CN103634913A - Clustering-based multi-hop multi-channel cognitive wireless sensor network communication method - Google Patents

Abstract

The invention relates to a clustering-based multi-hop multi-channel cognitive wireless sensor network communication method. In the method, the frame structure of a node is designed into a segmental structure, and a work period of the node is divided into a channel cognition and selection stage, a channel allocation stage, a data transmission stage and a dormant stage. According to the invention, a media access criteria is designed for the node in each cluster to provide intra-cluster uncontested communication, and at the same time, in order to alleviate problems, such as hidden terminals, etc., with inter-cluster communications, on the basis of the media access criteria, an assignment process in which channel weighing is performed according to the channel state is designed, and in order to better reduce the energy consumption, a channel selection criteria is further designed. The method of the invention can be used to realize advantages of high network throughput, low time delay and low packet loss rate on the basis of reducing the data packet transmission capacity, and is suitable for the spectrum access and spectrum switching strategy in a cognitive wireless sensor network.

Description

The communication means of the multi-hop multichannel cognition wireless sensing network based on sub-clustering

Technical field

What the present invention relates to is the method in a kind of wireless sensor network technology field, and the MAC agreement in especially a kind of cognition wireless sensor network based on sub-clustering is specifically related to the communication means of the multi-hop multichannel cognition wireless sensing network based on sub-clustering.

Background technology

The frequency range being operated in due to wireless sensor network is unauthorized frequency range, and along with this frequency range is more and more crowded, cognitive radio technology is introduced in wireless sensor network as solving the limited good technique of frequency spectrum resource.In conjunction with the feature of node energy finiteness in traditional wireless sensor networks, the MAC Protocol Design of the energy efficiency in cognition wireless sensor network (CRSN) is a primary study direction.

O.Boyinbode is disclosed in prior art, H.Le, A.Mbogho, M.Takizawa, with the document of R.Poliah " A survey on clustering algorithms for wireless sensor networks(Clustering Algorithm in Wireless Sensor Network outline) ", Proc.13th Int Network-Based Information Systems Conf, 2010, pp.358 – 364, it has carried out classification to the cluster algorithm of more current classical wireless sensor networks and has summed up and performance evaluation, describe the classical cluster algorithm such as LEACH and HEED in detail.

D.Cavalcanti is disclosed in prior art, S.Das, the document of Jianfeng Wang and K.Challapali " wireless sensor network of Cognitive Radio based Wireless Sensor Networks(based on cognitive radio) ", ICCCN, 2008, pp.1-6. with Ozgur B.Akan, Osman B.Karli, document " Cognitive Radio Sensor Networks (cognitive radio sensor network) " with Ozgur Ergul, IEEE Network, July/August 2009, pp.34-40, design idea and the main principle of CRSN network have been introduced, the framework of network, potential advantage and technical problem and challenge.Provided the research framework of CRSN network design.

In CRSN network owing to existing primary user and time user, in the MAC Protocol Design to CRSN network, a crucial technology is to make the wireless sensor network node as time user, when dynamic access channel, primary user is produced to least interference or noiseless.The document of Shui G and Shen S " in A New Multi-Channel MAC Protocol Combined with On-Demand Routing for Wireless Mesh Networks(wireless mesh network a kind of new combination as required the multi-channel MAC protocols of route) ", CSSE, 2008, pp.1-8, adopt the control channel method separated with data channel, by each node is equipped with to two antennas, the service condition of difference monitoring control channel and data channel, make time user can grasp constantly primary user's state, and can access in time available channel.In the situation that node is only equipped with single antenna, the document of Su Hang and Zhang Xi " CREAM-MAC:An efficient Cognitive Radio-EnAbled Multi-Channel MAC Protocol for Wireless Networks(CREAM-MAC: the efficient cognitive multi-channel MAC protocols that in wireless network, a kind of antenna enables) ", WOWMOM, 2008, pp.1-8, each node of middle employing is equipped with the transducer of an antenna and a plurality of detection primary user channels and avoids between inferior user and primary user and time user's mutual collision, thereby utilize better multichannel to communicate.

C.Li is disclosed in prior art, P.Wang, H.-H.Chen, with the document of M.Guizani " mono-kind of the A cluster based on-demand multi-channel mac protocol for wireless multimedia sensor networks(multichannel based on sub-clustering is wireless multimedia sensor network MAC agreement as required) ", Proc.IEEE Int.Conf.Communications ICC'08, 2008, pp.2371 – 2376, the MAC agreement of a WMSNs of multichannel on demand based on sub-clustering has been proposed, it is divided into three segmentations by the work period of whole node, make node support energy efficiency, the transfer of data of high-throughput and high reliability.

Compare with traditional MAC protocol for wireless sensor networks, the MAC agreement of at present relevant CRSN is wanted much less relatively.J.A.Han, W.S.Jeon, document " Energy-efficient channel management scheme for cognitive radio sensor networks (a kind of channel management mechanism of cognitive radio sensor network energy validity) " with D.G.Jeong, IEEE Trans-actions on Vehicular Technology, vol.60, no.4, pp.1905 – 1910, 2011, single bunch is analyzed, consider energy efficiency and primary user's impact, a new channel assignment scheme has been proposed, make CRSN can carry out according to the different result of channel-aware the selection of suitable mode of operation.

In sum, current most MAC agreement can not well directly be applied in multi-hop CRSN, reason is that the MAC protocol requirement node wherein having possesses a plurality of antennas, and due to the requirement of low-power consumption, wireless sensor network only allows node to be equipped with individual antenna conventionally; Some MAC agreements are not considered primary user's impact, can not be applied directly to the network coexisting with primary user; What have is to the single bunch of research of carrying out channel allocation Managed Solution, can not be applied in the CRSN of multi-hop.

Summary of the invention

The present invention is directed to existing MAC protocol application to above shortcomings in CRSN, taken into full account the interior primary user of network and time user's communication feature, under the environment existing primary user, inferior user effectively chance accesses channel and effectively communicates, alleviate user's interference that multi-hop multichannel concealed terminal brings simultaneously, and reduce time user's energy consumption as far as possible.The present invention can make wireless sensor node abundant access available channel in the network coexisting with primary user, on the basis that reduces transmission packet energy, realizes higher network throughput, lower time delay and packet loss.

Principle of the present invention is: the problem in cognition wireless sensor network has comprised multichannel hidden terminal problem and cognitive question, these two problems are all on channel, the former is that the node of two hop distances has been selected same channel, the latter is the existence due to primary user, and different channels no longer has same probability of use with respect to inferior user.Thereby by introducing the concept of channel weight (Channel Weight), be conducive to distinguish on the one hand different channels, channel that on the other hand also can envoy's point selection " better ", the i.e. larger channel of channel weight.

According to an aspect of the present invention, a kind of communication means of the multi-hop multichannel cognition wireless sensing network based on sub-clustering is provided, it is characterized in that, by the frame structure design of node, it is segmental structure, a work period (Interval) of node is divided into channel-aware and choice phase (Channel Sense and Selection Phase, CSSP), channel allocation stage (Channel Schedule Phase, CSP), data transfer phase (Data Transmission Phase, DTP) and the dormant stage (Sleep Phase, SP).

Preferably, each node has the channel weighted list of oneself, and the channel weight that channel weighted list is recorded reflects that node is for the behaviour in service of channel; The variation of channel weighted value is to occur while the different conditions of channel being detected by node, and with respect to node, channel has several states as follows:

-idle (Idle): this state appears at the channel-aware stage, time user awareness can be used to channel;

-busy (Busy): this state appears at the channel-aware stage equally, time user awareness is to the existence of primary user on channel, and channel is unavailable;

-communication (Communication): this state appears at time user and can use perceiving channel, and has normally used this channel to carry out transfer of data;

-collision (Collision): this state can occur in both cases, the first situation is that time user node in its adjacent cluster when data transfer phase is also being used this channel, the second situation is that time user primary user when transfer of data arrives suddenly, and both of these case all causes current channel unavailable.

Preferably, according to channel in different conditions, the concrete variation of the weighted value w of channel is not identical yet, when channel idle, weighted value increases W _idle; When channel is busy, weighted value reduces W _busy; When channel communication, weighted value increases W _com; When channel collision, corresponding the first and the second situation, weighted value reduces respectively W _secand W _pri.

Preferably, in channel-aware and choice phase, complete the process of channel-aware and selection, specifically comprise following steps:

Step (101): in the structure of time slot of channel-aware and choice phase, previous communication time slot in to be bunch head send to the own channel weighted list of safeguarding bunch all the other members to upgrade their channel weighted list, be now bunch head as transmit leg, a non-bunch of member is as recipient; A rear communication time slot is to prepare for channel associating perception, bunch in member own detected channel-aware result is all transferred to a bunch head, at bunch head, being in final channel-aware result carries out Combined Treatment;

Step (102): what the CSSP stage adopted is the mode of a part of channel in a perception N channel, thereby the quantity of minimizing channel-aware, but guarantee to choose best channel as far as possible.Particularly, from N channel set, { chN}, choose a channel subset as for can channel perception set (Polled-Channel Set) { S}, channel quantity in S is || S||=K, (K<N), node is determined available channel by the channel in perception S.The method that current K channel chosen can complete by methods such as enumerative technique or Dynamic Programmings.

Preferably, channel associating perception is the mode that adopts data fusion, particularly, one have M member bunch in, what after channel-aware that each bunch of interior nodes carries out, obtain judgment result is that D _i, final result of determination is by the data fusion mode D=Ψ (D adopting ₁, D ₂... D _m) decide.

Preferably, generally when inferior user detects primary user, use be the arrival that signal level is distinguished primary user, therefore the data fusion mode Ψ adopting here the M a detecting result is averaged, and with primary user's signal threshold value D _thresholdcompare, determine that whether available channel is; Bunch head is used final result of determination to complete the next channel allocation stage (Channel Schedule Phase, CSP).In some open schemes that propose, there is more mechanism how to predict primary user's behavior, therefore suppose that primary user's arrival probabilistic model is known here.

Preferably, in the channel allocation stage, bunch head has obtained final available channel list (Available Channel, AvaC) { AvaC}, bunch head is by { the processing of AvaC}, to bunch in each node carry out channel allocation, particularly, bunch head channel assignment scheme and transfer of data time slot allocation scheme are notified bunch in all the other nodes, wherein, the structure of time slot in CSP stage is exactly a communication time slot, and a bunch head is transmission side, and all the other nodes are recipients.

Preferably, in the channel allocation stage, bunch head also will for bunch in all the other nodes at data transfer phase (Data Transmission Phase, DTP), to distribute corresponding length be T _trtime slot.

Preferably, at data transfer phase, bunch interior nodes is carried out data communication at corresponding transfer of data time slot.The object of channel allocation is in order to complete channel handover mechanism, when node detects primary user in presently used channel, node is switched on other standby channel immediately, and wherein, the transmission side and the recipient that carry out data communication will understand the channel after the other side is switched mutually.

The present invention be first in each bunch for design of node medium access criterion with uncompetitive communication in providing bunch, this medium access criterion designs for more existing problems in multi-hop multi-channel wireless transducer, in order to alleviate the problems such as concealed terminal.Meanwhile, for reducing better energy consumption, also designed a channel selection criterion.This channel selection criterion is for so a kind of situation: at CRSN node, can in the situation that primary user exists, carry out under the condition of perception available channel, when CRSN node needs the channel number of perception more, its needed expense is larger, more with the energy that causes consuming.This channel selection criterion is exactly can obtain under the condition of some prioris at CRSN node, make CRSN node only need to carry out the channel-aware process of smaller amounts and obtain available channel, and guarantee that the result and the whole channels of perception that obtain approach, and reach the object of low-energy-consumption rate simultaneously as far as possible.The inventive method can realize higher network throughput on the basis that reduces transmission packet energy, and lower time delay and packet loss are applicable to frequency spectrum access and frequency spectrum switchover policy in cognition wireless sensor network.

Accompanying drawing explanation:

Fig. 1 is the frame assumption diagram of node in the present invention.

Fig. 2 is the structure of time slot figure of CSSP in the present invention.

Fig. 3 is CSP and the structure of time slot figure in DTP stage in the present invention.

Fig. 4 is channel assignment scheme sample chart figure in the present invention.

Fig. 5 is that scheme provided by the present invention and contrast scheme throughput are with the variation comparison diagram of data arrival rate.

Fig. 6 is that scheme provided by the present invention and contrast scheme packet loss are with the variation comparison diagram of data arrival rate.

Fig. 7 is that scheme provided by the present invention and contrast scheme average delay are with the variation comparison diagram of data arrival rate.

Fig. 8 is that scheme provided by the present invention and contrast scheme are successfully transmitted the average energy consumption of a packet with the variation comparison diagram of data arrival rate.

Fig. 9 is that scheme provided by the present invention and contrast scheme are successfully transmitted the average energy consumption of a packet with the variation comparison diagram of data arrival rate.

Embodiment:

Below in conjunction with drawings and Examples, technical scheme of the present invention is further described.

In one embodiment of the invention, each node has the channel weighted list of oneself, and channel weight reflects that node is for the behaviour in service of channel, thereby the maintenance that how to define channel weighted value is a very important problem.The variation of channel weighted value is to occur while the different conditions of channel being detected by node, and with respect to node, channel has several states as follows:

1) idle (Idle): this state appears at the channel-aware stage, time user awareness can be used to channel;

2) busy (Busy): this state appears at the channel-aware stage equally, time user awareness is to the existence of primary user on channel, and channel is unavailable;

3) communication (Communication): this state appears at time user and can use perceiving channel, and has normally used this channel to carry out transfer of data;

4) collision (Collision): this state can occur in both cases, the one, time user node in its adjacent cluster when data transfer phase is also being used this channel, another is exactly that time user primary user when transfer of data arrives suddenly, and both of these case all can cause current channel unavailable.

According to channel in different conditions, the concrete variation of the weighted value w of channel is not identical yet, when channel idle, weighted value increases W _idle; When channel is busy, weighted value reduces W _busy; When channel communication, weighted value increases W _com; When channel collision, corresponding two kinds of situations, weighted value reduces respectively W _secand W _pri.

By the frame structure design of node, it is segmental structure, a work period (Interval) of node is divided into channel-aware and choice phase (Channel Sense and Selection Phase, CSSP), channel allocation stage (Channel Schedule Phase, CSP), data transfer phase (Data Transmission Phase, DTP) and the dormant stage (Sleep Phase, SP).Each stage completes following function:

The CSSP stage: in CSSP structure of time slot, previous communication time slot in to be bunch head send to the own channel weighted list of safeguarding bunch all the other members to upgrade their channel weighted list, be now bunch head as transmit leg, a non-bunch of member is as recipient; A rear communication time slot is to prepare for channel associating perception, bunch in member own detected channel-aware result is all transferred to a bunch head, at bunch head, be in final channel-aware result and carry out Combined Treatment.Channel associating perception is the mode that adopts data fusion, as one have M member bunch in, what after channel-aware that each bunch of interior nodes carries out, obtain judgment result is that D _i, final result of determination is by the data fusion mode D=Ψ (D adopting ₁, D ₂... D _m) decide, generally when inferior user detects primary user, use be the arrival that signal level is distinguished primary user, therefore the data fusion mode Ψ adopting here the M a detecting result is averaged, and with primary user's signal threshold value D _thresholdcompare, determine that whether available channel is.Bunch head is used final result of determination to complete the next channel allocation stage (Channel Schedule Phase, CSP).In some open schemes that propose, there is more mechanism how to predict primary user's behavior, therefore suppose that primary user's arrival probabilistic model is known here.

Reasonably channel-aware and selection course should be combine with the dormancy awakening mechanism of node (Joint DCS and Sleep-Wake Strategy), reach and can choose best available channel, reduce again the object that energy consumes simultaneously as far as possible.What adopt here is the mode of a part of channel in a perception N channel, thereby reduces the quantity of channel-aware, but guarantees to choose best channel as far as possible.From N channel set, { chN}, choose a channel subset as can channel perception set (Polled-Channel Set) { S}, channel quantity in S || S||=K, (K<N), node is determined available channel by the channel in perception S.Whether a channel belongs to S is determined by the primary user's model on channel and channel weighted value, and specific implementation process is as follows:

If the N that node is safeguarded a channel set list is (ch _i, w _i), i=1,2 ... N, wherein w _ifor channel ch _icorresponding weighted value, the channel set list of Nodes has been to sort according to channel weighted value is descending.Suppose arbitrary channel ch _ion primary user arrive that to obey independent identically distributed parameter be λ _ipoisson distribution, in channel-aware part, the channel-aware result of carrying out at arbitrary channel-aware time slot obtains the probability that this channel can use and is

t wherein _plength for individual channel perception time slot.This probability is joined in the channel list that node safeguards, (ch _i, w _i, p _i), i=1,2 ... N, after the channel-aware process complete through, the probability that has obtained k channel and be available (Available Channel, AvaC) is

$P(\mathrm{AvaC}=k)=C_N^k\underset{\underset{\mathrm{lj}\&Element;\left\{\mathrm{chN}\right\}}{j=1}}{\overset{j=k}{\mathrm{\&Pi;}}}{p}_{l_j}\underset{\ln \&Element;\left\{\mathrm{chN}\right\}/\left\{l_j\right\}}{\mathrm{\&Pi;}}(1-p_{l_m})$

Making K is the desired value of available channel (Available Channel, AvaC) number, have,

$K=\underset{k=1}{\overset{k=N}{\mathrm{\&Sigma;}}}k\&CenterDot;P(\mathrm{AvaC}=k)=\underset{k=1}{\overset{k=N}{\mathrm{\&Sigma;}}}k\&CenterDot;C_N^k\underset{\underset{l_j\&Element;\left\{\mathrm{chN}\right\}}{j=1}}{\overset{j=k}{\mathrm{\&Pi;}}}{p}_{l_j}\underset{\mathrm{lm}\&Element;\left\{\mathrm{chN}\right\}/\left\{l_j\right\}}{\mathrm{\&Pi;}}(1-p_{l_m})$

With { S} replaces that { a rational prerequisite of chN} is, node is { available channel perceiving in chN} can only { S} just obtains by perception.Like this, than originally need to channel set, { chN} carries out the process of perception, can use to can channel perception set (Polled-Channel Set) that { S} carries out perception, the latter has reduced by N-K channel-aware time slot than the former, thereby reduced waking up period of node, reduced energy consumption.For guaranteeing that { S} has above-mentioned character, and { S} has specific character to needs.For try to achieve meet such condition { S} introduces following model here and carries out that { S} solves.

If the channel-aware strategy set that K channel forms is,

{S _K}={(ch ₁,p ₁,w ₁)，...(ch _i，p _i,w _i),...(ch _K,p _K，w _K)}

{ S _kby the channel in K stage, selected to form, making the strategy of the channel selection in i stage is (ch _i, x _i), ch wherein _ifor selected channel, x _iit is the impact on the current generation of the channel selection strategies in (i-1) individual stage.By node, successfully transmitting packet weighs and selects the energy of certain channel to consume.The energy of supposing the required consumption of each channel-aware time slot is E _p, node is the channel strategy (ch during stage at selected i _i, x _i) the needed energy of channel data be

, because node may successfully transmit data on this available channel, also may can not successfully transmit data due to the interference between primary user's unexpected arrival or all the other user, like this should be the function that on this channel, primary user arrives probability and channel weight,

.At { S _kdefinition in, { S _kin more perceived channel (supposing that its is simultaneously also perceived to available) can more easily by node use, be carried out transfer of data, i.e. { S _kin channel-aware sequentially reflected that node is used the priority level of channel.Like this, { S _kin the service condition of the selected channel of channel selection strategies in i stage be actually and be subject to that the channel selection strategies of its previous stage affects, this impact is designated as to x here _i.

Like this, the energy consumption in i stage can be defined as,

e _i(ch _i，x _i)=E _p+E _t(p _i，w _i)+E _e(x _i-1)

Like this, for the channel selection strategies { S in whole K stage _k, its energy consumption is

${\mathrm{<figure-callout\; id="1"\; label="E"\; filenames="HDA00002044841000031.png,HDA00002044841000032.png"\; state="\{\{state\}\}">E</figure-callout>}}_{1,K}\left[\right\{S_K\},x_1]=\underset{i=1}{\overset{K}{\mathrm{\&Sigma;}}}{e}_i({\mathrm{ch}}_i,x_i)$

E wherein _{1, K}[{ S _k, x ₁] in subscript represent the channel selection strategies from the channel selection strategies in the 1st stage to the K stage.For optimum { S _k, there is following formula to set up,

${\mathrm{<figure-callout\; id="1"\; label="E"\; filenames="HDA00002044841000031.png,HDA00002044841000032.png"\; state="\{\{state\}\}">E</figure-callout>}}_{1,K}\left[\right\{{S^{*}}_K\},{x^{*}}_1]=\underset{\&ForAll;\left\{S_K\right\}\&Element;\left\{\mathrm{chN}\right\}}{\min }{\mathrm{<figure-callout\; id="1"\; label="E"\; filenames="HDA00002044841000031.png,HDA00002044841000032.png"\; state="\{\{state\}\}">E</figure-callout>}}_{1,K}\left[\right\{S_K\},x_1]$

Optimum { S _kthe most directly to solve be enumerative technique, by enumerating one by one, altogether need to calculate

the inferior channel perception that can obtain K optimum.Obviously, this is a very large calculation of complex amount for the simple WSN node of hardware configuration, for reducing, solves optimum { S _kcomplexity, here to { S _kenergy consume definition and further analyze.

First, by following equation,

${\mathrm{<figure-callout\; id="1"\; label="E"\; filenames="HDA00002044841000031.png,HDA00002044841000032.png"\; state="\{\{state\}\}">E</figure-callout>}}_{1,K}\left[\right\{S_K\},x_1]=\underset{i=1}{\overset{K}{\mathrm{\&Sigma;}}}{e}_i({\mathrm{ch}}_i,x_i)=e_1({\mathrm{<figure-callout\; id="1"\; label="ch"\; filenames="HDA00002044841000031.png,HDA00002044841000032.png"\; state="\{\{state\}\}">ch</figure-callout>}}_1,x_1)+\underset{i=2}{\overset{K}{\mathrm{\&Sigma;}}}{e}_i({\mathrm{ch}}_i,x_i)$

$e=e_1({\mathrm{<figure-callout\; id="1"\; label="ch"\; filenames="HDA00002044841000031.png,HDA00002044841000032.png"\; state="\{\{state\}\}">ch</figure-callout>}}_1,x_1)+E_{2,K}\left[\right\{S_K\}/\{S_1\},x_2]$

Can obtain following formula,

${\mathrm{<figure-callout\; id="1"\; label="E"\; filenames="HDA00002044841000031.png,HDA00002044841000032.png"\; state="\{\{state\}\}">E</figure-callout>}}_{1,K}\left[\right\{{S^{*}}_K\},{x^{*}}_1]=\underset{\&ForAll;\left\{S_K\right\}\&Element;\left\{\mathrm{chN}\right\}}{\min }{\mathrm{<figure-callout\; id="1"\; label="E"\; filenames="HDA00002044841000031.png,HDA00002044841000032.png"\; state="\{\{state\}\}">E</figure-callout>}}_{1,K}\left[\right\{S_K\},x_1]$

$=\underset{\&ForAll;i\&Element;(1,K)}{\underset{\&ForAll;\left\{S_K\right\}\&Element;\left\{\mathrm{chN}\right\}}{\min }}\left\{{\mathrm{<figure-callout\; id="1"\; label="E"\; filenames="HDA00002044841000031.png,HDA00002044841000032.png"\; state="\{\{state\}\}">E</figure-callout>}}_{1,i}\right[\left\{S_i\right\},x_1]+E_{i+1,K}[\left\{S_K\right\}/\left\{S_i\right\},x_i\left]\right\}$

$=\underset{\&ForAll;i\&Element;(1,K)}{\underset{\&ForAll;\left\{S_i\right\}\&Element;\left\{\mathrm{chN}\right\}}{\min }}\left\{\underset{\&ForAll;\left\{S_K\right\}/\left\{S_i\right\}\&Element;\left\{\mathrm{chN}\right\}}{\min }\right\{{\mathrm{<figure-callout\; id="1"\; label="E"\; filenames="HDA00002044841000031.png,HDA00002044841000032.png"\; state="\{\{state\}\}">E</figure-callout>}}_{1,i}\left[\right\{S_i\},x_1]+E_{i+1,K}\left[\right\{S_K\}/\{S_i\},x_i]\left\}\right\}$

$=\underset{\&ForAll;i\&Element;(1,K)}{\underset{\&ForAll;\left\{S_i\right\}\&Element;\left\{\mathrm{chN}\right\}}{\min }}\left\{{\mathrm{<figure-callout\; id="1"\; label="E"\; filenames="HDA00002044841000031.png,HDA00002044841000032.png"\; state="\{\{state\}\}">E</figure-callout>}}_{1,i}\right[\left\{S_i\right\},x_1]+\underset{\&ForAll;\left\{S_K\right\}/\left\{S_i\right\}\&Element;\left\{\mathrm{chN}\right\}}{\min }{E}_{i+1,K}[\left\{S_K\right\}/\left\{S_i\right\},x_i\left]\right\}$

As can be seen from the above equation, for preferred channels selection strategy { S ^* _ksolve, in fact met the principle of optimality in Dynamic Programming (DP), i.e. " optimal policy of overall process has such character: no matter the state before certain state and decision-making are how in this optimal policy; for this state, remaining all decision-makings must form optimum substrategy ".Like this, for preferred channels selection strategy { S ^* _ksolve, can use Dynamic Programming (DP) algorithm to carry out Recursive Solution, compare like this method of exhaustion, can greatly save amount of calculation.

CSP and DTP stage: in the CSP stage, bunch head obtained final available channel list (Available Channel, AvaC) AvaC}, and bunch head by the processing of AvaC}, and to bunch in each node carry out channel allocation.Thereby all the other nodes in the CSP stage, to be bunch head notified channel assignment scheme and transfer of data time slot allocation scheme bunch, therefore the structure of time slot in CSP stage is exactly a communication time slot.For bunch head, be transmission side, all the other nodes are recipients.Data transfer phase (DTP) is that bunch interior nodes is carried out data communication at corresponding transfer of data time slot.The object of channel allocation is in order to complete channel handover mechanism, when node detects primary user in presently used channel, node should be switched on other standby channel immediately, and the transmission side and the recipient that carry out data communication will understand the channel after the other side is switched simultaneously mutually.In the CSP stage, bunch head also will for bunch in all the other nodes at data transfer phase (Data Transmission Phase, DTP), to distribute corresponding length be T _trtime slot.Concrete assigning process is as follows:

In bunch, (correspondence, bunch head has distributed k T for it in the DTP stage to k node _trtime slot) using available channel list { certain channel (ch of AvaC} _i, w _i, p _i) time (suppose that primary user on channel arrives that to obey parameter be λ _ipoisson distribution), this channel still available probability is

Obviously be that k node preferentially chosen a P _kilarger channel, its transmission success rate is higher, considers channel weighted value simultaneously, is that { channel in AvaC_k} is k node actual available channel list of distributing

wherein

for the weighted value after weighted average.Like this channel in AvaC_k} is actual be according to

value sort, wherein

to try to achieve P _kisituation under, use to P _kibe weighted acquisition.When in bunch, k node enters into the corresponding transfer of data time slot of DTP, it is according to { AvaC_k} carries out the selection of channel, when finding that collision appears in selected channel, transmission side and recipient can be according to { the next available channels of channel sequencing selection in AvaC_k}.

In another embodiment of the present invention, by C.Li, P.Wang, H.-H.Chen, document " A cluster based on-demand multi-channel mac protocol for wireless multimedia sensor networks " (Proc.IEEE Int.Conf.Communications ICC'08 with M.Guizani, 2008, pp.2371 – 2376.), in middle mentioned COM-MAC, add channel-aware process, be modified as the CR-COM agreement that can work under cognitive condition, with this CR-COM agreement scheme as a comparison.At the multichannel Clustering protocol CR-COM with basic, carry out on Performance Ratio basis, to different K values under KoN-MAC agreement can channel perception set { S _kperformance also compare.Wherein the primary user on different pieces of information channel can have different channel seizure ratios, and its arrival rate meets negative exponent and distributes, and adopts ON/OFF process simulation, and the primary user's occupancy on certain channel may have three kinds of situations, is respectively 0.4,0.5 and 0.6.Suppose to have 10 channels (comprise 1 control channel, and 9 data channels), the capacity of each channel is 2Mbps, and the time of carrying out a channel-aware is 0.4ms.In weighing the parameter of energy consumption, the energy consumption of node when receiving mode is 58.9mW, and the energy consumption when transmission mode is 46.5mW, and the energy consumption when channel-aware pattern is 58.9mW.Internodal maximum transmitted scope is 250m, and node produces and transmits a fixed rate data flow (CBR), and the size of packet is 210Byte, and the data pack buffer queue maximum of each Nodes is 50.If there is no specified otherwise, the primary user's model on data channel is chosen one at random from three kinds of channel seizure ratios (0.4/0.5/0.6), and each simulation run is the time of dry run 30s all.

Fig. 5 has compared can channel perception set { S in KoN-MAC agreement _knetwork throughput while getting different K values and under CR-COM agreement is along with the situation of change of data arrival rate.As shown in Figure 5, the throughput of KoN-MAC and CR-COM agreement all increases along with the increase of data arrival rate, and the state that finally reaches capacity, and reaches maximum.For arbitrary value (K=1,3,5,7,9) can channel perception set { S _k(K=1,3,5,7,9), the saturation throughput of KoN-MAC is all good than the performance of CR-COM, this is mainly that reason by two aspects causes: the channel-aware time of the one, KoN-MAC is determined by K value, remove while getting maximum (K=9) identical with the channel-aware time of CR-COM (being all the detecting period of whole 9 channels), channel-aware time during all the other K values is all few than CR-COM, makes a work period (Interval) of KoN-MAC shorter than CR-COM; Another reason only provides an available channel perception different from CR-COM exactly, KoN-MAC provides several available channel perceptions, while finding that a channel perception is unavailable in the transfer of data time slot of certain node in its distribution, it can be switched on standby channel perception, has improved the transmission success rate of data.For example, as KoN-MAC value K=1, KoN-MAC is now the same with CR-COM does not have standby channel perception, but the channel-aware time of KoN-MAC is 1/9 of CR-COM, the work period of KoN-MAC, (Interva) was about 3/4 of CR-COM, as shown in Figure 5, under saturation condition, the throughput of KoN-MAC is about 1.3 times of CR-COM.For KoN-MAC agreement, when the value of K makes node have standby channel perception (being K>1), along with reducing of K value, the saturation throughput of network increases thereupon, this is because guaranteed to have comprised as much as possible in a perceived K channel available channel in current all channels in the channel-aware of KoN-MAC and selection strategy, { AvaC_k} is the best available channel in current all channels as far as possible to make available channel list that certain node obtains, thereby make certain node can successfully transmit data in the transfer of data time slot of its distribution.As shown in Figure 5, when K=3, the saturation throughput of KoN-MAC is maximum, this is because channel-aware and the selection strategy of KoN-MAC just can guarantee that when K=3 node can obtain an available channel list { AvaC_k} as well as possible, its channel-aware time (3 channel-aware time) is than all the other K values (5 simultaneously, 7,9) to lack.

Fig. 6 has compared can channel perception set { S in KoN-MAC agreement _knetwork packet loss rate while getting different K values and under CR-COM agreement is along with the situation of change of data arrival rate.As shown in Figure 6, the variation tendency of packet loss is identical with the variation tendency of network throughput, for arbitrary value (K=1,3,5,7,9) can channel perception set { S _k(K=1,3,5,7,9), the packet loss of KoN-MAC is all good than the performance of CR-COM, and main cause is identical with the reason of variation that causes throughput.The packet loss of CR-COM just starts sharply to increase when data arrival rate is greater than 80kps, and when data arrival rate is greater than 170kps, just starts to increase for (K=3) in the best situation of KoN-MAC.In KoN-MAC, different K values is similar on the impact of throughput on the impact of packet loss and K value, and while removing K=1 (there is no standby channel perception), along with reducing of K value, the packet loss of KoN-MAC decreases.

Fig. 7 has compared can channel perception set { S in KoN-MAC agreement _kaverage delay while getting different K values and under CR-COM agreement is along with the situation of change of data arrival rate.The measured time delay of this emulation refers to from the transport layer transmission data of sending node and starts to receive to the transport layer of receiving node the time span that packet is used.As shown in Figure 7, the variation tendency of average delay is identical with the variation tendency of packet loss, for arbitrary value (K=1,3,5,7,9) can channel perception set { S _k(K=1,3,5,7,9), the average delay of KoN-MAC is all good than the performance of CR-COM, and main cause remains that the size of determined work period of channel-aware time (Interval) and probability size that node successfully transmits data in its transfer of data time slot cause.The average delay of CR-COM just starts sharply to increase when data arrival rate is greater than 80kps, and be finally stabilized in about about 0.93s, and when data arrival rate is greater than 170kps, just start to increase for (K=3) in the best situation of KoN-MAC, and be finally stabilized in 0.45s left and right.In KoN-MAC, different K values is similar on the impact of packet loss on the impact of average delay and K value, and while removing K=1 (there is no standby channel perception), along with reducing of K value, the average delay of KoN-MAC decreases.For KoN-MAC and CR-COM, it is stable that the average delay of the two all can keep when data arrival rate increases to a certain degree, this is because the data pack buffer queue of each node is to have peaked (being made as 50 in this emulation), when data arrival rate is enough large, the data pack buffer queue meeting of node reaches capacity, mean time is postponed a meeting or conference the same with throughput, be finally stabilized near certain fixed value.

Fig. 8 has compared can channel perception set { S in KoN-MAC agreement _kthe average energy consumption of a packet of successful transmission while getting different K values and under CR-COM agreement is along with the situation of change of data arrival rate.As shown in Figure 8, when data arrival rate less (being less than 40kps), for getting compared with for the KoN-MAC agreement of little K value, its average energy consumption of successfully transmitting a packet is more much smaller than CR-COM.When data arrival rate increases gradually, the average energy consumption of a packet of successful transmission of KoN-MAC and CR-COM starts to be tending towards a stationary value, for the difference between them of clear view more, the part that data arrival rate in Fig. 8 is greater than to 40kps is amplified, as shown in Figure 9.As shown in Figure 9, it is maximum that CR-COM agreement is successfully transmitted the average energy consumption of a packet, and the less KoN-MAC agreement of K value, and its average energy consumption of successfully transmitting a packet is less, and average energy consumption is minimum during K=1, and its least energy consumption is about 0.065mW.When KoN-MAC agreement value K=3 (now, network throughput, packet loss and average delay performance are best), the minimum average B configuration energy consumption of successfully transmitting a packet is 0.068mW left and right, while being about K=1 1.04 times.The average energy consumption that KoN-MAC agreement is successfully transmitted a packet occurs that the reason of this variation tendency is, the channel-aware of KoN-MAC and selection strategy have limited the extra energy consumption expense that channel-aware process produces, the saving (particularly when lower data arrival rate) that has brought significant energy to consume.

The result of complex chart 5, Fig. 6, Fig. 7, Fig. 8 and Fig. 9, the combination property of the KoN-MAC agreement when K=3 is best, and its saturation network throughput is 1.03 times of inferior good KoN-MAC agreement (K=5), is 1.99 times of CR-COM; Its maximum average delay is 0.938 times of inferior good KoN-MAC agreement (K=5), is 0.489 times of CR-COM; Its minimum average B configuration energy consumption of successfully transmitting a packet is 1.04 times of best KoN-MAC agreement (K=1), is 0.648 times of CR-COM.

Claims (9)

1. the communication means of the multi-hop multichannel cognition wireless sensing network based on sub-clustering, it is characterized in that, by the frame structure design of node, be segmental structure, a work period of node is divided into channel-aware and choice phase, channel allocation stage, data transfer phase and dormant stage.

2. the communication means of the multi-hop multichannel cognition wireless sensing network based on sub-clustering according to claim 1, it is characterized in that, each node has the channel weighted list of oneself, and the channel weight that channel weighted list is recorded reflects that node is for the behaviour in service of channel; The variation of channel weighted value is to occur while the different conditions of channel being detected by node, and with respect to node, channel has several states as follows:

-the free time: this state appears at the channel-aware stage, time user awareness can be used to channel;

-busy: this state appears at the channel-aware stage equally, and time user awareness is to the existence of primary user on channel, and channel is unavailable;

-communication: this state appears at time user and can use perceiving channel, and has normally used this channel to carry out transfer of data;

-collision: this state can occur in both cases, the first situation is that time user node in its adjacent cluster when data transfer phase is also being used this channel, the second situation is that time user primary user when transfer of data arrives suddenly, and both of these case all causes current channel unavailable.

3. the communication means of the multi-hop multichannel cognition wireless sensing network based on sub-clustering according to claim 2, is characterized in that, according to channel in different conditions, the concrete variation of the weighted value w of channel is not identical yet, when channel idle, weighted value increases W _idle; When channel is busy, weighted value reduces W _busy; When channel communication, weighted value increases W _com; When channel collision, corresponding the first and the second situation, weighted value reduces respectively W _secand W _pri.

4. the communication means of the multi-hop multichannel cognition wireless sensing network based on sub-clustering according to claim 2, is characterized in that, completes the process of channel-aware and selection in channel-aware and choice phase, specifically comprises following steps:

Step (101): in the structure of time slot of channel-aware and choice phase, previous communication time slot in to be bunch head send to the own channel weighted list of safeguarding bunch all the other members to upgrade their channel weighted list, be now bunch head as transmit leg, a non-bunch of member is as recipient; A rear communication time slot is to prepare for channel associating perception, bunch in member own detected channel-aware result is all transferred to a bunch head, at bunch head, being in final channel-aware result carries out Combined Treatment;

Step (102): a part of channel in a perception N channel, particularly, from N channel set, { chN}, choose a channel subset as for can channel perception set { S}, channel quantity in S is || S||=K, (K<N), node is determined available channel by the channel in perception S.

5. the communication means of the multi-hop multichannel cognition wireless sensing network based on sub-clustering according to claim 4, it is characterized in that, channel associating perception is the mode that adopts data fusion, particularly, one have M member bunch in, what after channel-aware that each bunch of interior nodes carries out, obtain judgment result is that D _i, final result of determination is by the data fusion mode D=Ψ (D adopting ₁, D ₂... D _m) decide.

6. the communication means of the multi-hop multichannel cognition wireless sensing network based on sub-clustering according to claim 5, is characterized in that, data fusion mode Ψ averages the M a detecting result, and with primary user's signal threshold value D _thresholdcompare, determine that whether available channel is; Bunch head is used final result of determination to complete the next channel allocation stage

7. the communication means of the multi-hop multichannel cognition wireless sensing network based on sub-clustering according to claim 2, it is characterized in that, in the channel allocation stage, bunch head has obtained final available channel list { AvaC}, bunch head is by { the processing of AvaC}, to bunch in each node carry out channel allocation, particularly, bunch head channel assignment scheme and transfer of data time slot allocation scheme are notified bunch in all the other nodes, wherein, the structure of time slot in CSP stage is exactly a communication time slot, and a bunch head is transmission side, and all the other nodes are recipients.

8. the communication means of the multi-hop multichannel cognition wireless sensing network based on sub-clustering according to claim 2, is characterized in that, in the channel allocation stage, bunch head also will for bunch in all the other nodes at data transfer phase, to distribute corresponding length be T _trtime slot.

9. the communication means of the multi-hop multichannel cognition wireless sensing network based on sub-clustering according to claim 1, it is characterized in that, at data transfer phase, bunch interior nodes is carried out data communication at corresponding transfer of data time slot, when node detects primary user in presently used channel, node is switched on other standby channel immediately, and wherein, the transmission side and the recipient that carry out data communication will understand the channel after the other side is switched mutually.

Images