CN106792973A - Cluster head election and its shift process in a kind of energy heterogeneous wireless sensor network - Google Patents

Abstract

The invention discloses the cluster head election in a kind of energy heterogeneous wireless sensor network and its shift process, including：Cluster head chooses the stage, and the node different for primary power, the threshold value of generation is also different, and the sensor node in active state can generate the random number between 01, if this number is less than its corresponding threshold values T (n), the node is elected as cluster head；The establishment stage of cluster, after cluster head node is produced, actively to this message of all node broadcasts in network；The stabilization sub stage of data is transmitted, for aggregation and processing data, route data；Data are merged by node gathered data.The present invention improves the overall equilibrium degree of offered load while WSN data transmission credibilities are ensured, effectively reduces data redundancy and reduces node energy loss, so as to extend the working life of WSN.

Description

Cluster head election and its shift process in a kind of energy heterogeneous wireless sensor network

Technical field

The invention belongs to wireless sensor network system technical field, more particularly to a kind of energy heterogeneous wireless sensor network Cluster head election and its shift process in network.

Background technology

Wireless sensor network is a kind of network of foundation-free facility, and it is by one group of sensor node with wireless Ad Hoc side Formula is constituted, the information of perceptive object the purpose is to collaborative sensing, collection and in treatment network coverage geographic area, and to these numbers According to being processed, most at last these information transmissions to the user for needing, and this requires that WSN possesses data fusion reliably and with long-term Function.The computing capability of WSN nodes, wireless bandwidth and limited battery capacity, cluster head route belong to hierarchical clustering association View, the data volume for being transferred to node is reduced by performing data aggregation and fusion, reduces each node energy consumption in class, is a kind of effective Improve the solution of energy efficiency.Data fusion technique is to be processed many numbers evidence or information, is combined into more effectively, more Meet the process of the data of user's request, it is possible to reduce the network traffics of redundancy.It is not all in most application Sensor is all isomorphism, and being directed to the routing algorithm of isomorphism can not well reduce the energy of whole network under heterogeneous network Amount consumption, easily causes local nodes overload and energy is consumed excessively, and the working life of WSN is greatly reduced.Towards In the application of data fusion, network working life and transmission reliability are primarily to pursue a goal.Sensor is data acquisition, data The critical component for the treatment of, a physical quantity in physical world can be mapped to a quantitative measured value, make people couple by it Physical world is formed and quantifies understanding.Sensor equipment has a sensing module such as sound, light, magnetic, vision etc. for perceiving surrounding generation Physical quantity in boundary a, digital signal processing module is used for the signal for processing sensing module collection and the work(for completing procotol Can, a radio transceiver chip is used to communicate and a power module is for the various operations of sensor provide energy.It is existing Technical scheme, the wireless sensor network system from special screne application.Wireless sensor network system substantially four The individual stage：1. substantial amounts of sensor node is randomly dispersed in Target monitoring area；2. pair network carries out network topology planning；3. , it is necessary to the valid data that each sensor is collected pass node and base station back in time, accurately in data-gathering process；4. pass Sensor node gathers substantial amounts of, real time data and is compressed treatment.There is following restrictive condition in technical scheme above：1. power supply It is limited：Because wireless sensor network is typically configured in remote districts or the more unattainable natural environment of staff, Therefore the power supply in sensor node is difficult to change.And the small volume of sensor node, so as to cannot be equipped with for it large-scale Energy supply device.2. disposal ability is limited：Due to the limitation of volume, processing unit adds sensor using microprocessor The limited power of node, therefore the disposal ability of processor is limited.Sensor node cannot run large-scale application program and place Manage substantial amounts of application data.3. storage capacity is limited：On the premise of volume, limited power, sensor node using capacity compared with Small microprocessor stores a small amount of sensed data and short and small application program.4. communication distance is limited：Sensor node is adopted Communicated with radio.Its communication range radius is the circle (r is maximum communication distance) of r in the ideal situation, in communication model All the sensors node in enclosing can be in communication with each other.Because communication distance is bigger, energy consumption is bigger, thus sensor node energy Amount is limited to cause its communication distance to be limited.Wireless sensor network node finite energy and typically no energy supplement, therefore road Needed efficiently to utilize energy by agreement.Energy consumption is mainly reflected in collection, communication and data processing this several respect, and energy loss Key is in network layer communication.Routing Protocol is the core technology of wireless sensor network layer, and node can only obtain local topology knot Structure information, Routing Protocol will can select suitable path in partial network information ground foundation.Traditional Routing Protocol can not be straight Connect for sensor networking, and be directed to MANET design networking and communication protocol be generally not suitable for sensor network.1. traditional The design object of Routing Protocol is mainly high-quality service, the consideration in terms of lacking to energy-conservation, and the master of sensor network protocol Be related to target be reduce energy loss, improve network lifecycle, it is considered to performance indications it is inconsistent.Conventional routing protocols are simultaneously It is inapplicable.2. legacy protocol causes to lead to the critical path for collecting node with shortest path or minimum delay as optimization aim, necessarily Node energy exhausts rapidly on footpath, shortens network lifecycle, while change in topology is too fast, does not meet the design mesh of Routing Protocol Mark.3. legacy protocol reaches the requirement of routing convergence in global scope can cause the limited sensor network nodes of computing capability Overload, same causes some node pre-mature exhaustion energy, is unfavorable for network stabilization.4. the group for being proposed for MANET Net and communication protocol, are also generally unfit in sensor network.

In sum, the consideration in terms of traditional Routing Protocol lacks to energy-conservation, and autgmentability does not reach sensor yet The requirement at networking, conventional routing protocols can not be applicable；Node energy exhausts rapidly on towards the critical path for collect node；It is limited Sensor network nodes overload.

The content of the invention

It is an object of the invention to provide the cluster head election in a kind of energy heterogeneous wireless sensor network and its rotation side Method, it is intended to solve the consideration in terms of traditional Routing Protocol lacks to energy-conservation, and autgmentability does not reach sensor networking yet It is required that, conventional routing protocols can not be applicable；Node energy exhausts rapidly on towards the critical path for collect node；Limited sensing The problem of device network node overload.

The present invention is achieved in that cluster head election and its rotation side in a kind of energy heterogeneous wireless sensor network Method, is roughly divided into startup stage and stabilization sub stage, and in startup stage, sensor node selection random number, random number is less than threshold value T (n_i) node as cluster head, transmitted in stabilization sub stage, node continuous collecting data and to cluster head, cluster head is by the node in the cluster The data for transmitting are sent to sink nodes after carrying out fusion treatment.Because each node energy is different, threshold value T (n_i) also different, often One wheel chooses cluster head using the method for cluster head rotation.Cluster head election and its rotation in the energy heterogeneous wireless sensor network Method is comprised the following steps：

Step one, cluster head chooses stage, the node different for primary power, the threshold value T (n of generation_i) also different, it is in The sensor node of active state can generate the random number between 0-1, if this number is less than its corresponding threshold values T (n), should Node is elected as cluster head；

Step 2, the establishment stage of cluster, after cluster head node is produced, actively to this message of all node broadcasts in network； The sensor node for being not selected as cluster head selects optimal cluster head, and sends the solicited message for adding the cluster；Cluster head is received please After asking, node is set to member in cluster；After cluster is formed, cluster head informs to be each node distribution one in cluster by broadcast mode Individual tdma communication time slot；Only in one's own time slot, member node just can send data to cluster head node；

Step 3, transmits the stabilization sub stage of data, for aggregation and processing data, route data；Node gathered data is right Data are merged.

Further, the step 2 is specifically included：

(1) P is elected in each round_optN number of leader cluster node, wherein P_optIt is optimization cluster head ratio, is also probability-weighted, N It is total nodal point number of whole network；N-th_iIndividual node is by following threshold value T (n_i) (probability threshold) come decide whether turn into cluster head Node：

Wherein, r is current wheel number, and G is nearestDo not have to turn into the node set of cluster head in wheel；Each Node has the opportunity to turn into the more leader cluster node of consumed energy in turn；

E₀Represent the primary power of ordinary node, a₁,a₂,...,a_nThe ratio shared by n kind special joints, b are represented respectively₁, b₂,...,b_nRepresent that special joint primary power exceedes the multiple of ordinary node primary power respectively；

a₁N,a₂N,...,a_nThe primary power of N number of special joint is respectively E₀(1+b₁),E₀(1+b₂),...,E₀(1+b_n), Remaining (1-a₁-a₂,...-a_n) N number of ordinary node primary power be E₀(1+b_n)；

The total primary power of multi-tier Heterogeneous network is：

It is assumed that N number of sensor node is evenly distributed on the border circular areas that a radius is A at random, sink nodes are located at region Centre, the gross energy that is consumed is during each round sends data to cluster head：

Wherein, l is cluster head number, E_elecRepresent the energy consumed per bit data during operation transmission circuit or receiving circuit Amount, E_DAThe cost of data fusion is performed for cluster head,It is cluster head to the average distance of sink nodes,It is bunch member node To the average distance of leader cluster node, ε_ampd⁴ _sinkWithIt is the energy of amplifier consumption.

It is calculated

To E_roundLocal derviation is sought on l, and it is 0 to make the partial derivative, then optimal cluster head number is

Obtain the energy sum E that network is consumed in each round_round.Meanwhile,

The initial total energy E of network_total, it is known that R_totalIt is the estimate of network lifecycle, also can obtain

R_tatal=E_total/E_round

Different probability-weighted P are taken according to its primary power to this n+1 kinds node_i：

(2) the present energy E that node i is taken turns in r_iR () chooses its cluster head T turnaround time_i,Represent network in r The average energy of wheel, withAs reference energy and the present energy E of node_iR () is made comparisons, obtain：

Wherein, P_optIt is optimization cluster head ratio；

The average energy of each node of network is after r wheels：

Substitute intoIt is calculated probability-weighted H_i(r)；

Obtain the probability threshold of each node.

Cluster head election and its shift process in the energy heterogeneous wireless sensor network that the present invention is provided, cluster set up rank Section, only in one's own time slot, member node just can send data to cluster head node, it is ensured that the communication in cluster is not There is conflict；The stabilization sub stage of data is transmitted, for aggregation and processing data, route data；Data are entered by node gathered data Row fusion, so can not only reduce the transmission of redundant data, while improving transmission reliability.

The present invention improves the overall equilibrium degree of offered load while WSN data transmission credibilities are ensured, effectively reduces Data redundancy and reduction node energy loss, so as to extend the working life of WSN；Combine sensor technology, embedding people's formula meter Calculation technology, modern network and wireless communication technology, distributed information processing etc., can be by all kinds of integrated miniature biographies Sensor collaboratively real-time monitoring, the data danger for perceiving and gathering various monitoring objects, data are wirelessly sent, and with User terminal is sent to from the network mode of group multi-hop, so as to realize that people study the purpose of nature.

Brief description of the drawings

Fig. 1 is cluster head election and its shift process in energy heterogeneous wireless sensor network provided in an embodiment of the present invention Flow chart.

Fig. 2 and Fig. 3 are provided in an embodiment of the present invention and LESCH algorithms contrast schematic diagram.

Specific embodiment

In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to embodiments, to the present invention It is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not used to Limit the present invention.

Application principle of the invention is explained in detail below in conjunction with the accompanying drawings.

As shown in figure 1, cluster head election and its wheel in energy heterogeneous wireless sensor network provided in an embodiment of the present invention The method of changing is comprised the following steps：

S101：Cluster head chooses the stage, and the node different for primary power, the threshold value of generation is also different, in enlivening shape The sensor node of state can generate the random number between 0-1, if this number is less than its corresponding threshold values T (n), the node is worked as Elect cluster head as；

S102：The establishment stage of cluster, after cluster head node is produced, actively to this message of all node broadcasts in network；Not The sensor node for being selected as cluster head selects optimal cluster head, and sends the solicited message for adding the cluster；Cluster head receives request Afterwards, node is set to member in cluster；After cluster is formed, cluster head informs to be each node distribution one in cluster by broadcast mode Tdma communication time slot；Only in one's own time slot, member node just can send data to cluster head node；

S103：The stabilization sub stage of data is transmitted, for aggregation and processing data, route data；Node gathered data, logarithm According to being merged.

Specifically include in step s 102：

N number of sensor node is had in network to be evenly distributed at random in the border circular areas that a radius is A, cluster covering Area size is limited in the region that radius is r, i.e., the node in cluster head communication radius r could turn into this cluster member.

Assuming that sensor network has following property:

A) sensor node is fixed.

B) Sink node is deployed in a fixed position beyond region A × A, and Sink node is unique.

C) network need not be safeguarded artificially after disposing.

D) sensor network interior joint is energy heterogeneous.

E) sensor node is in the direct communications range of other nodes, and directly can be communicated with Sink node.

F) node location has randomness, and without equipment GPS, can not know its specific position by the method for measuring Put.

The present invention is obtained on the basis of homogeneous network cluster head routing algorithm.

During isomorphism, in order to be balanced the load in all-network node, about PN leader cluster node is elected in each round, its Middle P is optimization cluster head ratio, is also herein probability-weighted.Each node is decided whether to turn into by following probability threshold Leader cluster node：

Wherein, r is current wheel number, and G is nearestDo not have to turn into the node set of cluster head in wheel；Each Node has the opportunity to turn into the more leader cluster node of consumed energy in turn.

Ordinary node (Normal node) and n kinds special joint (Special node) are included in network, they have not Same primary power.

If E₀Represent the primary power of ordinary node, a₁,a₂,...,a_nThe ratio shared by n kind special joints is represented respectively, b₁,b₂,...,b_nRepresent that special joint primary power exceedes the multiple of ordinary node primary power respectively.

a₁N,a₂N,...,a_nThe primary power of N number of special joint is respectively E₀(1+b₁),E₀(1+b₂),...,E₀(1+b_n), Remaining (1-a₁-a₂,...-a_n) N number of ordinary node primary power be E₀(1+b_n)。

The total primary power of multi-tier Heterogeneous network is：

Different probability-weighted P are taken according to its primary power to this n+1 kinds node_i：

The wheel number that special joint turns into cluster head is less than ordinary node so that special joint is as the chance of cluster head more than general Logical node turns into the chance of cluster head.

First run data transfer is completed, into next round.

Changed by the energy of first round part of nodes, in addition it is also necessary to counted again with reference to the present energy situation of each node New probability-weighted H_i(r)。

The present energy E that the present invention is taken turns based on node i in r_iR () chooses its cluster head T turnaround time_i.Due to node tool There are different present energies, the probability-weighted H of high-energy node_iR () should be bigger than low energy node, can just make network uniform Ground consumed energy is extending the life cycle of network.

IfThe average energy that network is taken turns in r is represented, withAs reference energy and the present energy E of node_i R () is made comparisons, obtain：

Wherein, P_optIt is optimization cluster head ratio.

The present energy E of node_iR () can directly measure, and average energyIt is difficult to obtain, it is necessary to be estimated.

It is assumed that sink nodes are located in the middle of border circular areas, the data of k bits are sent to cluster head, in the energy of each round consumption Measure and be：

Wherein, l is cluster head number, E_elecRepresent the energy consumed per bit data during operation transmission circuit or receiving circuit Amount, E_DAThe cost of data fusion is performed for cluster head,It is cluster head to the average distance of sink nodes,It is bunch member node To the average distance of leader cluster node, ε_ampd⁴ _sinkWithIt is the energy of amplifier consumption.

It is calculated

To E_roundLocal derviation is sought on l, and it is 0 to make the partial derivative, then optimal cluster head number is：

Obtain the energy sum E that network is consumed in each round_round。

The initial total energy E of network_totalIt is known, then the estimate of network lifecycle also can obtain：

R_tatal=E_total/E_round；

Ideally each node is roughly the same in the energy that each round is consumed, and each node of network is flat after r wheels Equal energy is：

Substitute intoIt is calculated probability-weighted H_i(r)。

Substitute into, obtain the probability threshold of each node.

To ensure that sensor node cluster head will not be overweight and dead because of traffic load, the dump energy threshold value for setting cluster head is E_min, when node present energy is less than threshold value, it is no longer participate in the selection of cluster head.

Step S104 of the invention is specifically included：

After cluster is formed, into the stabilization sub stage, stable state operation is made up of several time frames.

Each leader cluster node serves as expeditor and control centre's effect of data transfer in respective cluster, and leader cluster node is The data transmission conflict between cluster interior nodes is avoided, the packet that cluster interior nodes are gathered is received successively using time division multiplex strategy.

Be processed the data of each node in cluster such as data fusion by leader cluster node, centralized calculation etc., then will be after processing Data is activation give certain gateway node, the gateway node is transferred to s ink nodes again after data are reprocessed.

The step of data fusion, specifically includes：

Aggregators of the sensor node as the cluster, i.e. cluster head are selected in each cluster；

Determine the data fusion mode of the network, including：

(1) data fusion in layer.Data fusion is that the sensor that receives in the cluster of each cluster head of this layer is adopted in layer The all information for collecting, and the data fusion for carrying out；

(2) inter-layer data fusion.Inter-layer data fusion is that the aggregators of this layer receive other layer of all letters of transmission Breath, and the data fusion for carrying out；

Determine the data transfer mode of network, including：

The comentropy that sensor in each cluster is collected it is transferred to the aggregators of the cluster；

The cluster head of each cluster gives the information transfer after data fusion in the layer and inter-layer data fusion than bag Layer where including the cluster apart from sink nodes closer to layer aggregators；

The aggregators of each layer only have in the completion layer after data fusion and inter-layer data fusion Continue to transmit the information after the data fusion；

Comentropy after the data fusion is finally delivered to the aggregation node.

Presently preferred embodiments of the present invention is the foregoing is only, is not intended to limit the invention, it is all in essence of the invention Any modification, equivalent and improvement made within god and principle etc., should be included within the scope of the present invention.

Claims (5)

1. the cluster head in a kind of energy heterogeneous wireless sensor network is elected and its shift process, it is characterised in that the energy Cluster head election and its shift process in heterogeneous wireless sensor network are divided into startup stage and stabilization sub stage；In startup stage, Sensor node selects random number, random number to be less than threshold value T (n_i) node as cluster；In stabilization sub stage, node continuous collecting Data are simultaneously transmitted to cluster head, and the data that cluster head transmits the node in the cluster are sent to sink nodes after carrying out fusion treatment；By Different, the threshold value T (n in each node energy_i) also different, each round chooses cluster head using the method for cluster head rotation.

2. the cluster head in energy heterogeneous wireless sensor network as claimed in claim 1 is elected and its shift process, its feature It is that the cluster head election and its shift process in the energy heterogeneous wireless sensor network are comprised the following steps：

Step one, cluster head chooses the stage, and the node different for primary power, the threshold value of generation is also different, in active state Sensor node can generate random number between 0-1, if this number is less than its corresponding threshold values T (n), the node is elected to It is cluster head；

Step 2, the establishment stage of cluster, after cluster head node is produced, actively to this message of all node broadcasts in network；Not by The sensor node for being selected as cluster head selects optimal cluster head, and sends the solicited message for adding the cluster；After cluster head receives request, Node is set to member in cluster；After cluster is formed, cluster head informs to be each node distribution one in cluster by broadcast mode Tdma communication time slot；Only in one's own time slot, member node just can send data to cluster head node；

Step 3, transmits the stabilization sub stage of data, for aggregation and processing data, route data；Node gathered data, to data Merged.

3. the cluster head in energy heterogeneous wireless sensor network as claimed in claim 2 is elected and its shift process, its feature It is that the step 2 is specifically included：

(1) PN leader cluster node is elected in each round, wherein P is optimization cluster head ratio, is also probability-weighted；Each node Decided whether to turn into leader cluster node by following probability threshold：

$T\left(n_i\right)=\left\{\begin{array}{c}\frac{P_{opt}}{1-P_{opt}\[r\mathrm{mod}(1/P_{opt})\]},\∀n_i\∈G\\ 0,\∀n_i\∉G\end{array}\right.;$

Wherein, r is current wheel number, and G is nearestDo not have to turn into the node set of cluster head in wheel；Each node Have the opportunity to turn into the more leader cluster node of consumed energy in turn；

E₀Represent the primary power of ordinary node, a₁,a₂,...,a_nThe ratio shared by n kind special joints, b are represented respectively₁, b₂,...,b_nRepresent that special joint primary power exceedes the multiple of ordinary node primary power respectively；

a₁N,a₂N,...,a_nThe primary power of N number of special joint is respectively E₀(1+b₁),E₀(1+b₂),...,E₀(1+b_n), it is left (1-a₁-a₂,...-a_n) N number of ordinary node primary power be E₀(1+b_n)；

The total primary power of multi-tier Heterogeneous network is：

$\begin{array}{c}{E}_{total}=E_0(1+b_1)a_{1}N+E_0(1+b_2)a_{2}N+...+E_0(1+b_n)a_{n}N\\ +E_0(1-a_1-a_2-...-a_n)N\\ =E_{0}N(1+a_1{b}_1+a_2{b}_2+...+a_n{b}_n)\\ =E_{0}N(1+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i{b}_i)\end{array};$

N number of sensor node is evenly distributed on the border circular areas that a radius is A at random, and sink nodes are located in the middle of region, often One take turns send data to cluster head during the gross energy that is consumed be：

$E_{round}=k(2{\mathrm{NE}}_{elec}+{\mathrm{NE}}_{DA}+{\mathrm{l\ϵ}}_{amp}{d^4}_{\sin k}+{\mathrm{N\ϵ}}_{fs}{{\stackrel{\‾}{d}}_{CH}}^2)$

Wherein, l is cluster head number, E_elecThe energy consumed per bit data during operation transmission circuit or receiving circuit is represented, E_DAThe cost of data fusion is performed for cluster head,It is cluster head to the average distance of sink nodes,It is bunch member node to cluster The average distance of head node, ε_ampd⁴ _sinkWithIt is the energy of amplifier consumption：

It is calculated

To E_roundLocal derviation is sought on l, and it is 0 to make the partial derivative, then optimal cluster head number is：

$l_{opt}=\sqrt{\frac{N}{2}}\sqrt{\frac{{\mathrm{\ϵ}}_{fs}}{{\mathrm{\ϵ}}_{amp}}}\frac{A}{{d^2}_{\sin k}};$

Obtain the energy sum E that network is consumed in each round_round；Meanwhile,

The initial total energy E of network_total, it is known that R_totalIt is the estimate of network lifecycle, also can obtain：

R_tatal=E_total/E_round；

Different probability-weighted P are taken according to its primary power to this n+1 kinds node_i：

$P_1=\frac{1+b_1}{1+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i{b}_i},P_2=\frac{1+b_2}{1+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i{b}_i},...,P_n=\frac{1+b_n}{1+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i{b}_i},P_0=\frac{1}{1+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i{b}_i},$

(2) the present energy E that node i is taken turns in r_iR () chooses its cluster head T turnaround time_i,Represent what network was taken turns in r Average energy, withAs reference energy and the present energy E of node_iR () is made comparisons, obtain：

$H_i\left(r\right)=P_{opt}\frac{E_i\left(r\right)}{\stackrel{\‾}{E}\left(r\right)}{P}_i;$

Wherein, P_optIt is optimization cluster head ratio；

The average energy of each node of network is after r wheels：

$\stackrel{\‾}{E}\left(r\right)=\frac{E_{total}}{N}(1-\frac{r}{R_{total}});$

Substitute intoIt is calculated probability-weighted H_i(r)；

$H_1\left(r\right)=P_{opt}\frac{E_1\left(r\right)}{\stackrel{\‾}{E}\left(r\right)}\frac{1+b_1}{1+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i{b}_i},H_2\left(r\right)=P_{opt}\frac{E_2\left(r\right)}{\stackrel{\‾}{E}\left(r\right)}\frac{1+b_2}{1+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i{b}_i},...,$

$H_n\left(r\right)=P_{opt}\frac{E_n\left(r\right)}{\stackrel{\‾}{E}\left(r\right)}\frac{1+b_n}{1+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i{b}_i},H_2\left(r\right)=P_{opt}\frac{E_0\left(r\right)}{\stackrel{\‾}{E}\left(r\right)}\frac{1}{1+\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{a}_i{b}_i},$

Obtain the probability threshold of each node.

4. the cluster head in energy heterogeneous wireless sensor network as claimed in claim 2 is elected and its shift process, its feature It is that the step 4 data fusion is specifically included：

(1) data fusion mode of the network is determined, including：

Data fusion in layer, data fusion is the institute that collects of the sensor that receives in the cluster of each cluster head of this layer in layer There is information, and the data fusion for carrying out；

Inter-layer data is merged, and inter-layer data fusion is that the aggregators of this layer receive other layer of all information of transmission, and is entered Capable data fusion；

(2) data transfer mode of network is determined, including：

The comentropy that sensor in each cluster is collected it is transferred to the aggregators of the cluster；

The cluster head of each cluster gives the information transfer after data fusion in the layer and inter-layer data fusion than including institute Layer where stating cluster apart from sink nodes closer to layer aggregators；

The aggregators of each layer only have data fusion and the inter-layer data in the completion layer just to continue after merging Transmit the information after the data fusion；

Comentropy after the data fusion is finally delivered to the aggregation node.

5. in energy heterogeneous wireless sensor network described in a kind of utilization claim 1~2 any one cluster head election and its The wireless sensor network of shift process.