CN102625404B - Distributed routing protocol method for three-dimensional underwater acoustic sensor networks - Google Patents

Abstract

Disclosed is a distributed routing protocol method for three-dimensional underwater acoustic sensor networks. Supposing that each node knows its own geographic location and channels are bidirectional and symmetrical, transponding cost is calculated and transponding nodes are selected according to factors such as the node location, utilization rate of the channel, energy, and the like when data transponding is required. Node energy is used as effectively as possible on the premise that other network performances are not affected, and balanced use of the energy of each node and the utilization rate of the underwater acoustic channel are considered. The purpose is to use the channel effectively so as to minimize energy consumption, improve the utilization rate of the energy and the channel of the networks, and prolong the service life of the whole networks.

Description

A kind of distributed routing protocol method that is applied to three-dimensional water sound sensor network

Technical field

The present invention relates to the Routing Protocol of sensor network network layer, relate in particular to a kind of distributed routing protocol method that is applied to three-dimensional water sound sensor network, belong to Underwater acoustic signal processing technical field.

Background technology

Network end-to-end postpones, throughput and energy consumption are the basic technical indicators that embodies network performance, tradition Routing Protocol adopts shortest path first more, between two nodes, find a path that distance is the shortest as optimal path, as dijkstra's algorithm, A* (A-star) algorithm, D* (D-star) algorithm, bellman-ford algorithm etc., this type of algorithm is mostly only taking internodal distance as cost restrictive condition, do not consider some limiting factors of other side, therefore, only adopt a few paths the shortest to send information, in network, be easy to bump and cause packets lost, and the re-transmission of packets of information can make energy consumption increase, the unbalanced use of energy, can make some node exhaust too early energy, cause link failure, reduce the end-to-end delay of network, throughput and energy consume performance, shorten the life span of network.

Summary of the invention

The object of the invention is to overcome the defect of prior art, a kind of distributed routing protocol method that is applied to three-dimensional water sound sensor network is proposed, the method is calculated and is forwarded cost selection forward node according to factors such as the position of node, channel utilization and energy, and the computational methods of the forwarding cost of having derived, can effectively improve capacity usage ratio, the channel utilization of network, extend the life span of network.

Above-mentioned purpose of the present invention is achieved in that a kind of distributed routing protocol method that is applied to three-dimensional water sound sensor network, it is characterized in that: this method supposes that each node knows the geographical position of oneself, and channel is bi-directional symmetrical, in the time of needs forwarding data, on the basis of existing network topology structure, calculate and forward cost selection forward node according to factors such as the position of node, channel utilization and energy, if this forward node is destination node, data are directly sent to this destination node, data transmission procedure finishes; If this forward node is not destination node, send the data to this forward node, this forward node is received after data, continue to calculate according to factors such as the position of node, channel utilization and energy the forward node that forwards cost selection oneself, until data are finally sent to destination node, data transmission procedure finishes.

Specifically can carry out according to the following steps:

1) network carries out initialization, adopts the distributed node of three-dimensional water sound sensor network to lay algorithm and sets up network topology structure information, mainly contains following four-stage:

A) sub-clustering

In this stage, when node is initial, be placed on seabed by random cloth, carry out sub-clustering according to the group character of node (ID).Basic thought is that the each node in adjacent domain is selected the ID (CID) of ID maximum in this region as this bunch.This just means that each node will jump neighbors exchange id with one, and preserves a neighbors ID table.

B) grouping

After node has divided bunch, bunch head is overlapping according to the covering between bunch interior nodes, in each bunch, node is divided into groups.Need to be existed two overlapping nodes of sensor coverage to distribute different packet ID (GID) in the grouping stage, their cloth is put into the different degree of depth overlapping to eliminate this.Know and ID and the position of bunch interior nodes therefore start to distribute successively GID due to bunch head, select a new node at every turn, check that this node and other have distributed the internodal sensor coverage of GID overlapping.In this way, bunch head can ensure to distribute different GID to having arbitrarily two overlapping nodes of sensor coverage.Once a bunch head completes this process, the each node in will giving bunch sends the information that includes node GID.

C) depth assignment

Overlapping in order to reduce to cover, improve three-dimension integrally coverage, join the different degree of depth will to each component, computational process is as follows: any two not the space between be on the same group wherein, the degree of depth that D is seawater, G is the number of grouping.This space will guarantee that the distance between every layer (having the xy plane of certain degree of depth) and other layer is between top layer and sea, between bottom and seabed also all within this distance.And this grouping allows not cover overlapping node cloth and is placed on identical depth layer.This computational process is completed by a bunch head, and each node in depth information being sent to bunch by bunch head.

D) additional cycles

Each node is determined from own nearest neighbors according to distance, and is checked between own and this neighbors whether exist sensor coverage overlapping.If this arest neighbors node is in the awareness coverage of oneself, node continues mobile.In the time not having obvious coverage can promote or reach certain periodicity, movement will stop.In general, after two cycles the overall coverage of network improve no longer obvious, so additional cycles number is made as to 2.

2) setting total call duration time is T;

3) all nodes of network enter communication process, have first judged whether that data need to send, if had, enter transmission state, execution step 4); Otherwise, enter accepting state, execution step 5);

4) at transmission state:

A) first judge whether to arrive call duration time T, if so, whole communication process finishes, execution step 6); Otherwise, continue down to carry out;

B) network node has data to need to send, in step 1) on the network topology structure basis set up, to calculate and forward cost and select forward node according to the factor such as the position of node, channel utilization and energy, the computational methods that forward cost are as follows:

$E_i^{\left(j\right)}=E_{\mathrm{ij}}\·{\hat{N}}_{\mathrm{ij}}^t\·{\hat{N}}_i^{\mathrm{hop}}---\left(1\right)$

In formula, estimate for the forwarding cost of link (i, j), while representing that node i selects node j to send packet as forward node to destination node, successfully send the required energy of 1bit and estimate, formed by following three factors:

1. E _ijrepresent that node i sends the required energy of 1bit data to node j.Can be expressed from the next:

$E_{\mathrm{ij}}=2\·E_{\mathrm{elec}}^b+\min (\frac{P_{T_i}^{\max }}{R},{\mathrm{TL}}_{\mathrm{ij}}\·E_{\min }^b)---\left(2\right)$

In formula, the meaning of variable is as follows:

wherein, represent respectively send or receive the intrinsic energy loss of every bit data electronic device, with range-independence.

the maximum transmit power of node i while being R for data transfer rate.

TL _ijfor node i is to the loss of node j, can be calculated by underwater sound propagation energy consumption model, that is:

In shallow sea, loss computing formula is:

TL＝10logd+α(f)d×10 ^-3 (3)

In formula, d represents the distance between source node and receiving node, and α (f) represents absorption coefficient, means that sound intensity propagation per unit distance is because of attenuation by absorption α (f) decibel, and unit is dB/km, and the unit of TL is dB.

In deep-sea, loss computing formula is:

TL＝20logd+α(f)d×10 ^-3 (4)

In formula, absorption coefficient (f) numerical values recited is relevant with frequency used.

For frequency more than hundreds of hertz, α (f) computing formula is:

$\mathrm{\α}\left(f\right)=\frac{0.1f^2}{1+f^2}+\frac{40f^2}{4100+f^2}+2.75\×{10}^{-4}{f}^2+0.003---\left(5\right)$

In formula, f represents signal post's frequency, and unit is kHz.

For low frequency, α (f) computing formula is:

$\mathrm{\α}\left(f\right)=\frac{0.11f^2}{1+f^2}+0.011f^2+0.002---\left(6\right)$

for a given noise grade, the least energy loss that the every bit data of recipient's decoding success are required.

2. being illustrated in link (i, j) bit error rate is BER _ijsituation under, until before recipient j decoding success, the average time of the required transmission of packet of long L.The utilance of underwater acoustic channel, the energy efficiency of link depend primarily on required average transmission times.Can be expressed from the next:

${\hat{N}}_{\mathrm{ij}}^t=(1-{\mathrm{PER}}_{\mathrm{ij}})\·\underset{n=1}{\overset{\∞}{\mathrm{\Σ}}}n\·{{\mathrm{PER}}_{\mathrm{ij}}}^{n-1}---\left(7\right)$

Wherein:

PER _ij＝1-(1-BER _ij) ^L (8)

BER _ij, PER _ijbe respectively bit error rate and the Packet Error Ratio of link (i, j), n is total transmission times of a bag.

Underwater acoustic channel utilance can be defined as:

$\mathrm{\η}=\frac{\frac{L^D}{{\hat{N}}_{\mathrm{ij}}^t\·T_{\mathrm{rtt}}}}{R}---\left(9\right)$

In formula, represent the time that round trip is used, wherein, represent respectively the transmission time of packet header, valid data, ACK bag, T _prepresent the propagation time, suppose that wanting the valid data length of transmission package is L ^d, packet header length is L ^h, ACK packet length is L ^a, node communication scope is r, and the velocity of sound is q, and data transfer rate is R, has definition for the active link capacity between sender and recipient, represent accessible mean bit rate under uncontested medium access control scheme.So formula (9) can be written as again

$\mathrm{\η}=\frac{L^D}{{\hat{N}}_{\mathrm{ij}}^t\·[L^H+L^D+L^A+R\·\left(2\frac{r}{q}\right)]}---\left(10\right)$

3. represent the average number of hops from node i to destination node, can be expressed from the next:

${\hat{N}}_i^{\mathrm{hop}}=\max (\frac{\underset{k=1}{\overset{m}{\mathrm{\Σ}}}{H}_k}{m},1)---\left(11\right)$

M is the total path number of node i to destination node, H _kfor node i is to the jumping figure of the k paths of destination node.

Node i is selected node j ^*as forward node, if:

$j^{*}=\arg {\min }_{j^{*}\∈S_i\∩P_i^s}{E}_i^{\left(j\right)}---\left(12\right)$

In formula, S _ifor the neighbors collection of node i, for the set of node nearer apart from destination node s than node i apart from destination node s, if i.e. d _js< d _is, be forward node j ^*for the neighbors repeating cost of node i minimum node set;

C) whether intercept channel busy, when channel idle, send data;

D) network node sends after data, waits for the other side's response within a two-way time;

If e) receive the other side's response message within a two-way time, illustrate that this secondary data sends successfully, reenter communication process, execution step 3);

5) in accepting state:

A) first judge whether to arrive call duration time T, if so, whole communication process finishes, execution step 6); Otherwise, continue down to carry out;

If b) receive data, first judge these data be to oneself data or need to be transmitted to the data of other node;

C) if give the data of oneself, just these data are preserved, and reply response message, reenter communication process, execution step 3);

D) if need to be transmitted to the data of other node, just that these data are temporary, reply response message, and enter into transmission state, forward this data, execution step 4);

6) sign off.

Compared with prior art, tool of the present invention has the following advantages and remarkable result:

In cordless communication network, Routing Protocol adopts shortest path first more, between two nodes, find a path that distance is the shortest as optimal path, this type of algorithm is mostly only taking internodal distance as cost restrictive condition, do not consider some limiting factors of other side, but in water sound sensor network communication environment complexity, the energy content of battery of node is finite sum preciousness extremely, therefore, on route optimization criterion, not only to consider and adopt shortest path or time delay, delay inequality criterion, also must consider channel utilization, the equilibrium of each node battery consumption between amount of battery consumption and network, the Routing Protocol of using for reference existing wireless communications network carries out brand-new design to the Routing Protocol of water sound sensor network.The inventive method is according to selecting factors forward node such as the position of node, channel utilization and energy, the computational methods of forward node have been provided by mathematical derivation, do not affecting under the prerequisite of other network performance, effectively utilize as far as possible node energy, and consider that each node energy equilibrium is used and the problem of underwater acoustic channel utilance, improve capacity usage ratio, the channel utilization of network, extend the life span of network.The present invention easily realizes, and has higher actual application value.

1) according to selecting factors forward node such as the position of node, channel utilization and energy, effectively utilize as far as possible node energy, and consider that each node energy equilibrium is used and the problem of underwater acoustic channel utilance, more traditional Routing Protocol method has higher network performance;

2), by mathematical derivation, the computational methods of forward node have been provided, for Project Realization has been established theoretical foundation;

3) this method has good realizability.

Brief description of the drawings

Fig. 1 is FB(flow block) of the present invention;

Fig. 2 is that distributed node lays algorithm four-stage schematic diagram, wherein, a) being initial network schematic diagram, is b) to node clustering schematic diagram according to node ID, c) being overlapping to node grouping schematic diagram according to sensor coverage, is d) to adjust degree of depth schematic diagram according to grouping;

Fig. 3 is that distributed node lays algorithm flow block diagram;

Fig. 4 is the topological structure schematic diagram that network simulation adopts;

Fig. 5 is the contrast that two kinds of average end-to-end delays of Routing Protocol lower network change with average load;

Fig. 6 is the contrast that two kinds of Routing Protocol lower network throughputs change with average load;

Fig. 7 is the contrast that two kinds of Routing Protocol lower network average energy consumption change with average load.

Embodiment

Referring to Fig. 1, suppose that each node knows the geographical position of oneself, and channel is bi-directional symmetrical, in the time of needs forwarding data, on the basis of existing network topology structure, calculate and forward cost selection forward node according to factors such as the position of node, channel utilization and energy, if this forward node is destination node, data are directly sent to this destination node, data transmission procedure finishes; If this forward node is not destination node, send the data to this forward node, this forward node is received after data, continue to calculate according to factors such as the position of node, channel utilization and energy the forward node that forwards cost selection oneself, until data are finally sent to destination node, data transmission procedure finishes.

Specifically can carry out according to the following steps:

1) network carries out initialization, adopts the distributed node of three-dimensional water sound sensor network to lay algorithm and sets up network topology structure information, mainly contains following four-stage:

A) sub-clustering

In this stage, when node is initial, be placed on seabed by random cloth, carry out sub-clustering according to the group character of node (ID).Basic thought is that the each node in adjacent domain is selected the ID (CID) of ID maximum in this region as this bunch.This just means that each node will jump neighbors exchange id with one, and preserves a neighbors ID table.

B) grouping

After node has divided bunch, bunch head is overlapping according to the covering between bunch interior nodes, in each bunch, node is divided into groups.Need to be existed two overlapping nodes of sensor coverage to distribute different packet ID (GID) in the grouping stage, their cloth is put into the different degree of depth overlapping to eliminate this.Know and ID and the position of bunch interior nodes therefore start to distribute successively GID due to bunch head, select a new node at every turn, check that this node and other have distributed the internodal sensor coverage of GID overlapping.In this way, bunch head can ensure to distribute different GID to having arbitrarily two overlapping nodes of sensor coverage.Once a bunch head completes this process, the each node in will giving bunch sends the information that includes node GID.

C) depth assignment

Overlapping in order to reduce to cover, improve three-dimension integrally coverage, join the different degree of depth will to each component, computational process is as follows: any two not the space between be on the same group wherein, the degree of depth that D is seawater, G is the number of grouping.This space will guarantee that the distance between every layer (having the xy plane of certain degree of depth) and other layer is between top layer and sea, between bottom and seabed also all within this distance.And this grouping allows not cover overlapping node cloth and is placed on identical depth layer.This computational process is completed by a bunch head, and each node in depth information being sent to bunch by bunch head.

D) additional cycles

Each node is determined from own nearest neighbors according to distance, and is checked between own and this neighbors whether exist sensor coverage overlapping.If this arest neighbors node is in the awareness coverage of oneself, node continues mobile.In the time not having obvious coverage can promote or reach certain periodicity, movement will stop.In general, after two cycles the overall coverage of network improve no longer obvious, so additional cycles number is made as to 2.

2) setting total call duration time is T;

3) all nodes of network enter communication process, have first judged whether that data need to send, if had, enter transmission state, execution step 4); Otherwise, enter accepting state, execution step 5);

4) at transmission state:

A) first judge whether to arrive call duration time T, if so, whole communication process finishes, execution step 6); Otherwise, continue down to carry out;

B) network node has data to need to send, in step 1) on the network topology structure basis set up, to calculate and forward cost and select forward node according to the factor such as the position of node, channel utilization and energy, the computational methods that forward cost are as follows:

$E_i^{\left(j\right)}=E_{\mathrm{ij}}\&CenterDot;{\hat{N}}_{\mathrm{ij}}^t\&CenterDot;{\hat{N}}_i^{\mathrm{hop}}---\left(1\right)$

In formula, estimate for the forwarding cost of link (i, j), while representing that node i selects node j to send packet as forward node to destination node, successfully send the required energy of 1bit and estimate, formed by following three factors:

1. E _ijrepresent that node i sends the required energy of 1bit data to node j.Can be expressed from the next:

$E_{\mathrm{ij}}=2\&CenterDot;E_{\mathrm{elec}}^b+\min (\frac{P_{T_i}^{\max }}{R},{\mathrm{TL}}_{\mathrm{ij}}\&CenterDot;E_{\min }^b)---\left(2\right)$

In formula, the meaning of variable is as follows:

wherein, represent respectively send or receive the intrinsic energy loss of every bit data electronic device, with range-independence.

the maximum transmit power of node i while being R for data transfer rate.

TL _ijfor node i is to the loss of node j, can be calculated by underwater sound propagation energy consumption model, that is:

In shallow sea, loss computing formula is:

TL＝10logd+α(f)d×10 ^-3 (3)

In formula, d represents the distance between source node and receiving node, and α (f) represents absorption coefficient, means that sound intensity propagation per unit distance is because of attenuation by absorption α (f) decibel, and unit is dB/km, and the unit of TL is dB.

In deep-sea, loss computing formula is:

TL＝20logd+α(f)d×10 ^-3 (4)

In formula, absorption coefficient (f) numerical values recited is relevant with frequency used.

For frequency more than hundreds of hertz, α (f) computing formula is:

$\mathrm{\&alpha;}\left(f\right)=\frac{0.1f^2}{1+f^2}+\frac{40f^2}{4100+f^2}+2.75\&times;{10}^{-4}{f}^2+0.003---\left(5\right)$

In formula, f represents signal post's frequency, and unit is kHz.

For low frequency, α (f) computing formula is:

$\mathrm{\&alpha;}\left(f\right)=\frac{0.11f^2}{1+f^2}+0.011f^2+0.002---\left(6\right)$

for a given noise grade, the least energy loss that the every bit data of recipient's decoding success are required.

2. being illustrated in link (i, j) bit error rate is BER _ijsituation under, until before recipient j decoding success, the average time of the required transmission of packet of long L.The utilance of underwater acoustic channel, the energy efficiency of link depend primarily on required average transmission times.Can be expressed from the next:

${\hat{N}}_{\mathrm{ij}}^t=(1-{\mathrm{PER}}_{\mathrm{ij}})\&CenterDot;\underset{n=1}{\overset{\&infin;}{\mathrm{\&Sigma;}}}n\&CenterDot;{{\mathrm{PER}}_{\mathrm{ij}}}^{n-1}---\left(7\right)$

Wherein:

PER _ij＝1-(1-BER _ij) ^L (8)

BER _ij, PER _ijbe respectively bit error rate and the Packet Error Ratio of link (i, j), n is total transmission times of a bag.

Underwater acoustic channel utilance can be defined as:

$\mathrm{\&eta;}=\frac{\frac{L^D}{{\hat{N}}_{\mathrm{ij}}^t\&CenterDot;T_{\mathrm{rtt}}}}{R}---\left(9\right)$

In formula, represent the time that round trip is used, wherein, represent respectively the transmission time of packet header, valid data, ACK bag, T _prepresent the propagation time, suppose that wanting the valid data length of transmission package is L ^d, packet header length is L ^h, ACK packet length is L ^a, node communication scope is r, and the velocity of sound is q, and data transfer rate is R, has definition for the active link capacity between sender and recipient, represent accessible mean bit rate under uncontested medium access control scheme.So formula (9) can be written as again

$\mathrm{\&eta;}=\frac{L^D}{{\hat{N}}_{\mathrm{ij}}^t\&CenterDot;[L^H+L^D+L^A+R\&CenterDot;\left(2\frac{r}{q}\right)]}---\left(10\right)$

3. represent the average number of hops from node i to destination node, can be expressed from the next:

${\hat{N}}_i^{\mathrm{hop}}=\max (\frac{\underset{k=1}{\overset{m}{\mathrm{\&Sigma;}}}{H}_k}{m},1)---\left(11\right)$

M is the total path number of node i to destination node, H _kfor node i is to the jumping figure of the k paths of destination node.

Node i is selected node j ^*as forward node, if:

$j^{*}=\arg {\min }_{j^{*}\&Element;S_i\&cap;P_i^s}{E}_i^{\left(j\right)}---\left(12\right)$

In formula, S _ifor the neighbors collection of node i, for the set of node nearer apart from destination node s than node i apart from destination node s, if i.e. d _js< d _is, be forward node j ^*for the neighbors repeating cost of node i minimum node set;

C) whether intercept channel busy, when channel idle, send data;

D) network node sends after data, waits for the other side's response within a two-way time;

If e) receive the other side's response message within a two-way time, illustrate that this secondary data sends successfully, reenter communication process, execution step 3);

5) in accepting state:

A) first judge whether to arrive call duration time T, if so, whole communication process finishes, execution step 6); Otherwise, continue down to carry out;

If b) receive data, first judge these data be to oneself data or need to be transmitted to the data of other node;

C) if give the data of oneself, just these data are preserved, and reply response message, reenter communication process, execution step 3);

D) if need to be transmitted to the data of other node, just that these data are temporary, reply response message, and enter into transmission state, forward this data, execution step 4);

6) sign off.

Referring to Fig. 2, a) initial network schematic diagram, b) according to node ID to node clustering schematic diagram, c) according to sensor coverage overlapping to node grouping schematic diagram and d) according to grouping adjust the degree of depth schematic view illustrating distributed node lay the four-stage of algorithm.In the sub-clustering stage, at the bottom of Two-dimensional Sea, node is carried out to sub-clustering according to the ID of node.In the grouping stage, overlapping according to the covering between node and its neighbors, in each bunch, node is divided into groups.In the depth assignment stage, bunch head of each bunch calculates packet depth, and each node in this depth information is sent to bunch, covering in two-dimensional network is overlapped in three-dimensional network and reduced.Finally, in the additional cycles stage, node further moves, overlapping to reduce the covering that still exists of after date of initial movable week.

Referring to Fig. 3, for distributed node lays algorithm flow block diagram.Initialization node ID, carries out sub-clustering.Bunch head is carried out a task of distributing GID to bunch interior nodes, distributes after GID, calculates the degree of depth of each GID, and node in giving bunch sends this depth information, notifies their positions that should move for the first time.Meanwhile, bunch head is also carried out moving for the first time of oneself, adjusts the degree of depth.Non-leader cluster node, receives after the depth information that bunch hair send, and carries out moving for the first time of oneself, adjusts the degree of depth.After initial movable completes, carry out additional cycles to improve overall coverage range.The additional cycles number of times of repetition is made as to 2, and in additional cycles, each node determination, from own nearest neighbors position, if distance is less than the sensing range s of transducer, moves away from this neighbors.

Referring to Fig. 4, for the topological structure that network simulation adopts, lay algorithm by distributed node and calculate, node is distributed in the scope of 8km × 8km × 5km under water, and the communication range of node is r=2km, and the distance between node is all in communication range.In network, there are two category nodes: sensor node and gateway node, totally 16.In emulation, MAC layer adopts identical agreement, and physical layer adopts identical wireless pipe model, and network layer contrasts distributed routing protocol and shortest path Routing Protocol.Simulation parameter is with reference to the parameter setting of LinkQuest modulator-demodulator: the velocity of sound is q=1500m/s, and data transfer rate is R=320bps, and valid data length is L ^d=960bits, packet header length is L ^h=64bits, ACK packet length is L ^a=64bits, total length of data packets is L=L ^h+ L ^d=1024bits, central task frequency is f=10kHz, transmitted power is P _tX=15W, received power is P _rX=0.9W, intercepting power is P _sEN=8mW.Packet generation time interval of emulation transaction flow obeys and is uniformly distributed, altogether emulation 4 groups of transaction flow, the time of advent interval average 30 seconds respectively, 60 seconds, 90 seconds, 120 seconds.The average end-to-end delay of network, throughput and the average energy consumption of distributed routing protocol and shortest path Routing Protocol are obtained respectively by emulation with the comparison diagram 5～7 of average load variation.

Referring to Fig. 5, it is the contrast that two kinds of average end-to-end delays of Routing Protocol lower network change with average load.As seen from the figure, under low load, both end-to-end average retardations are close, along with the increase of offered load, network conflict aggravation, Frame is constantly retransmitted, and end-to-end delay all increases widely, after increasing to a certain degree, all tend towards stability, but distributed routing protocol reach the load ability to bear of network bottleneck will be apparently higher than shortest path Routing Protocol.Generally speaking, the average end-to-end delay of the network of distributed routing protocol is all the time lower than shortest path Routing Protocol, and the average end-to-end delay performance of distributed routing protocol is better than shortest path Routing Protocol.

Referring to Fig. 6, it is the contrast that two kinds of Routing Protocol lower network throughputs change with average load.As seen from the figure, under low load, network does not almost have the conflict of Frame, both have shown approximate performance, and along with the increase of load, both throughputs all constantly increase, in the time that load increases to a certain degree, network conflict aggravation, throughput is all tending towards saturated rear decline, but the throughput of distributed routing protocol is all the time higher than shortest path Routing Protocol.As can be seen here, distributed routing protocol has more excellent network throughput performance with respect to shortest path Routing Protocol.

Referring to Fig. 7, it is the contrast that two kinds of Routing Protocol lower network average energy consumption change with average load.Because only adopting a few paths the shortest, shortest path Routing Protocol transmits information, in network, be easy to bump and cause packets lost, and the re-transmission of packets of information can make energy consumption increase, the unbalanced use of energy, can make some node exhaust too early energy, cause link failure.And distributed routing protocol takes into full account the selecting factors forward node such as position, channel utilization and the energy of node, the blindness that reduces information forwards, and energy consumption balance, to each node, has been reduced to the average energy consumption of network.Generally speaking, the network average energy consumption of distributed routing protocol is all the time lower than shortest path Routing Protocol, and the energy efficiency of distributed routing protocol is better than shortest path Routing Protocol.

Can find out by above simulation result and analysis: distributed routing protocol is according to selecting factors forward node such as the position of node, channel utilization and energy, make network obtain optimization in various degree in the performances such as average end-to-end delay, throughput, average energy consumption, improve the utilance of channel, the balanced energy consumption of network, realize the steady decline of network energy, the life span that has effectively extended network.

Claims (1)

1. one kind is applied to the distributed routing protocol method of three-dimensional water sound sensor network, it is characterized in that: suppose that each node knows the geographical position of oneself, and channel is bi-directional symmetrical, in the time of needs forwarding data, on the basis of existing network topology structure, according to the factor including position, channel utilization and the energy of node, forward cost by calculating, select forward node, if this forward node is destination node, data are directly sent to this destination node, data transmission procedure finishes; If this forward node is not destination node, send the data to this forward node, this forward node is received after data, continue according to the factor including position, channel utilization and the energy of node, calculate and forward cost, select the forward node of oneself, until data are finally sent to destination node, data transmission procedure finishes, and comprises the following steps:

1) network carries out initialization, adopts the distributed node of three-dimensional water sound sensor network to lay algorithm and sets up network topology structure information, has following four-stage:

A) sub-clustering

In this stage, when node is initial, be placed on seabed by random cloth, carry out sub-clustering according to the group character ID of node, basic thought is that the each node in adjacent domain is selected the ID of ID maximum in this region as this bunch, bunch mark CID, this just means that each node will jump neighbors exchange id with one, and preserves a neighbors ID table;

B) grouping

After node has divided bunch, bunch head is overlapping according to the covering between bunch interior nodes, in each bunch, node is divided into groups, need to be existed two overlapping nodes of sensor coverage to distribute different packet ID in the grouping stage, the i.e. mark GID of group, their cloth is put into the different degree of depth overlapping to eliminate this, because bunch head is known ID and the position of bunch interior nodes, therefore start to distribute successively GID, a new node of each selection, check that this node and other have distributed the internodal sensor coverage of GID overlapping, in this way, bunch head can ensure to distribute different GID to having arbitrarily two overlapping nodes of sensor coverage, once a bunch head completes this process, each node in will giving bunch sends the information that includes node GID,

C) depth assignment

Overlapping in order to reduce to cover, improve three-dimension integrally coverage, join the different degree of depth will to each component, computational process is as follows: any two not the space between be on the same group wherein, the degree of depth that D is seawater, G is the number of grouping, and this space will be guaranteed every layer, and the distance having between xy plane and other layer of certain degree of depth is between top layer and sea, between bottom and seabed also all within this distance, and, this grouping allows not cover overlapping node cloth and is placed on identical depth layer, and this computational process is completed by a bunch head, and each node in depth information being sent to bunch by bunch head;

D) additional cycles

Each node is determined from own nearest neighbors according to distance, and check between own and this neighbors whether exist sensor coverage overlapping, if this arest neighbors node is in the awareness coverage of oneself, node continues mobile, in the time not having obvious coverage can promote or reach certain periodicity, movement will stop, and after two cycles, the overall coverage of network improves no longer obviously, and additional cycles number is made as to 2;

2) setting total call duration time is T;

3) all nodes of network enter communication process, have first judged whether that data need to send, if had, enter transmission state, execution step 4); Otherwise, enter accepting state, execution step 5);

4) at transmission state:

A) first judge whether to arrive call duration time T, if so, whole communication process finishes, execution step 6); Otherwise, continue down to carry out;

B) network node has data to need to send, in step 1) on the network topology structure basis set up, calculate and forward cost according to the position of node, channel utilization and capacity factor, select forward node, the computational methods that forward cost are as follows:

$E_i^{\left(j\right)}=E_{\mathrm{ij}}\&CenterDot;{\hat{N}}_{\mathrm{ij}}^t\&CenterDot;{\hat{N}}_i^{\mathrm{hop}}---\left(1\right)$

In formula, estimate for the forwarding cost of link (i, j), while representing that node i selects node j to send packet as forward node to destination node, successfully send the required energy of 1bit and estimate, formed by following three factors:

1. E _ijrepresent that node i sends the required energy of 1bit data to node j, is expressed from the next:

$E_{\mathrm{ij}}=2\&CenterDot;E_{\mathrm{elec}}^b+\min (\frac{P_{T_i}^{\max }}{R},{\mathrm{TL}}_{\mathrm{ij}}\&CenterDot;E_{\min }^b)---\left(2\right)$

In formula, the meaning of variable is as follows:

wherein, represent respectively send or receive the intrinsic energy loss of every bit data electronic device, with range-independence;

the maximum transmit power of node i while being R for data transfer rate;

TL _ijfor node i is to the loss of node j, can be calculated by underwater sound propagation energy consumption model, that is:

In shallow sea, loss computing formula is:

TL＝10logd+α(f)d×10 ^-3 (3)

In formula, d represents the distance between source node and receiving node, and α (f) represents absorption coefficient, means that sound intensity propagation per unit distance is because of attenuation by absorption α (f) decibel, and unit is dB/km, and the unit of TL is dB;

In deep-sea, loss computing formula is:

TL＝20logd+α(f)d×10 ^-3 (4)

In formula, absorption coefficient (f) numerical values recited is relevant with frequency used;

For frequency more than hundreds of hertz, α (f) computing formula is:

$\mathrm{\&alpha;}\left(f\right)=\frac{0.1f^2}{1+f^2}+\frac{40f^2}{4100+f^2}+2.75\&times;{10}^{-4}{f}^2+0.003---\left(5\right)$

In formula, f represents signal post's frequency, and unit is kHz;

For low frequency, α (f) computing formula is:

$\mathrm{\&alpha;}\left(f\right)=\frac{0.11f^2}{1+f^2}+0.011f^2+0.002---\left(6\right)$

for a given noise grade, the least energy loss that the every bit data of recipient's decoding success are required;

2. being illustrated in link (i, j) bit error rate is BER _ijsituation under, until before recipient j decoding success, the average time of the required transmission of packet of long L, the utilance of underwater acoustic channel, the energy efficiency of link depend primarily on required average transmission times, are expressed from the next:

${\hat{N}}_{\mathrm{ij}}^t=(1-{\mathrm{PER}}_{\mathrm{ij}})\&CenterDot;\underset{n=1}{\overset{\&infin;}{\mathrm{\&Sigma;}}}n\&CenterDot;{{\mathrm{PER}}_{\mathrm{ij}}}^{n-1}---\left(7\right)$

Wherein:

PER _ij＝1-(1-BER _ij) ^L (8)

BER _ij, PER _ijbe respectively bit error rate and the Packet Error Ratio of link (i, j), n is total transmission times of a bag;

Underwater acoustic channel utilance can be defined as:

$\mathrm{\&eta;}=\frac{\frac{L^D}{{\hat{N}}_{\mathrm{ij}}^t\&CenterDot;T_{\mathrm{rtt}}}}{R}---\left(9\right)$

In formula, represent the time that round trip is used, wherein, represent respectively the transmission time of packet header, valid data, ACK bag, T _prepresent the propagation time, suppose that wanting the valid data length of transmission package is L ^d, packet header length is L ^h, ACK packet length is L ^a, node communication scope is r, and the velocity of sound is q, and data transfer rate is R, has definition for the active link capacity between sender and recipient, represent accessible mean bit rate under uncontested medium access control scheme, so formula (9) can be written as again

$\mathrm{\&eta;}=\frac{L^D}{{\hat{N}}_{\mathrm{ij}}^t\&CenterDot;[L^H+L^D+L^A+R\&CenterDot;\left(2\frac{r}{q}\right)]}---\left(10\right)$

3. represent the average number of hops from node i to destination node, can be expressed from the next:

${\hat{N}}_i^{\mathrm{hop}}=\max (\frac{\underset{k=1}{\overset{m}{\mathrm{\&Sigma;}}}{H}_k}{m},1)---\left(11\right)$

M is the total path number of node i to destination node, H _kfor node i is to the jumping figure of the k paths of destination node;

Node i is selected node j ^*as forward node, if:

$j^{*}=\arg {\min }_{j^{*}\&Element;S_i\&cap;P_i^s}{E}_i^{\left(j\right)}---\left(12\right)$

In formula, S _ifor the neighbors collection of node i, for the set of node nearer apart from destination node s than node i apart from destination node s, if i.e. d _js< d _is, be forward node j ^*for the neighbors repeating cost of node i minimum node set;

C) whether intercept channel busy, when channel idle, send data;

D) network node sends after data, waits for the other side's response within a two-way time;

If e) receive the other side's response message within a two-way time, illustrate that this secondary data sends successfully, reenter communication process, execution step 3);

5) in accepting state:

A) first judge whether to arrive call duration time T, if so, whole communication process finishes, execution step 6); Otherwise, continue down to carry out;

If b) receive data, first judge these data be to oneself data or need to be transmitted to the data of other node;

C) if give the data of oneself, just these data are preserved, and reply response message, reenter communication process, execution step 3);

D) if need to be transmitted to the data of other node, just that these data are temporary, reply response message, and enter into transmission state, forward this data, execution step 4);

6) sign off.