consideration dump energies when election bunch head of improving one's methods as the people such as Md.Solaiman Ali propose, its threshold value is:

$P_i=\frac{k*C_i}{n-k*\left(r\mathrm{mod}\frac{n}{k}\right)}+\frac{E_i}{E_{\max }}*\frac{k}{n}---\left(3\right)$

But its bunch of head desired value is greater than optimum bunch head number, and along with the increase of wheel number, the ENERGY E of node i _ireduce gradually, make P _ireduce gradually, this is by few for bunch head growing number causing expecting.But bunch head number that ALEACH algorithm destroys the verified LEACH of the people such as W.Heinzelman is expected, has occurred the problem that bunch head number is successively decreased within the same cycle, during each week, zigzag wave is dynamic.LEACH algorithm does not consider residue energy of node when election bunch head, and the node that such dump energy is few also can be chosen as a bunch head, thus this bunch of head energy pre-mature exhaustion, cause this bunch to lose efficacy.

In order to address this problem, the threshold value P of the CHEEB algorithm that the present invention proposes _ibe set to:

$P_i=\frac{k*C_i}{n-k*\left(r\mathrm{mod}\frac{n}{k}\right)}*\frac{E_i}{\stackrel{\&OverBar;}{E_{C_i=1}}}---\left(4\right)$

Because the node only having current period not become bunch head can participate in election of cluster head, so the denominator of the energy proportion factor is not the average energy of all nodes in (4) formula, but be eligible for the node average energy of election.(4) formula can make to expect that bunch head number keeps k constant simultaneously.Bunch head expects that formula is:

$E[\#\mathrm{CH}]=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{P}_i---\left(5\right)$

(4) formula is substituted into (5) formula obtain:

$\begin{array}{c}E[\#\mathrm{CH}]\\ =\frac{k}{n-k*\left(r\mathrm{mod}\frac{n}{k}\right)}*\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}(C_i*E_i)}{\stackrel{\&OverBar;}{E_{C_i=1}}}\\ =\frac{k}{n-k\left(r\mathrm{mod}\frac{n}{k}\right)}*\frac{\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}(C_i*E_i)}{\frac{\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}(C_i*E_i)}{\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{C}_i}}\\ =\frac{k*\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{C}_i}{n-k\left(r\mathrm{mod}\frac{n}{k}\right)}\end{array}---\left(6\right)$

The document " An application-specific protocol architecture forwireless microsensor networks " studied from people such as W.Heinzelman:

$E\left[\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{C}_i\right]=n-k*\left(r\mathrm{mod}\frac{n}{k}\right)---\left(7\right)$

By (6) formula and (7) Shi Ke get:

$E[\#\mathrm{CH}]=\frac{k*[n-k*\left(r\mathrm{mod}\frac{n}{k}\right)]}{n-k*\left(r\mathrm{mod}\frac{n}{k}\right)}=k---\left(8\right)$

To sum up, prove that the expectation of bunch head number of the CHEEB algorithm that the present invention proposes is all k, illustrated that CHEEB algorithm maintains the optimum bunch head number of LEACH algorithm.

In order to node energy is balanced, high-energy node be allowed to be elected to bunch head, the address location of self, dump energy, node and the information broadcast such as the distance of bunch head and the distance of bunch head and base station are given other nodes by each candidate cluster head more.The primary power of node i is made to be E _i0, take turns the dump energy weights F (E before sub-clustering at r _ir) be

$F\left(E_{i^r}\right)=\left\{\begin{array}{c}1,r=1\\ \frac{E_{i^{r-1}}}{\frac{1}{N(m,r-1)}\underset{j=1}{\overset{N(m,r-1)}{\mathrm{\&Sigma;}}}{E}_{j^{r-1}}}\&CenterDot;\frac{E_{i^{r-1}}}{E_{i^0}},r>1\end{array}\right.---\left(9\right)$

F (E _ir) larger, illustrate that this node dump energy is at this moment larger.

Node i is to a bunch head CH _icommunication range less, the energy consumption bunch between head and node is less; If same node i is less to the communication range of base station BS, the energy consumption of transfer of data is also less.According to Free propagation energy model, then comprehensive distance weights can be expressed as

$D\left(v_i\right)=1-\frac{d{(i,{\mathrm{CH}}_i)}^2}{\underset{j=1}{\overset{N(m,r-1)}{\mathrm{\&Sigma;}}}d{(j,{\mathrm{CH}}_i)}^2}-\frac{d{(i,\mathrm{BS})}^2}{\underset{j=1}{\overset{N(m,r-1)}{\mathrm{\&Sigma;}}}d{(j,\mathrm{BS})}^2}---\left(10\right)$

In conjunction with formula (8), add comprehensive distance weights simultaneously and make bunch head as far as possible near base station, reduce the energy that transfer of data consumes.The probability that node i becomes bunch head is calculated by formula (11).

$P_{i-\mathrm{ch}}=\mathrm{\&alpha;F}\left(E_{i^r}\right)+\mathrm{\&beta;D}\left(v_i\right)---\left(11\right)$

If α, β probability ratio that to be adjustment node dump energy weights and comprehensive distance weights shared when bunch head competition, and alpha+beta=1.

Embodiment 3

The false code of CHEEB algorithm proposed by the invention is as follows:

Embodiment 4

In order to prove validity of the present invention, adopting MATLAB emulation tool to test, comparing ALEACH, EEUC and CHEEB tri-kinds of routing algorithms.Concrete simulation parameter is as shown in table 1.

Table 1 optimum configurations table

(1) network energy consumption

The height of network energy consumption directly affects the performance of route, and network energy consumption is less relative to the slope of wheel number, and energy consumption is less, and energy is more efficient.Fig. 3 compares the total energy consumption that three kinds of algorithms are often taken turns, and as can be seen from the figure, the energy consumption of CHEEB is all less than the energy consumption of ALEACH and EEUC.Because CHEEB not only considers the dump energy of node when electing bunch head, also assures that each takes turns bunch head number at expected range, so CHEEB is more efficient than the energy of ALEACH and EEUC.

(2) network lifecycle

From the first round of WSN to the timing definition of first node failure be network lifecycle.Fig. 4 compares the network lifecycle of three kinds of algorithms, and as can be seen from the figure, the life cycle of CHEEB is all longer than the life cycle of ALEACH and EEUC.Wherein CHEEB is about 25%, CHEEB than the life cycle of ALEACH and is about 10% than the life cycle of EEUC.From the quantity of surviving node, the surviving node of CHEEB when the 1000th takes turns is more than 90%, and the node of ALEACH all lost efficacy, and the surviving node of EEUC is about 40%.This illustrates the balanced well energy ezpenditure of nodes of CHEEB.

(3) volume of transmitted data

The information that sensor node collects finally all will send to base station, and volume of transmitted data also becomes one of index of routing algorithm efficiency, and when energy ezpenditure is identical, volume of transmitted data is The more the better.Fig. 5 compares the volume of transmitted data of three kinds of algorithms, and as can be seen from the figure, CHEEB volume of transmitted data is maximum, and EEUC takes second place, and ALEACH is minimum.When all node energies exhaust, CHEEB is EEUC relative to 1.5 times of transfer of data increment of ALEACH relative to the transfer of data increment of ALEACH, illustrates that CHEEB is higher than the efficiency of algorithm of ALEACH and EEUC.

(4) bunch head distributed number

In expected range, more stable bunch head quantity will make load more balanced.Fig. 6 compares bunch head quantity of three kinds of algorithms, as can be seen from the figure, relative to ALEACH and EEUC, bunch head quantity that CHEEB algorithm produces more concentrates on a bunch desired value for head quantity, main cause is that the threshold value expected based on bunch head that proposes and distance weights more adequately can describe network characteristic, and therefore bunch head distribution and quantity can be more stable.

As can be seen from Figure 7, size entirety in CHEEB algorithm bunch is more even, position simultaneously bunch in application scenarios is relatively suitable, and this all serves certain effect to the rate of energy dissipation of entirety in the energy ezpenditure gap of balance leader cluster node and non-leader cluster node and network

Above-described embodiment is only used for further illustrating a kind of balancing energy of wireless sensor network routing optimization method expected based on bunch head of the present invention; but the present invention is not limited to embodiment; every above embodiment is done according to technical spirit of the present invention any simple modification, equivalent variations and modification, all fall in the protection range of technical solution of the present invention.

Claims (5)

1. the balancing energy of wireless sensor network routing optimization method expected based on bunch head, it is characterized in that, comprise: by number of clusters order, residue energy of node and bunch the factor such as load balancing join in the election of bunch head, using the load Distribution situation of sub-clustering and residue energy of node two leading indicators as election bunch head; The operating time unit of the election of described bunch of head is wheel, and each is taken turns and is divided into bunch establishment stage and data transfer phase two parts, ensures that each takes turns bunch head number at expected range, solve bunch unbalanced problem of head energy ezpenditure by adjustment threshold value; By controlling the coverage of diverse location bunch head, calculating distance weights and dump energy weights, making node member's distribution of bunch head comparatively even, to improve node energy efficiency.

2. a kind of balancing energy of wireless sensor network routing optimization method expected based on bunch head according to claim 1, is characterized in that: Ci is whether node i became the indicator function of bunch head at current period, and the step of described bunch of establishment stage is:

A1, travel through each inefficacy and the node of Ci=1, the real number j between node i stochastic generation [0,1];

The threshold value Pi of A2, computing node i, judges the magnitude relationship of Pi and j; If j<Pi, then enter steps A 3, otherwise forward steps A 4 to;

A3, node i are elected as epicycle bunch head, and Ci is set to 0; The elected bunch header of bunch head broadcast; Enter steps A 5;

The not elected epicycle bunch head of A4, node i, Ci is set to 1; Receive the elected bunch header that all bunches of hairs come; Enter steps A 5;

The signal strength signal intensity of the elected bunch header that A5, the non-leader cluster node of not being elected as bunch head come according to each bunch of hair received, select the maximum bunch head of signal strength signal intensity as epicycle to add bunch; Connectivity request message is fed back to selected bunch head by non-leader cluster node;

A6, bunch head receive the connectivity request message of non-leader cluster node, and according to the quantity of bunch interior nodes, bunch head creates the timetable when an advised nodes can transmit data, and the node in this timetable being broadcast to bunch;

A7, bunch in node time of reception table and enter data transfer phase.

3. a kind of balancing energy of wireless sensor network routing optimization method expected based on bunch head according to claim 2, is characterized in that: the step of described data transfer phase is:

Whether B1, decision node i are bunch heads; If not then entering step B2, if yes then enter step B3;

The radio of B2, node in each bunch is closed until the transmission time distributing to this node arrives; The node of Ci=1 sends the packet of dump energy information to a bunch head in last time slot of oneself; The node of Ci=0 send in last time slot of oneself not containing the packet of dump energy information to a bunch head, forward step B3 to;

B3, bunch head open the packet that receiver reception bunch interior nodes is sent; Advanced row data fusion after bunch head receives the packet of all bunches of interior nodes, more incidentally dump energy information will send to base station;

B4, base station receive the next information of bunch hair and calculate the average energy of Ci=1 node, then being broadcast to the whole network;

B5, node receive the necessary average energy of calculated threshold;

B6, a new round start, and judge whether it is the new cycle, if the Ci of all nodes is then reset to 1 by the new cycle, proceed to steps A 1; Then proceed to step B1 if the judgment is No.

4. a kind of balancing energy of wireless sensor network routing optimization method expected based on bunch head according to claim 2, is characterized in that: in steps A 3 bunch elected bunch header of head broadcast comprise elected leader cluster node ID and for distinguishing the stem whether this message be notice information.

5. a kind of balancing energy of wireless sensor network routing optimization method expected based on bunch head according to claim 2, is characterized in that: the computing formula of the threshold value Pi in described steps A 2 is:

$P_i=\frac{k*C_i}{n-k*\left(r\mathrm{mod}\frac{n}{k}\right)}*\frac{E_i}{E_{C_i=1}}---\left(1\right)$

Because the node only having current period not become bunch head can participate in election of cluster head, so the denominator of the energy proportion factor is not the average energy of all nodes in (1) formula, but be eligible for the node average energy of election; (1) formula can make to expect that bunch head number keeps k constant simultaneously; Bunch head expects that formula is:

$E[\#\mathrm{CH}]=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{P}_i---\left(2\right)$

(1) formula is substituted into (2) formula obtain:

$\begin{array}{c}E[\#\mathrm{CH}]\\ =\frac{k}{n-k*\left(r\mathrm{mod}\frac{n}{k}\right)}*\frac{\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}(C_i*E_i)}{E_{C_i=1}}\\ =\frac{k}{n-k\left(r\mathrm{mod}\frac{n}{k}\right)}*\frac{\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}(C_i*E_i)}{\frac{\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}(C_i*E_i)}{\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{C}_i}}\\ =\frac{k*\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{C}_i}{n-k\left(r\mathrm{mod}\frac{n}{k}\right)}\end{array}---\left(3\right)$

The document " An application-specific protocol architecture forwireless microsensor networks " studied from people such as W.Heinzelman:

$E\left[\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{C}_i\right]=n-k*\left(r\mathrm{mod}\frac{n}{k}\right)---\left(4\right)$

By (4) formula and (3) Shi Ke get:

$E[\#\mathrm{CH}]=\frac{k*[n-k*\left(r\mathrm{mod}\frac{n}{k}\right)]}{n-k*\left(r\mathrm{mod}\frac{n}{k}\right)}---\left(5\right)$

In order to node energy is balanced, high-energy node be allowed to be elected to bunch head, the address location of self, dump energy, node and the information broadcast such as the distance of bunch head and the distance of bunch head and base station are given other nodes by each candidate cluster head more; The primary power of node i is made to be E _i0, take turns the dump energy weights F (E before sub-clustering at r _ir) be

$F\left(E_{i^r}\right)=\left\{\begin{array}{c}1,r=1\\ \frac{E_{i^{r-1}}}{\frac{1}{N(m,r-1)}\underset{j=1}{\overset{N(m,r-1)}{\mathrm{\&Sigma;}}}{E}_{j^{r-1}}}\&CenterDot;\frac{E_{i^{r-1}}}{E_{i^0}},r>1\end{array}\right.---\left(6\right)$

F (E _ir) larger, illustrate that this node dump energy is at this moment larger;

Node i is to a bunch head CH _icommunication range less, the energy consumption bunch between head and node is less; If same node i is less to the communication range of base station BS, the energy consumption of transfer of data is also less.According to Free propagation energy model, then comprehensive distance weights can be expressed as

$D\left(v_i\right)=1-\frac{d{(i,{\mathrm{CH}}_i)}^2}{\underset{j=1}{\overset{N(m,r-1)}{\mathrm{\&Sigma;}}}d{(j,{\mathrm{CH}}_i)}^2}-\frac{d{(i,\mathrm{BS})}^2}{\underset{j=1}{\overset{N(m,r-1)}{\mathrm{\&Sigma;}}}d{(j,\mathrm{BS})}^2}---\left(7\right)$

In conjunction with formula (5), add comprehensive distance weights simultaneously and make bunch head as far as possible near base station, reduce the energy that transfer of data consumes; The probability that node i becomes bunch head is calculated by formula (8):

P _i-ch＝αF(E _ir)+βD(v _i) (8)

If α, β probability ratio that to be adjustment node dump energy weights and comprehensive distance weights shared when bunch head competition, and alpha+beta=1.