CN104093196A - LEACH alternating time dynamic optimization method based on energy consumption - Google Patents

Abstract

The invention relates to an LEACH alternating time dynamic optimization method based on energy consumption and belongs to the technical field of the wireless sensor network. The LEACH alternating time dynamic optimization method based on energy consumption includes the following steps of calculating energy consumption of nodes in an LEACH protocol every turn firstly, calculating the time of duration of each turn according to the energy consumption of the turn, and dynamically adjusting the time of duration of clusters every turn according to the principle of balance of the energy consumption of the turns. By applying the alternating time dynamic optimization method based on energy consumption to the LEACH protocol, the cluster selection period can be effectively optimized, inter-node energy consumption can be effectively balanced, and the life cycle of the network can be effectively prolonged.

Description

A kind of LEACH rotation time dynamic optimization method based on energy consumption

Technical field

The dynamic optimization method that the present invention relates to a kind of Leach rotation time based on energy consumption, belongs to wireless sensor network technology field.

Background technology

The life cycle that extends network is important problem in a radio sensing network.A lot of scholars have also proposed the life cycle that a lot of energy-efficient agreements extend radio sensing network.At present, the equilibrium that realizes energy by network cluster dividing realizes and extends the focus direction that network lifecycle is research with this.LEACH algorithm is the most popular energy-efficient communication protocol based on sub-clustering being proposed by people such as Heinzelman, and agreement is by carrying out sub-clustering and balancing energy being loaded to the consumption that reduces gross energy in network in each bunch to node.In order to ensure the equilibrium consumption of energy, LEACH periodically selects at random node to serve as a bunch head in all nodes.LEACH agreement realizes with " wheel ".Each is taken turns and comprises: establishment stage and stable operation stage.Establishment stage to complete bunch forming process adding of the selection of leader cluster node, the broadcast of leader cluster node, non-leader cluster node and leader cluster node be bunch in the TDMA scheduling process of member's distribution T DMA time slot; In stable operation stage, bunch head is accepted to send to base station from its member's message and by the data after polymerization.But many parameters of LEACH can affect the performance of agreement, these parameters have to be optimized, as, threshold values, bunch number etc.Therefore the parameter that, many scholars are devoted to optimize LEACH improves the performance of LEACH.Although a lot of people have done different optimization to LEACH, each takes turns the but few people's research of lasting time LEACH.Each length of taking turns the duration is very crucial for the whole performance of network.If the time is oversize, bunch head is just for a long time in active state so, and the energy of leader cluster node will very fast being exhausted; If the time is too short, the selection of bunch head too frequently causes too much energy dissipation at a bunch establishment stage, because this one-phase is to send data.

Summary of the invention

The object of the invention is the defect existing in order to overcome prior art, proposed a kind of Leach rotation time dynamic optimization method based on energy consumption.

Thought of the present invention is the difference consuming according to bunch interior nodes quantity and energy, adjusts dynamically every duration of taking turns, and optimizes energy consumption between choosing bunch cycle, the balance node of LEACH, extends the life cycle of network with this.

The object of the invention is to be achieved through the following technical solutions:

A LEACH rotation time dynamic optimization method based on energy consumption, comprises the following steps:

Step 1, calculate in LEACH agreement that node is every takes turns energy consumption, comprise following content:

The 1.1st step: the energy consumption of calculating a bunch node in every wheel according to following formula:

$\begin{array}{c}{E}_{i\_\mathrm{CH}/\mathrm{frame}}=l{E}_{\mathrm{elec}}(n_i-1)+l{E}_{\mathrm{DA}}{n}_i+l{E}_{\mathrm{elec}}+l{\mathrm{\ϵ}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4\\ =l{E}_{\mathrm{elec}}{n}_i+l{E}_{\mathrm{DA}}{n}_i+l{\mathrm{\ϵ}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4\end{array};---\left(1\right)$

E _{i_CH/round}＝E _{i_CH/frame}×N _frames/round； (2)

Wherein E _{i_CH/frame}represent the energy consumption of every frame of leader cluster node, E _{i_CH/round}represent the every energy consumption of taking turns of leader cluster node, n _irepresent the number of nodes that bunch i comprises, l represents bit number contained in each packets of information, E _elecrepresent to propagate a Bit data institute consumed energy, E _dArepresent polymerization one Bit data institute energy requirement, d _{i_toBS}represent the distance of leader cluster node i to base station, ε _amprepresent as d>=d ₀time radio frequency amplifier transmission 1bit unit square rice square energy consuming, d ₀for the threshold values distance presupposing, N _frames/roundrepresent every quantity of taking turns transmitting data frame;

The 1.2nd step: the energy consumption of calculating a non-bunch of node in every wheel according to following formula:

$E_{k\_\mathrm{non}-\mathrm{CH}/\mathrm{frame}}=l{E}_{\mathrm{elec}}+l{\mathrm{\ϵ}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2;---\left(3\right)$

E _{k_non-CH/round}＝E _{k_non-CH/frame}×N _frames/round； (4)

Wherein E _{k_non-CH/frame}the energy consumption of every frame of member node in representing bunch, d _{k_toCH}in representative bunch, member node k is to the distance of a bunch node, ε _fsrepresent as d < d ₀time the emission amplifier transmission 1bit unit square rice energy that consumes, E _{k_non-CH/round}represent bunch in the every energy consumption of taking turns of member node;

Step 2, the every wheel duration of calculating LEACH agreement, comprise following content:

The 2.1st step: calculate the gross energy that in every wheel, bunch i consumes according to following formula:

$\begin{array}{c}{E}_{i\_\mathrm{total}}=E_{i\_\mathrm{CH}/\mathrm{round}}+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}{E}_{k\_\mathrm{non}-\mathrm{CH}/\mathrm{round}}\\ =N_{\mathrm{frames}/\mathrm{round}}\&times;[(l{E}_{\mathrm{elec}}\&CenterDot;n_i+l{E}_{\mathrm{DA}}\&CenterDot;n_i+l{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4)+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(l{E}_{\mathrm{elec}}+l{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)]\end{array};---\left(5\right)$

Wherein, E _{i_total}represent the every energy consumption of taking turns of bunch i;

The 2.2nd step: the data volume of calculating bunch i transmission in every wheel according to following formula:

$N_{\mathrm{frames}/\mathrm{round}}=E_{i\_\mathrm{total}}/[(l{E}_{\mathrm{elec}}\&CenterDot;n_i+l{E}_{\mathrm{DA}}\&CenterDot;n_i+l{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4)+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(l{E}_{\mathrm{elec}}+l{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)];---\left(6\right)$

The 2.3rd step: calculate the duration that each is taken turns according to following formula:

$t_{i\_\mathrm{round}}$ $=\frac{l}{r_b}\&CenterDot;n_i\&CenterDot;\frac{E_{i\_\mathrm{total}}}{l{E}_{\mathrm{elec}}\&CenterDot;n_i+l{E}_{\mathrm{DA}}\&CenterDot;n_i+l{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(l{E}_{\mathrm{elec}}+l{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)};---\left(7\right)$

Wherein r _bfor bit rate;

Step 3, dynamically adjust each bunch of every duration of taking turns, comprise following content:

The 3.1st step: the further energy consumption of compute cluster i in one takes turns:

The every energy consumption of taking turns of bunch i by formula (7) is:

$E_{i\_\mathrm{total}}=\frac{[E_{\mathrm{elec}}\&CenterDot;n_i+E_{\mathrm{DA}}\&CenterDot;n_i+{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(E_{\mathrm{elec}}+{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)]\&CenterDot;r_b\&CenterDot;t_{i\_\mathrm{round}}}{n_i};---\left(8\right)$

The 3.2nd step: dynamically adjust every the wheel duration of each bunch, comprise content:

Find out the conduct maximum bunch that bunch internal segment is counted at most, dump energy is maximum in each bunch of front-wheel, be designated as a bunch j, every duration t that takes turns of bunch j _{j_round}adjust according to formula (10), the time of other bunch is adjusted according to formula (11):

t _{j_round}＝t _round(E _{nj_current}/E _{nj_init})； (10)

$t_{i\_\mathrm{round}}=\frac{E_{\mathrm{elec}}\&CenterDot;n_j+E_{\mathrm{DA}}\&CenterDot;n_j+{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{j\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_j-1}{\mathrm{\&Sigma;}}}(E_{\mathrm{elec}}+{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)}{E_{\mathrm{elec}}\&CenterDot;n_i+E_{\mathrm{DA}}\&CenterDot;n_i+{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(E_{\mathrm{elec}}+{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)}\&CenterDot;\left(\frac{E_{\mathrm{ni}\_\mathrm{current}}}{E_{\mathrm{ni}\_\mathrm{init}}}\right)\&CenterDot;{\left(\frac{n_i}{n_j}\right)}^2\&CenterDot;t_{j\_\mathrm{round}};---\left(11\right)$

By t _{j_round}be set as the choosing bunch cycle of each node, the non-maximum bunch every duration t that takes turns in its regulation _{i_round}inside carry out transfer of data, in official hour after end of transmission, dormancy t _{j_round}-t _{i_round}time, to save node energy, then enters next round process together with maximum cluster knot point.

Beneficial effect

Contrast traditional LEACH agreement, according to provided by the invention bunch of rotation time dynamic adjusting method, the energy consumption of balanced net interior nodes effectively, extend the quantity of network life cycle and base station reception packets of information; In addition, the method is applicable to fixing all basic LEACH agreement and the improved protocol thereof of wheel time proposing at present, applied widely.

Brief description of the drawings

Fig. 1 is LEACH agreement topology diagram.

Fig. 2 is the flow chart of the inventive method.

Fig. 3 is in the scene of 50,100,200 nodes, the comparison diagram of the situation lower life cycle of different rotation times.

Fig. 4 is under different rotation times, the comparison diagram of the node life span of 100 node scenes.

Fig. 5 is under different rotation times, the comparison diagram of the rate of energy dissipation of 100 node scenes.

Fig. 6 is under different rotation times, and the base station of 100 node scenes receives the comparison diagram of data volume.

Fig. 7 is rotation time while being 20s, the comparison diagram of the LEACH algorithm before and after improving in network life cycle.

Fig. 8 is rotation time while being 20s, the comparison diagram of the LEACH algorithm before and after improving in node life span.

Fig. 9 is rotation time while being 20s, the comparison diagram of the LEACH algorithm before and after improving in node average energy consumption.

Figure 10 is rotation time while being 20s, and the LEACH algorithm before and after improving receives the comparison diagram in data volume in base station.

Figure 11 is rotation time while being 20s, the comparison diagram of the LEACH algorithm before and after improving on first node death time and half node death time.

Embodiment

Below in conjunction with drawings and Examples, the present invention will be further described.

As shown in Figure 1, each leader cluster node intercoms with base station LEACH agreement topological diagram of the present invention mutually, and each bunch of interior nodes communicated by letter mutually with leader cluster node in this bunch.Introduce principle, implementation process and the assessment result of the inventive method below.

As shown in Figure 2, principle is as follows for the flow chart of the inventive method:

Step 1, calculate in LEACH agreement that node is every takes turns energy consumption:

Its operating procedure comprises the 1.1st and 1.2 steps, is specially:

The 1.1st step: calculate in every wheel the energy consumption of a bunch node:

Because base station is far from sensitive zones, we think this distance much larger than energy transmission consume critical distance, therefore multichannel consumption models is followed in the consumption of energy.The energy consumption of leader cluster node is mainly used in receiving the data of member node transmission and the data after polymerization being sent to base station, so the consumption of leader cluster node energy is shown below:

$\begin{array}{c}{E}_{i\_\mathrm{CH}/\mathrm{frame}}=l{E}_{\mathrm{elec}}(n_i-1)+l{E}_{\mathrm{DA}}{n}_i+l{E}_{\mathrm{elec}}+l{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4\\ =l{E}_{\mathrm{elec}}{n}_i+l{E}_{\mathrm{DA}}{n}_i+l{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4\end{array}---\left(1\right)$

E _{i_CH/round}＝E _{i_CH/frame}×N _frames/round (2)

Wherein E _{i_CH/frame}represent the energy consumption of every frame of leader cluster node, E _{i_CH/round}represent the every energy consumption of taking turns of leader cluster node, n _irepresent the number of nodes that bunch i comprises, l represents bit number contained in each packets of information, E _elecrepresent to propagate a Bit data institute consumed energy, E _dArepresent polymerization one Bit data institute energy requirement, d _{i_toBS}represent the distance of leader cluster node i to base station, ε _amprepresent as d>=d ₀time radio frequency amplifier transmission 1bit unit square rice square energy consuming, d ₀for the threshold values distance presupposing, N _frames/roundrepresent every quantity of taking turns transmitting data frame.

The 1.2nd step: calculate in every wheel the energy consumption of a non-bunch of node:

In bunch, member node is just responsible for data to pass to a corresponding bunch of node.We think bunch in distance between member node and a corresponding bunch of node relatively short, Friss free-space model is followed in the consumption of energy.So the consumption of member node energy is shown below in bunch:

$E_{k\_\mathrm{non}-\mathrm{CH}/\mathrm{frame}}=l{E}_{\mathrm{elec}}+l{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2---\left(3\right)$

E _{k_non-CH/round}＝E _{k_non-CH/frame}×N _frames/round (4)

Wherein E _{k_non-CH/frame}the energy consumption of every frame of member node in representing bunch, d _{k_toCH}in representative bunch, member node k is to the distance of a bunch node, ε _fsrepresent as d < d ₀time the emission amplifier transmission 1bit unit square rice energy that consumes, E _{k_non-CH/round}represent bunch in the every energy consumption of taking turns of member node.

Step 2, the every wheel duration of calculating LEACH agreement:

On the basis of step 1 operation, further calculate that LEACH agreement is every takes turns the duration, its operating procedure comprises the 2.1st step to the 2.3 steps, is specially:

The 2.1st step: calculate the gross energy that in every wheel, bunch i consumes:

Draw leader cluster node and bunch in after the every energy of taking turns of member node consumes, we can draw the gross energy that bunch i consumes in a rotational cycle:

$\begin{array}{c}{E}_{i\_\mathrm{total}}=E_{i\_\mathrm{CH}/\mathrm{round}}+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}{E}_{k\_\mathrm{non}-\mathrm{CH}/\mathrm{round}}\\ =N_{\mathrm{frames}/\mathrm{round}}\&times;[(l{E}_{\mathrm{elec}}\&CenterDot;n_i+l{E}_{\mathrm{DA}}\&CenterDot;n_i+l{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4)+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(l{E}_{\mathrm{elec}}+l{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)]\end{array}---\left(5\right)$

Wherein, E _{i_total}represent the every energy consumption of taking turns of bunch i.

The 2.2nd step: the data volume of calculating bunch i transmission in every wheel:

Can be obtained by formula (5), in a rotational cycle, the data volume of bunch i transmission is:

$N_{\mathrm{frames}/\mathrm{round}}=E_{i\_\mathrm{total}}/[(l{E}_{\mathrm{elec}}\&CenterDot;n_i+l{E}_{\mathrm{DA}}\&CenterDot;n_i+l{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4)+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(l{E}_{\mathrm{elec}}+l{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)]---\left(6\right)$

The 2.3rd step: calculate the duration that each is taken turns:

Suppose r _bfor bit rate, this is the parameter just setting before algorithm starts, and, in a distributed time slot, the data of node transmission l bit are consuming time is one contains n _ibunch time that transmission one frame data will consume of individual node is the time of bunch i work one-period:

$t_{i\_\mathrm{round}}=t_{\mathrm{frame}}\&CenterDot;N_{\mathrm{frames}/\mathrm{round}}$ $=\frac{l}{r_b}\&CenterDot;n_i\&CenterDot;\frac{E_{i\_\mathrm{total}}}{l{E}_{\mathrm{elec}}\&CenterDot;n_i+l{E}_{\mathrm{DA}}\&CenterDot;n_i+l{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(l{E}_{\mathrm{elec}}+l{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)}---\left(7\right)$

Step 3, dynamically adjust each bunch of every duration of taking turns:

On the basis of step 2, each bunch according to the difference of energy consumption and bunch interior nodes quantity, dynamically adjusts every wheel the duration, and its operating procedure comprises the 3.1st and 3.2 steps, is specially:

The 3.1st step: the further energy consumption of compute cluster i in one takes turns:

From formula (7), the every energy consumption of taking turns of bunch i is:

$E_{i\_\mathrm{total}}=\frac{[E_{\mathrm{elec}}\&CenterDot;n_i+E_{\mathrm{DA}}\&CenterDot;n_i+{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(E_{\mathrm{elec}}+{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)]\&CenterDot;r_b\&CenterDot;t_{i\_\mathrm{round}}}{n_i}---\left(8\right)$

For the energy consumption between each bunch of equilibrium, need to be according to dynamic every the wheel duration of adjusting each bunch of number of nodes.

The 3.2nd step: dynamically adjust every the wheel duration of each bunch:

The interstitial content that each bunch of front-wheel is worked as in order is set { n ₁, n ₂..., n _k, wherein i=1,2 ..., k, k is bunch number when front-wheel.Suppose that bunch j is for maximum bunch, its contained nodes is n _j=max{n ₁, n ₂..., n _k, if there is the nodes of multiple bunches to be n _j, maximum bunch of the conduct of dump energy maximum in selecting bunch.The initial wheel set of time of maximum bunch is t _round, this is the fiducial time of all bunches of references.It is n that bunch i comprises node number _i, its wheel time is t _{i_round}, balanced for node energy is consumed, therefore the average energy consumption of member node equates there is following formula in order bunch:

$\frac{E_{i\_\mathrm{total}}}{n_i}=\frac{E_{j\_\mathrm{total}}}{n_j}\&DoubleRightArrow;\frac{n_j}{n_i}=\frac{E_{j\_\mathrm{total}}}{E_{i\_\mathrm{total}}}---\left(9\right)$

Formula (8) is brought in formula (9), can be according to t fiducial time _roundthe every time of taking turns that draws i bunch is t _{i_round}, consider in network that each bunch every takes turns the duration and should reduce and reduce along with the dump energy of each bunch, therefore, every duration t that takes turns of maximum bunch j _{j_round}adjust according to formula (10), similar, the wheel time t of other bunches _{i_round}adjust according to formula (11):

t _{j_round}＝t _round(E _{nj_current}/E _{nj_init}) (10)

$t_{i\_\mathrm{round}}=\frac{E_{\mathrm{elec}}\&CenterDot;n_j+E_{\mathrm{DA}}\&CenterDot;n_j+{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{j\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_j-1}{\mathrm{\&Sigma;}}}(E_{\mathrm{elec}}+{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)}{E_{\mathrm{elec}}\&CenterDot;n_i+E_{\mathrm{DA}}\&CenterDot;n_i+{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(E_{\mathrm{elec}}+{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)}\&CenterDot;\left(\frac{E_{\mathrm{ni}\_\mathrm{current}}}{E_{\mathrm{ni}\_\mathrm{init}}}\right)\&CenterDot;{\left(\frac{n_i}{n_j}\right)}^2\&CenterDot;t_{j\_\mathrm{round}}---\left(11\right)$

From above formula, it is all to change with the quantity of bunch interior nodes and the difference of dump energy that different bunches every takes turns the duration, so does more even that the node energy that can make in takes turns in network consumes.But the choosing bunch cycle of the each node of LEACH protocol requirement is consistent, therefore we select the wheel time t of maximum bunch j _{j_round}for the choosing bunch cycle of each node, remaining bunch is at every duration t that takes turns of its regulation _{i_round}inside carry out transfer of data, in official hour, after end of transmission, understand dormancy t _{j_round}-t _{i_round}time, to save node energy, until be waken up the choosing bunch of carrying out next round.

The implementation process of the inventive method is as follows:

Step 1, revise original LEACH protocol code:

As shown in Figure 2, former LEACH agreement is made improvements.After improvement, we can carry out emulation experiment to improved LEACH algorithm on network analog platform NS2.

The workflow of LEACH Routing Protocol after improvement in one takes turns is:

1., before algorithm starts, it is t that the initial wheel time is set _round;

2. each node is selected a value between 0-1 at random, if selected value is less than some threshold values, this node becomes leader cluster node so; After selected leader cluster node, inform whole network by broadcast; Other nodes in network according to the signal strength signal intensity of the information of reception determine subordinate bunch, and add information to its transmission, this packets of information is containing node ID, node location and residue energy of node;

3. leader cluster node is to residue gross energy and the nodes information of all member node in base station report bunch;

4. base station is according to all bunches of information collecting, descending to a bunch sequence according to nodes and residue gross energy, and bunch elects maximum bunch as by what rank the first, maximum bunch of note bunch number be j;

5. base station is used following formula dynamically to adjust the epicycle duration of maximum bunch according to the residue gross energy of maximum bunch:

t _{j_round}＝t _round(E _{j_current}/E _{j_init})

Wherein, t _roundfor reference wheel time, t _{j_round}for bunch j is at the duration of epicycle, E _{j_current}for the residue gross energy of bunch j, E _{j_init}for the initial gross energy of bunch j;

Use following formula dynamically to adjust the data transmission period of other bunch in epicycle:

$t_{i\_\mathrm{round}}=\frac{E_{\mathrm{elec}}\&CenterDot;n_j+E_{\mathrm{DA}}\&CenterDot;n_j+{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{j\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_j-1}{\mathrm{\&Sigma;}}}(E_{\mathrm{elec}}+{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)}{E_{\mathrm{elec}}\&CenterDot;n_i+E_{\mathrm{DA}}\&CenterDot;n_i+{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(E_{\mathrm{elec}}+{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)}\&CenterDot;\left(\frac{E_{\mathrm{ni}\_\mathrm{current}}}{E_{\mathrm{ni}\_\mathrm{init}}}\right)\&CenterDot;{\left(\frac{n_i}{n_j}\right)}^2\&CenterDot;t_{j\_\mathrm{round}}$

Wherein, t _{i_round}for bunch i is at the data transmission period of epicycle, t _{j_round}for bunch j is at the duration of epicycle, E _{i_current}for the residue gross energy of bunch i, E _{i_init}for the initial gross energy of bunch i, n _ifor the nodes that bunch i comprises, n _jfor the nodes that bunch j comprises, E _elecrepresent to propagate the energy that a Bit data consumes, E _dArepresent polymerization one Bit data institute energy requirement, d _{i_toBS}the leader cluster node of representative bunch i is to the distance of base station, d _{k_toCH}in representative bunch, member node k is to the distance of a bunch node, ε _amprepresent as d>=d ₀time radio frequency amplifier transmission 1bit unit square rice square energy consuming, d ₀for the threshold values distance presupposing, ε _fsrepresent as d < d ₀time the radio frequency amplifier transmission 1bit unit square rice energy that consumes;

6. base station is by t _{i_round}and t _{j_round}beam back the leader cluster node of each bunch;

7. the leader cluster node of maximum bunch is at t _{j_round}in time, carry out TDMA scheduling, for bunch in each member node distribute data transmission time slot; The leader cluster node of other bunch is at t _{i_round}in time, carry out TDMA scheduling, for bunch in each member node distribute data transmission time slot;

8. in the time slot that member node is its distribution at leader cluster node in each bunch, to bunch head transmission data, leader cluster node carries out sending information to base station again after data fusion to the data that receive;

9. each bunch of i is at process t _{i_round}after time, dormancy t _{j_round}-t _{i_round}after time, be waken up; After this whole network enters bunch foundation and the stable operation stage of a new round, forwards step 2 to and restarts, until the residue node number of the whole network is less than set value k.

Step 2, corresponding experiment parameter is set:

The present invention tests scene: in the scope of 100 × 100m, and 50,100 or 200 nodes of random distribution, base station is positioned at the position of (50,175).The primary power of base station is unlimited, and the primary power of ordinary node is 2mJ, and follow-up can not supply.Radio reception data consumes energy E _elecbe set to 50nJ/bit, the required calculating energy of data fusion is set to 5nJ/bit, and a bunch ratio N/k is 5%, and data package size l is 4000bits, bit rate R _bfor 1Mbps, the energy ε of radio frequency amplifier _ampand ε _fsbe respectively 0.0013pJ/bit/m ⁴and 10pJ/bit/m ².

Step 3, operation agreement, analyze its performance:

With reference to Fig. 3, draw the network lifecycle of LEACH improvement algorithm under different rotation times.As can be seen from the figure, in the case of identical network configuration, the quantity difference of node, the position that the peak value of network lifecycle occurs is just different, and the few peak value of node morning of occurring, the evening that the peak value that node is many occurs.This is because along with the increasing of number of nodes, produce containing bunch interior nodes many bunch probability will increase, according to the every algorithm of taking turns the duration of dynamic adjustment in this paper, should suitably increase and just establish a bunch time, even to realize node energy consumption.As can be seen from the figure, peak value appears at respectively 16s, 20s and 26s.

With reference to Fig. 4, draw in the situation of different initial rotation times the existence situation of nodes.Can find out, in the time that bunch set of time of just establishing is 20s, the dead speed of node is the slowest, from figure, also can find out, even if just establish, bunch time changes to some extent, but the time of occurrence of first dead node differs and is not very large, this be because, adopt the LEACH consultation of dynamic wheel time dynamically to adjust every wheel the duration according to the energy of the residue of node and consumption, after some wheel, the wheel time can be adjusted to level proper under current energy environment, and the generation time of first dead node has been delayed in the energy consumption that this can balanced node.

With reference to Fig. 5, draw the situation that 100 meshed network energy consume, as can be seen from the figure, when bunch set of time of establishing is originally 20s, it is the slowest that energy consumes.And can also find out in the time that network energy is low-down, network can also continue longer a period of time of operation, this is because taked the every method of taking turns the time of dynamic change, when energy reduces to a certain degree, every duration of taking turns has reduced a lot, this just causes node energy can reduce slowly, therefore, in network energy approach exhaustion, can also move longer a period of time.

With reference to Fig. 6, draw in the situation that difference is just established bunch time, the comparison of the data packet number that base station receives, as can be seen from the figure, the data packet number that base station receives in the time that bunch time of just establishing is 20s is maximum.If take turns long that the duration arranges every, a bunch node energy consumes too fast and dead, leader cluster node place so, and the data of polymerization will tail off, thus the data that receive base station will tail off.If take turns the duration every too short, have so too much energy dissipation in the choosing bunch stage, and in this one-phase, node can not send data, thereby cause base station to receive that data reduce.

The performance of the LEACH algorithm before and after step 4, comparative analysis improve:

From the simulation result of former LEACH agreement, can find, in the scene of 100 nodes, former LEACH agreement is in the time that initial rotation time is 20s, and network life cycle is the longest.Therefore be necessary both, in the time impelling rotation time to be 20s, to do performance comparison analysis.In order conveniently to contrast, we are called S-LEACH (Static LEACH) algorithm the algorithm before optimizing, and the algorithm after optimization is called D-LEACH (Dynamic LEACH), and we analyze energy consumption, death time of data volume, first node and half node that base station receives etc. of life cycle to node, node.

With reference to Fig. 7, S-LEACH and the D-LEACH comparison diagram on network lifecycle, can find out, D-LEACH algorithm is set various just establishing under bunch time, network lifecycle all has larger improvement than S-LEACH algorithm, and the in the situation that of 100 nodes, both peak values all appear at bunch time of building when being 20s.

With reference to Fig. 8, S-LEACH and the D-LEACH comparison diagram in node rate of death, can find out, the life cycle of D-LEACH is than long 40% left and right of S-LEACH, and in there is first node death, S-LEACH general 200 take turns after network completely dead, but D-LEACH has but used 250 to take turns, this be because D-LEACH every take turns the duration along with bunch in dump energy and the dynamically adjustment of bunch interior nodes number, along with the continuous death of node, the number of bunch interior nodes can reduce relatively, every duration of taking turns also can decline relatively, so just can allow node energy reduce relatively slow, therefore in there is node death, D-LEACH can also continue for some time than S-LEACH more.

With reference to Fig. 9, S-LEACH and the D-LEACH comparison diagram in average energy consumption, can find, it is slow that the energy consumption rate S-LEACH of D-LEACH wants, the slope of S-LEACH is substantially constant in whole network life cycle, but D-LEACH is constantly to reduce at 510s front slope, this is due to the energy consumption along with node, often take turns the duration and decline to some extent, this will inevitably fall low-energy-consumption, but at 510s between 600s, the energy consumption of D-LEACH is accelerated suddenly, this be because, D-LEACH is in the time of 510s, first node death, and the threshold values of D-LEACH is chosen formula based on dump energy, along with the propelling of time, residue energy of node can be lower, cause the threshold values of node smaller, it is just smaller that node is chosen as the probability of bunch head, in there is no node death, this consequence is not also clearly, but in the time having node death, can be chosen as the quantity of leader cluster node just still less, be elected to bunch head when there is no node in a network, or node does not join in any bunch, node all can directly send information to base station, this process can consume a large amount of energy.The ratio that therefore can cause energy consumption to increase is very fast.And in 600s, the energy consumption of node can slow down, this be because, the average residual energy comparison of node is low, causes every duration of taking turns very short, the energy that node is every takes turns consumption just seldom, therefore can also continue to grow a period of time.

With reference to Figure 10, comparison diagram aspect the data volume that S-LEACH and D-LEACH receive in base station, in figure, the data volume that D-LEACH receives is compared and is remained basically stable with S-LEACH, but the growth rate of data volume is smaller, this be because, in order to allow, node energy is balanced to be consumed D-LEACH, containing node few bunch often take turns the duration and will lack, the data volume sending will correspondingly reduce, this can cause the every data total amount of taking turns transmission of whole network fewer than S-LEACH, and As time goes on, the slope of D-LEACH slowly reduces, during to about 640s, data volume no longer increases, although network is also continuing operation, but do not play due effect, therefore we can think under present case, the life cycle of D-LEACH is 640s.

With reference to Figure 11, S-LEACH and D-LEACH are at the comparison diagram aspect first node death time and half node death time two, can find out, the time of first node death of D-LEACH is late more a lot of than S-LEACH, this be because the equilibrium of D-LEACH algorithm the consumption of node energy, this can postpone the time of first node death greatly.Under the contrast of half node death time, D-LEACH is also better than S-LEACH, but also can find, what between the HNA of D-LEACH and FND, differ is less, this is because due to even energy consumption, in there is first node death, the energy level of other nodes has also reduced a lot, and therefore the ratio S-LEACH of part of nodes death is fast.

In sum, the present invention is better than former LEACH agreement in multinomial performance index.

Above-described specific descriptions; object, technical scheme and beneficial effect to invention further describe; institute is understood that; the foregoing is only specific embodiments of the invention; the protection range being not intended to limit the present invention; within the spirit and principles in the present invention all, any amendment of making, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.

Claims (1)

1. the LEACH rotation time dynamic optimization method based on energy consumption, is characterized in that, comprises the following steps:

Step 1, calculate in LEACH agreement that node is every takes turns energy consumption, comprise following content:

The 1.1st step: the energy consumption of calculating a bunch node in every wheel according to following formula:

$\begin{array}{c}{E}_{i\_\mathrm{CH}/\mathrm{frame}}=l{E}_{\mathrm{elec}}(n_i-1)+l{E}_{\mathrm{DA}}{n}_i+l{E}_{\mathrm{elec}}+l{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4\\ =l{E}_{\mathrm{elec}}{n}_i+l{E}_{\mathrm{DA}}{n}_i+l{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4\end{array};---\left(1\right)$

E _{i_CH/round}＝E _{i_CH/frame}×N _frames/round； (2)

Wherein E _{i_CH/frame}represent the energy consumption of every frame of leader cluster node, E _{i_CH/round}represent the every energy consumption of taking turns of leader cluster node, n _irepresent the number of nodes that bunch i comprises, l represents bit number contained in each packets of information, E _elecrepresent to propagate a Bit data institute consumed energy, E _dArepresent polymerization one Bit data institute energy requirement, d _{i_toBS}represent the distance of leader cluster node i to base station, ε _amprepresent as d>=d ₀time radio frequency amplifier transmission 1bit unit square rice square energy consuming, d ₀for the threshold values distance presupposing, N _frames/roundrepresent every quantity of taking turns transmitting data frame;

The 1.2nd step: the energy consumption of calculating a non-bunch of node in every wheel according to following formula:

$E_{k\_\mathrm{non}-\mathrm{CH}/\mathrm{frame}}=l{E}_{\mathrm{elec}}+l{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2;---\left(3\right)$

E _{k_non-CH/round}＝E _{k_non-CH/frame}×N _frames/round； (4)

Wherein E _{k_non-CH/frame}the energy consumption of every frame of member node in representing bunch, d _{k_toCH}in representative bunch, member node k is to the distance of a bunch node, ε _fsrepresent as d < d ₀time the emission amplifier transmission 1bit unit square rice energy that consumes, E _{k_non-CH/round}represent bunch in the every energy consumption of taking turns of member node;

Step 2, the every wheel duration of calculating LEACH agreement, comprise following content:

The 2.1st step: calculate the gross energy that in every wheel, bunch i consumes according to following formula:

$E_{i\_\mathrm{total}}=E_{i\_\mathrm{CH}/\mathrm{round}}+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}{E}_{k\_\mathrm{non}-\mathrm{CH}/\mathrm{round}}$ $=N_{\mathrm{frames}/\mathrm{round}}\&times;[(l{E}_{\mathrm{elec}}\&CenterDot;n_i+l{E}_{\mathrm{DA}}\&CenterDot;n_i+l{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4)+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(l{E}_{\mathrm{elec}}+l{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)];---\left(5\right)$

Wherein, E _{i_total}represent the every energy consumption of taking turns of bunch i;

The 2.2nd step: the data volume of calculating bunch i transmission in every wheel according to following formula:

$N_{\mathrm{frames}/\mathrm{round}}=E_{i\_\mathrm{total}}/[(l{E}_{\mathrm{elec}}\&CenterDot;n_i+l{E}_{\mathrm{DA}}\&CenterDot;n_i+l{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4)+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(l{E}_{\mathrm{elec}}+l{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)];---\left(6\right)$

The 2.3rd step: calculate the duration that each is taken turns according to following formula:

$t_{i\_\mathrm{round}}=\frac{l}{r_b}\&CenterDot;n_i\&CenterDot;\frac{E_{i\_\mathrm{total}}}{l{E}_{\mathrm{elec}}\&CenterDot;n_i+l{E}_{\mathrm{DA}}\&CenterDot;n_i+l{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(l{E}_{\mathrm{elec}}+l{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)};---\left(7\right)$

Wherein r _bfor bit rate;

Step 3, dynamically adjust each bunch of every duration of taking turns, comprise following content:

The 3.1st step: the further energy consumption of compute cluster i in one takes turns:

The every energy consumption of taking turns of bunch i by formula (7) is:

$E_{i\_\mathrm{total}}=\frac{[E_{\mathrm{elec}}\&CenterDot;n_i+E_{\mathrm{DA}}\&CenterDot;n_i+{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(E_{\mathrm{elec}}+{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)]\&CenterDot;r_b\&CenterDot;t_{i\_\mathrm{round}}}{n_i};---\left(8\right)$

The 3.2nd step: dynamically adjust every the wheel duration of each bunch, comprise content:

Find out the conduct maximum bunch that bunch internal segment is counted at most, dump energy is maximum in each bunch of front-wheel, be designated as a bunch j, every duration t that takes turns of bunch j _{j_round}adjust according to formula (10), the time of other bunch is adjusted according to formula (11):

t _{j_round}＝t _round(E _{nj_current}/E _{nj_init})； (10)

$t_{i\_\mathrm{round}}=\frac{E_{\mathrm{elec}}\&CenterDot;n_j+E_{\mathrm{DA}}\&CenterDot;n_j+{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{j\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_j-1}{\mathrm{\&Sigma;}}}(E_{\mathrm{elec}}+{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)}{E_{\mathrm{elec}}\&CenterDot;n_i+E_{\mathrm{DA}}\&CenterDot;n_i+{\mathrm{\&epsiv;}}_{\mathrm{amp}}{d}_{i\_\mathrm{toBS}}^4+\underset{k=1}{\overset{n_i-1}{\mathrm{\&Sigma;}}}(E_{\mathrm{elec}}+{\mathrm{\&epsiv;}}_{\mathrm{fs}}{d}_{k\_\mathrm{toCH}}^2)}\&CenterDot;\left(\frac{E_{\mathrm{ni}\_\mathrm{current}}}{E_{\mathrm{ni}\_\mathrm{init}}}\right)\&CenterDot;{\left(\frac{n_i}{n_j}\right)}^2\&CenterDot;t_{j\_\mathrm{round}};---\left(11\right)$

By t _{j_round}be set as the choosing bunch cycle of each node, the non-maximum bunch every duration t that takes turns in its regulation _{i_round}inside carry out transfer of data, in official hour after end of transmission, dormancy t _{j_round}-t _{i_round}time, to save node energy, then enters next round process together with maximum cluster knot point.