CN104735721A - Method for transmitting image signal based on compressed sensing technology in multi-hop wireless network - Google Patents

Abstract

The invention provides a method for transmitting an image signal based on the compressed sensing technology in a multi-hop wireless network. The method includes the steps that (1) original signals are packaged into a plurality of description packages with the equivalent information amount and the same significance through an encoder in a compressed sensing mode; (2) transmission benefits of each transmission link are calculated; (3) link connection needing to be built is determined according to the transmission benefits of the transmission links; (4) the transmission links are built according to the scheduling scheme, and transmission of the description packages is started. According to the method, encoding, decoding and route scheduling in image signal transmission are comprehensively considered, the high implementability is achieved, the method can be widely applied to existing wireless network transmission, the network transmission capacity is further improved, and the image signal transmission instantaneity and the image signal transmission effect are effectively improved.

Description

Based on the method for compressed sensing technical transmission picture signal in multi-hop wireless network

Technical field

What the present invention relates to is a kind of method of wireless communication technology field, especially the coded image data packing for the purpose of network output optimum, an algorithm for packet routing scheduling, be specifically related to the method based on compressed sensing technical transmission picture signal in multi-hop wireless network.

Background technology

Along with the development of radio network technique, the universal of mobile device, the demand of the mobile data service that people are right is growing.Thus the mobile data transfer of large data, real-time more and more receives publicity.This topic not only obtains academicly and studies widely, is also quite paid attention in the application of industrial application.Effectively utilizing existing resource for user provides real-time is better, capacity is larger mobile data service to be an important topic urgently to be resolved hurrily.

On the one hand, along with the increase of data volume, the increasing the weight of of network load, the routing scheduling problem of packet will become and become increasingly complex.For letter signal, the data volume of picture signal (as picture, video etc.) is larger, requirement of real-time is higher, and picture signal promotes the indispensable part of Consumer's Experience exactly, will day by day increase the transmission demand of image signal transmission.On the other hand, invention before only considered the problem of information source end coding, receiving terminal decoding mostly, or only considered packet routing scheduling problem in a network.A set of coding and decoding and the mutual adaptive algorithm of routing scheduling, the transmission performance of optimized image signal in coding and transmission, can improve Internet Transmission effect, improve network capacity.

Summary of the invention

For defect of the prior art, the object of this invention is to provide the problem that a kind of method solves the image signal transmission in multi-hop wireless network.The present invention has considered coding and decoding in image signal transmission and routing scheduling problem, and has very strong realizability, can be widely deployed in wireless network transmissions till now.The present invention can help us to improve network capacity, effectively improves image signal transmission real-time and laser propagation effect.

According to provided by the invention in multi-hop wireless network the method based on compressed sensing technical transmission picture signal, comprise the steps:

The first step, encoder utilizes compressed sensing mode original signal to be packaged into the description bag that several amount of information are suitable, importance is identical;

Second step, calculates the transmission benefit of every bar transmission link;

According to the transmission benefit of transmission link, 3rd step, determines that the link needing to set up connects;

4th step, sets up transmission link by scheduling scheme and starts transmission description bag.

Preferably, described step 1, is specially:

First, encoder carries out M (M < N) secondary sampling to the original image signal F that size is N' × N', obtains sampled value y _m, wherein, N=N' × N', F=[F _i,j] (i, j≤N'), F _i,jfor the value in original image corresponding points; The essence of sampling measures original image signal F=[F _i,j] at sampling matrix φ ' _m(m=1,2 ..., M) on projection

y _m＝F,φ' _m

Wherein, m=1,2 ..., M illustrates sampling number, original image signal F=[F _i,j] and sampling matrix size be all N' × N';

By two dimensional image signal F=[F _i,j] by arranging building up a dimensional vector x=[x ₁... x _n], (n=1,2 ..., N), wherein N=N' × N'; And by two-dimentional sampling matrix φ ' _mone dimension row vector φ is stacked as by row transposition _m, then have

y _m＝F,φ' _m＝φ _mx

；

By y _mpile the matrix y that size is M × 1, namely $y=\left[\begin{array}{c}{y}_1\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ y_M\end{array}\right];$ By sampling matrix φ _minversion φ _m ^tpiling size is by row M × N matrix Φ, namely $\mathrm{\&Phi;}=\left[\begin{array}{c}{\mathrm{\&phi;}}_1\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ {\mathrm{\&phi;}}_M\end{array}\right],$ Then have

y＝Φx

Wherein, sampling matrix Φ is that obedience ± 1 grade is generally with the random matrix of distribution (i.i.d.);

Then, encoder is by sampled value y _mand the sampling matrix φ ' of correspondence _minformation package become one describe bag;

The information of sampling matrix Φ is that transmitting terminal and receiving terminal are shared, and namely sampling matrix Φ is pre-stored in the decoder of transmitting terminal and the decoder of receiving terminal, and encoder only needs when packing describes bag to point out sampling matrix φ ' in description bag _mcorresponding sampling number m.

Preferably, described step 2, is specially:

(1) the transmission demand factor of link is calculated

If a certain moment t, for comprising sending node N _twith receiving node N _rlink l, sending node N _tdescription bag set in buffer memory is receiving node N _rdescription bag set in buffer memory is suppose that each node can only forward at most the same description bag received once; So, will two kinds of different description bags be had in nodal cache: a kind of is that node receives and forwarded over description bag, be called ash bag; Another kind is that node receives but not yet forwarded over description bag, is called white bag; Note sending node N _tthe set of the Hui Bao in buffer memory, white bag is respectively then have

$U_{N_t}=U_{N_t}^G\&cup;U_{N_t}^W$

If set up link l, then can be that those belong to set for the packet of transmission and do not belong to set the set of description bag; Link l can for the set U of the packet of transmission _pfor

$U_P=\left\{d\right|d\&Element;U_{N_t}^W,d\&NotElement;U_{N_r}\}$

Consider link transmission capacity, then the packet number p that each transmission time slot link l can transmit is

p=min(|U _P|,q(l))

Wherein, q (l)=q (N _t, N _r) represent the description bag number that a transmission time slot link l can transmit, | U _p| represent set U _pin the number of description bag that comprises;

If set up link l, at the reduction Δ RMS of the root-mean-square error of receiving node place restored image be

$\mathrm{\&Delta;PMS}=f_{\mathrm{RMS}}\left(\right|U_{N_r}\left|\right)-f_{\mathrm{RMS}}\left(\right|U_{N_r}|+p)$

Wherein, function f _rMS() is the root-mean-square error RMS of compressed sensing restored image and the relation of the sampled value number for restoring;

Transmission demand factor TDF (l) defining this moment t link l is

$\mathrm{TDF}\left(l\right)=\mathrm{\&Delta;RMS}=f_{\mathrm{RMS}}\left(\right|U_{N_r}\left|\right)-f_{\mathrm{RMS}}\left({|U}_{N_r}\right|+p)$

The transmission demand factor of link has weighed the importance that link is about to the packet of transmission;

(2) the transmission supply factor of link is calculated;

(3) transmission income, the transmission cost of link is calculated

For link transmission requirements factor R evenue (l) and transmission are supplied, factor TSF (l) is to be amassed in transmission income Revenue (l) of definition link l;

Revenue(l)＝TDF(l)·TSF(l)

The transmission income of the interference set link that transmission cost Cost (l) defining link l is l, namely

$\mathrm{Cost}\left(l\right)=\underset{l_i\&Element;S_{\mathrm{int}}\left(l\right)}{\mathrm{\&Sigma;}}\mathrm{Revenue}\left(l_i\right)$

Wherein, the interference S set of link l _int(l) for the interference brought because of link l and the link set that cannot normally work, namely

S _int(l)＝{l _int|l _int≠l,Dis(N _t(l _int),N _r(l))＜r,Dis(N _r(l _int),N _t(l))＜r}

Wherein, N _tl () is the sending node of link l, N _rl () is the receiving node of link l, Dis (N _a, N _b) be two node N _a, N _bbetween geographic distance or channel quality, r is the interference radius of node or the threshold value of channel quality, l _intrepresent the link that can cause interference for link l;

(4) the transmission benefit of link is obtained

Transmission benefit Utility (l) of link l is defined as

$\mathrm{Utility}\left(l\right)=\frac{\mathrm{Revenue}\left(l\right)}{\mathrm{Cost}\left(l\right)}.$

Preferably, the transmission supply factor of described calculating link, is specially:

First calculate the node weights of each via node, be specially:

A the joint behavior value Q of () each via node is initialized as 0, node weight weight values W is initialized as 0, and node is initialized as just infinite to the jumping figure H of receiving terminal node; The Q of receiving terminal node is just infinite, and W is just infinite, and H is 0;

B () travels through all node N except receiving terminal node successively _iif have and node N _ithe node N be directly connected _jmeet

$W\left(N_i\right)<\frac{\min (q(N_i,N_j),Q\left(N_j\right))}{(1+H\left(N_j\right))}$

Then

$W\left(N_i\right)\&LeftArrow;\frac{\min (q(N_i,N_j),Q\left(N_j\right))}{(1+H\left(N_j\right))}$

Q(N _i)←min(q(N _i,N _j),Q(N _j))

H(N _i)←1+H(N _j)

Namely N is worked as _iwith N _jfor transmission down hop, and node N _iwhen obtaining more excellent node weight weight values, then more new node N _ijoint behavior value Q, node weight weight values W and node to the jumping figure H of receiving terminal node; Wherein, q (l)=q (N _i, N _j) represent with N _ifor sending node, N _jfor the description bag number that the link l of receiving node can transmit in a transmission time slot, W (N _i) represent node N _iweighted value, Q (N _i) represent node N _iperformance number, H (N _i) represent node N _ito the jumping figure of receiving terminal node;

C () repeats (b) until joint behavior value Q, the node weight weight values W of all nodes and node no longer change in once traversal to the jumping figure H of receiving terminal node; Obtain the node weight weight values W of each node;

Node weights has weighed the ability of each via node to receiving terminal transmission data;

Be N for receiving node _rlink l, link l transmit supply factor TSF (l) be defined as

TSF(l)=W(N _r)

Wherein, W (N _r) represent node N _rweighted value.

Preferably, described 3rd step, is specially:

(1), when starting, transmission benefit is that positive link is all divided into the link that may be established, and joins the set U of the link that will be established _estLinkin;

(2) the link l that transmission benefit is maximum is found out _max, then from U _estLinkremoving is to link l _maxproduce other links S of interference _int(l _max); And by link l _maxtransmission benefit zero, i.e. Utility (l _max)=0;

(3) (2) are repeated until U _estLinkin link do not interfere with each other, namely then U _estLinkfor the link set that this timeslot scheduling scheme will be set up.

Preferably, in described 4th step, receiving terminal is decoded to paid-in description bag according to actual conditions.

Preferably, in described 4th step, if at time t, it is { k that receiving terminal receives K corresponding sampling number ₁, k ₂..., the description bag of K}, wherein if describe the sampled value comprised in bag be $\{y_{k_1},y_{k_2},...,y_K\},$ Be stacked as by sampling number $y_r={\left[\begin{array}{cccc}{y}_{k_1}& y_{k_2}& \&CenterDot;\&CenterDot;\&CenterDot;\&CenterDot;& y_K\end{array}\right]}^T,$ Corresponding sampling matrix is ${\mathrm{\&Phi;}}_r={\left[\begin{array}{cccc}{\mathrm{\&phi;}}_{k_1}^T& {\mathrm{\&phi;}}_{k_2}^T& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{\&phi;}}_K^T\end{array}\right]}^T;$ Restore original signal, i.e. solving-optimizing problem

min TV(F)

s.t.y _r＝Φ _rV _F

Wherein, column vector V _fbe that two dimensional image signal F is stacking by row, TV (F) represents V _fthe total variance of corresponding two-dimentional original image F.

Compared with prior art, the present invention has following beneficial effect:

(1) the present invention utilizes the recovery technique of compressed sensing, and receiving terminal can be decoded according to paid-in description bag at any time, and without the need to waiting for that some disappearance describes bag or request transmitting terminal retransmits, the time delay that minimizing is waited for or retransmission operation is brought.

(2) picture signal of the present invention to real-time, the large data of transmission has very strong realizability and specific aim.

(3), in routing algorithm of the present invention, the link that transmittability is high can preferential busy channel.Can, under the prerequisite avoiding channel confliction, whole network be made to have higher output like this.

(4) the present invention has considered coding and decoding in image signal transmission and routing scheduling problem.In this transmission method, the coding and decoding of picture signal and the routing scheduling of packet adaptive mutually, coding and transmission work in coordination with the transmission performance of optimized image signal, can improve better Internet Transmission effect, improve network capacity.

(5) the present invention can ensure good real-time property, and efficient routing scheduling, and when link establishment is transmitted, does not have channel confliction, and the network simultaneously obtaining relative good exports performance.

Accompanying drawing explanation

By reading the detailed description done non-limiting example with reference to the following drawings, other features, objects and advantages of the present invention will become more obvious:

Fig. 1 is the contrast effect schematic diagram of restored map root-mean-square error (RMS) the time dependent curve of network scenarios 1 and network scenarios 2.

Fig. 2 is the contrast effect schematic diagram of the time dependent curve of data packet number in the receiving terminal node buffer memory of network scenarios 1 and network scenarios 2.

Embodiment

Below in conjunction with specific embodiment, the present invention is described in detail.Following examples will contribute to those skilled in the art and understand the present invention further, but not limit the present invention in any form.It should be pointed out that to those skilled in the art, without departing from the inventive concept of the premise, some distortion and improvement can also be made.These all belong to protection scope of the present invention.

The invention provides a kind of in multi-hop wireless network the method based on compressed sensing technical transmission picture signal, transmission method of the present invention include match each other data packing, transfer of data, data recovery link, the problem of picture signal transmitting real-time, large data in the wireless network can be completed efficiently.Step of the present invention is as follows: first transmitting terminal is packed to view data according to compressed sensing, then each node first calculates the node weights of oneself, then according to the node weights of receiving node in the packet state in both link ends nodal cache and link, calculate the transmission benefit of link, and in units of link, carry out routing scheduling according to link transmission benefit.

The network system model of discussion of the present invention is a unicast networks having a transmitting terminal, a corresponding receiving terminal and multiple relay transmission node.Because transmitting terminal and receiving terminal are outside communication range, therefore forward from the signal demand of transmitting terminal could arrive receiving terminal by via node.

The present invention includes following steps:

The first step, encoder is responsible for utilizing compressed sensing technology original signal to be packaged into the description bag that several amount of information are suitable, importance is identical.

First, the original image signal F that encoder will be N' × N' to size carries out M (M < N) secondary sampling, obtains sampled value y _m, wherein, N=N' × N', F=[F _i,j] (i, j≤N'), F _i,jfor the value in original image corresponding points.The essence of sampling measures original image signal F=[F _i,j] at sampling matrix φ ' _m(m=1,2 ..., M) on projection

y _m＝F,φ' _m

Wherein, m=1,2 ..., M illustrates sampling number, original image signal F=[F _i,j] and sampling matrix size be all N' × N'.

By two dimensional image signal F=[F _i,j] by arranging building up a dimensional vector x=[x ₁... x _n], (n=1,2 ..., N), wherein N=N' × N'.And by two-dimentional sampling matrix φ ' _mone dimension row vector φ is stacked as by row transposition _m, then have

y _m＝F,φ' _m＝φ _mx。

By y _mpile the matrix y that size is M × 1, namely $y=\left[\begin{array}{c}{y}_1\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ y_M\end{array}\right];$ By sampling matrix φ _minversion φ _m ^tpiling size is by row M × N matrix Φ, namely $\mathrm{\&Phi;}=\left[\begin{array}{c}{\mathrm{\&phi;}}_1\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ {\mathrm{\&phi;}}_M\end{array}\right],$ Then have

y＝Φx

Wherein, sampling matrix Φ is that obedience ± 1 grade is generally with the random matrix of distribution (i.i.d.).

Then, encoder is by sampled value y _mand the sampling matrix φ ' of correspondence _minformation package become one describe bag.

The information of sampling matrix Φ is that transmitting terminal and receiving terminal are shared, and namely sampling matrix Φ is pre-stored in the decoder of transmitting terminal and the decoder of receiving terminal.Therefore, encoder, when packing describes bag, describes in bag without the need to being joined by the particular content of sampling matrix, only needs to point out sampling matrix φ ' in description bag _mcorresponding sampling number m.Therefore, describing bag is exactly the packet comprising sampled value and sampling number.

Second step, to every bar transmission link, calculates its transmission benefit.Algorithm concrete steps are as follows:

(1) the transmission demand factor of link is calculated.

If a certain moment t, for comprising sending node N _twith receiving node N _rlink l, sending node N _tdescription bag set in buffer memory is receiving node N _rdescription bag set in buffer memory is suppose that each node can only forward at most the same description bag received once.So, will two kinds of different description bags be had in nodal cache: a kind of is that node receives and forwarded over description bag, be called ash bag; Another kind is that node receives but not yet forwarded over description bag, is called white bag.Note sending node N _tthe set of the Hui Bao in buffer memory, white bag is respectively then have

$U_{N_t}=U_{N_t}^G\&cup;U_{N_t}^W$

If set up link l, should be able to be not present in receiving node buffer memory for the packet of transmission, the white bag in sending node; Namely those belong to set and do not belong to set the set of description bag.Link l can for the set U of the packet of transmission _pfor

Consider link transmission capacity, then the packet number p that each transmission time slot link l can transmit is

p=min(|U _P|,q(l))

Wherein, q (l)=q (N _t, N _r) represent the description bag number that a transmission time slot link l can transmit, | U _p| represent set U _pin the number of description bag that comprises.

If set up link l, at the reduction Δ RMS of the root-mean-square error of receiving node place restored image be

$\mathrm{\&Delta;PMS}=f_{\mathrm{RMS}}\left(\right|U_{N_r}\left|\right)-f_{\mathrm{RMS}}\left(\right|U_{N_r}|+p)$

Wherein, function f _rMS() is the root-mean-square error RMS of compressed sensing restored image and the relation of the sampled value number for restoring.

Transmission demand factor TDF (l) defining this moment t link l is

$\mathrm{TDF}\left(l\right)=\mathrm{\&Delta;RMS}=f_{\mathrm{RMS}}\left(\right|U_{N_r}\left|\right)-f_{\mathrm{RMS}}\left({|U}_{N_r}\right|+p)$

The transmission demand factor of link has weighed the importance that link is about to the packet of transmission.

(2) the transmission supply factor of link is calculated.

First calculate the node weights of each via node, computational algorithm is

A the joint behavior value Q of () each via node is initialized as 0, node weight weight values W is initialized as 0, and node is initialized as just infinite to the jumping figure H of receiving terminal node.The Q of receiving terminal node is just infinite, and W is just infinite, and H is 0.

B () travels through all node N except receiving terminal node successively _iif have and node N _ithe node N be directly connected _jmeet

$W\left(N_i\right)<\frac{\min (q(N_i,N_j),Q\left(N_j\right))}{(1+H\left(N_j\right))}$

Then

$W\left(N_i\right)\&LeftArrow;\frac{\min (q(N_i,N_j),Q\left(N_j\right))}{(1+H\left(N_j\right))}$

Q(N _i)←min(q(N _i,N _j),Q(N _j))

H(N _i)←1+H(N _j)

Namely N is worked as _iwith N _jfor transmission down hop, and node N _iwhen obtaining more excellent node weight weight values, then more new node N _ijoint behavior value Q, node weight weight values W and node to the jumping figure H of receiving terminal node.Wherein, q (l)=q (N _i, N _j) represent with N _ifor sending node, N _jfor the description bag number that the link l of receiving node can transmit in a transmission time slot, W (N _i) represent node N _iweighted value, Q (N _i) represent node N _iperformance number, H (N _i) represent node N _ito the jumping figure of receiving terminal node.

C () repeats (b) until joint behavior value Q, the node weight weight values W of all nodes and node no longer change in once traversal to the jumping figure H of receiving terminal node.Obtain the node weight weight values W of each node.

Node weights has weighed the ability of each via node to receiving terminal transmission data.

Be N for receiving node _rlink l, its transmission supply factor TSF (l) be defined as

TSF(l)=W(N _r)

Wherein, W (N _r) represent node N _rweighted value.

(3) transmission income, the transmission cost of link is calculated.

For link transmission requirements factor R evenue (l) and transmission are supplied, factor TSF (l) is to be amassed in transmission income Revenue (l) of definition link l.

Revenue(l)＝TDF(l)·TSF(l)

The transmission income of the interference set link that transmission cost Cost (l) defining link l is l, namely

$\mathrm{Cost}\left(l\right)=\underset{l_i\&Element;S_{\mathrm{int}}\left(l\right)}{\mathrm{\&Sigma;}}\mathrm{Revenue}\left(l_i\right)$

The interference S set of its link l _int(l) for the interference brought because of link l and the link set that cannot normally work, namely

S _int(l)＝{l _int|l _int≠l,Dis(N _t(l _int),N _r(l))＜r,Dis(N _r(l _int),N _t(l))＜r}

Wherein, N _tl () is the sending node of link l, N _rl () is the receiving node of link l, Dis (N _a, N _b) be two node N _a, N _bbetween geographic distance or channel quality, r is the interference radius of node or the threshold value of channel quality, l _intrepresent the link that can cause interference for link l.

(4) the transmission benefit of link is obtained.

Transmission benefit Utility (l) of link l is defined as

$\mathrm{Utility}\left(l\right)=\frac{\mathrm{Revenue}\left(l\right)}{\mathrm{Cost}\left(l\right)}.$

According to the transmission benefit of link, 3rd step, determines that the link needing to set up connects.Concrete algorithm is as follows:

(1) set of the link that will be established is U _estLink.During beginning, transmission benefit is that positive link is all divided

For the link that may be established, join set U _estLinkin.

(2) the link l that transmission benefit is maximum is found out _max, then from U _estLinkremoving is to link l _maxproduce other links S of interference _int(l _max).And by link l _maxtransmission benefit zero, i.e. Utility (l _max)=0.Due to l _maxthe link interfered with each other is removed in this stage, therefore l _maxbe not affected in follow-up iteration.

(3) (2) are repeated until U _estLinkin link do not interfere with each other, namely then U _estLinkfor the link set that this timeslot scheduling scheme will be set up.

After above algorithm travels through all links, masked institute noisy may, system reaches one can the state of stable transfer.This stable transfer state refers to, in each interference range, under the prerequisite not producing interference, the link that channel is all transmitted value of utility local optimum takies.

4th step, set up transmission link by scheduling scheme and start transmission description bag, receiving terminal can be decoded to paid-in description bag according to actual conditions.

If at time t, it is { k that receiving terminal receives K corresponding sampling number ₁, k ₂..., the description bag of K}, wherein $\{k_1,k_2,...,K\}\&Subset;\{\mathrm{1,2,3},...,M\}.$ If describe the sampled value comprised in bag be $\{y_{k_1},y_{k_2},...,y_K\},$ Be stacked as by sampling number $y_r={\left[\begin{array}{cccc}{y}_{k_1}& y_{k_2}& \&CenterDot;\&CenterDot;\&CenterDot;& y_K\end{array}\right]}^T,$ Corresponding sampling matrix is ${\mathrm{\&Phi;}}_r={\left[\begin{array}{cccc}{\mathrm{\&phi;}}_{k_1}^T& {\mathrm{\&phi;}}_{k_2}^T& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{\&phi;}}_K^T\end{array}\right]}^T,$ According to compressive sensing theory, restore original signal, i.e. solving-optimizing problem by solving TV Minimization method

min TV(F)

s.t.y _r＝Φ _rV _F

Wherein, column vector V _fthat two dimensional image signal F is stacking by row.

In an experiment, wireless network is distributed in the square area of a 80*80, and the mid point on the both sides that square is relative is a sender node and receiving terminal node respectively; In addition, 18 via nodes are randomly dispersed in square area.The interference radius of radio communication and communication radius are all r=30 rice, the effect of the wireless signal that two nodes namely in 30 meters of radius can be sent each other.Between two nodes in communication range, the signal to noise ratio of communication link meets Gaussian Profile.We know simultaneously, and in the interference radius of transmission node, other nodes except receiving node can not receive data; In the interference radius of receiving node, other nodes except transmission node can not send data.And arbitrary node can only receive a node to its data simultaneously, and can not receive simultaneously and send data.According to these relations, we can set up the interference relationships matrix between this network scenarios link.

Specific implementation step comprises following steps:

The first step, the node weight weight values W of each via node is calculated according to network topology, the node weight weight values W of receiving terminal node is set to this 100 special numerical requirements of 100(, just needs the node weight weight values of receiving terminal node enough large relative to the node weight weight values of other nodes).

Second step, the algorithm introduced according to us, calculates the transmission demand factor of each link, the transmission income of link, transmission cost and transmission benefit.

According to the transmission benefit of link, 3rd step, determines that the link needing to set up connects, sets up communication link, upgrade the packet state in each nodal cache.

4th step, records following data:

(1) data packet number that receives at each time slot of receiving terminal node.

(2) in each time slot, receiving terminal node carries out the mean square deviation of restored map relative to former figure of signal restoring according to the packet received.

In two different network scenarios, by recording above data, compare the result that contrast experiment obtains, we obtain Fig. 1, Fig. 2.Wherein, contrast experiment's content is in same wireless network, utilize the transmission method based on conventional discrete cosine transform to transmit same picture signal.

In Fig. 1, (a) is network scenarios 1 recovery effect, and (b) is network scenarios 2 recovery effect.In Fig. 2, (a) is network scenarios 1 recovery effect, and (b) is network scenarios 2 recovery effect.

Above specific embodiments of the invention are described.It is to be appreciated that the present invention is not limited to above-mentioned particular implementation, those skilled in the art can make various distortion or amendment within the scope of the claims, and this does not affect flesh and blood of the present invention.

Claims (7)

1. in multi-hop wireless network based on a method for compressed sensing technical transmission picture signal, it is characterized in that, comprise the steps:

The first step, encoder utilizes compressed sensing mode original signal to be packaged into the description bag that several amount of information are suitable, importance is identical;

Second step, calculates the transmission benefit of every bar transmission link;

According to the transmission benefit of transmission link, 3rd step, determines that the link needing to set up connects;

4th step, sets up transmission link by scheduling scheme and starts transmission description bag.

2. according to claim 1 in multi-hop wireless network based on the method for compressed sensing technical transmission picture signal, it is characterized in that, described step 1, is specially:

First, encoder carries out M (M < N) secondary sampling to the original image signal F that size is N' × N', obtains sampled value y _m, wherein, N=N' × N', F=[F _i,j] (i, j≤N'), F _i,jfor the value in original image corresponding points; The essence of sampling measures original image signal F=[F _i,j] at sampling matrix φ ' _m(m=1,2 ..., M) on projection

y _m＝<F,φ' _m>

Wherein, m=1,2 ..., M illustrates sampling number, original image signal F=[F _i,j] and sampling matrix size be all N' × N';

By two dimensional image signal F=[F _i,j] by arranging building up a dimensional vector x=[x ₁... x _n], (n=1,2 ..., N), wherein N=N' × N'; And by two-dimentional sampling matrix φ ' _mone dimension row vector φ is stacked as by row transposition _m, then have

y _m＝<F,φ' _m>＝φ _mx；

By y _mpile the matrix y that size is M × 1, namely $y=\left[\begin{array}{c}{y}_1\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ y_M\end{array}\right];$ By sampling matrix φ _minversion φ _m ^tpiling size is by row M × N matrix Φ, namely $\mathrm{\&Phi;}=\left[\begin{array}{c}{\mathrm{\&phi;}}_1\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ {\mathrm{\&phi;}}_M\end{array}\right],$ Then have

y＝Φx

Wherein, sampling matrix Φ is that obedience ± 1 grade is generally with the random matrix of distribution (i.i.d.);

Then, encoder is by sampled value y _mand the sampling matrix φ ' of correspondence _minformation package become one describe bag;

The information of sampling matrix Φ is that transmitting terminal and receiving terminal are shared, and namely sampling matrix Φ is pre-stored in the decoder of transmitting terminal and the decoder of receiving terminal, and encoder only needs when packing describes bag to point out sampling matrix φ ' in description bag _mcorresponding sampling number m.

3. according to claim 1 in multi-hop wireless network based on the method for compressed sensing technical transmission picture signal, it is characterized in that, described step 2, is specially:

(1) the transmission demand factor of link is calculated

If a certain moment t, for comprising sending node N _twith receiving node N _rlink l, sending node N _tdescription bag set in buffer memory is receiving node N _rdescription bag set in buffer memory is suppose that each node can only forward at most the same description bag received once; So, will two kinds of different description bags be had in nodal cache: a kind of is that node receives and forwarded over description bag, be called ash bag; Another kind is that node receives but not yet forwarded over description bag, is called white bag; Note sending node N _tthe set of the Hui Bao in buffer memory, white bag is respectively then have

$U_{N_t}=U_{N_t}^G\&cup;U_{N_t}^W$

If set up link l, then can be that those belong to set for the packet of transmission and do not belong to set the set of description bag; Link l can for the set U of the packet of transmission _pfor

$U_P=\left\{d\right|d\&Element;U_{N_t}^W,d\&NotElement;U_{N_r}\}$

Consider link transmission capacity, then the packet number p that each transmission time slot link l can transmit is

p=min(|U _P|,q(l))

Wherein, q (l)=q (N _t, N _r) represent the description bag number that a transmission time slot link l can transmit, | U _p| represent set U _pin the number of description bag that comprises;

If set up link l, at the reduction Δ RMS of the root-mean-square error of receiving node place restored image be

$\mathrm{\&Delta;PMS}=f_{\mathrm{RMS}}\left(\right|U_{N_r}\left|\right)-f_{\mathrm{RMS}}\left(\right|U_{N_r}|+p)$

Wherein, function f _rMS() is the root-mean-square error RMS of compressed sensing restored image and the relation of the sampled value number for restoring;

Transmission demand factor TDF (l) defining this moment t link l is

$\mathrm{TDF}\left(l\right)=\mathrm{\&Delta;RMS}=f_{\mathrm{RMS}}\left(\right|U_{N_r}\left|\right)-f_{\mathrm{RMS}}\left({|U}_{N_r}\right|+p)$

The transmission demand factor of link has weighed the importance that link is about to the packet of transmission;

(2) the transmission supply factor of link is calculated;

(3) transmission income, the transmission cost of link is calculated

For link transmission requirements factor R evenue (l) and transmission are supplied, factor TSF (l) is to be amassed in transmission income Revenue (l) of definition link l;

Revenue(l)＝TDF(l)·TSF(l)

The transmission income of the interference set link that transmission cost Cost (l) defining link l is l, namely

$\mathrm{Cost}\left(l\right)=\underset{l_i\&Element;S_{\mathrm{int}}\left(l\right)}{\mathrm{\&Sigma;}}\mathrm{Revenue}\left(l_i\right)$

Wherein, the interference S set of link l _int(l) for the interference brought because of link l and the link set that cannot normally work, namely

S _int(l)＝{l _int|l _int≠l,Dis(N _t(l _int),N _r(l))＜r,Dis(N _r(l _int),N _t(l))＜r}

Wherein, N _tl () is the sending node of link l, N _rl () is the receiving node of link l, Dis (N _a, N _b) be two node N _a, N _bbetween geographic distance or channel quality, r is the interference radius of node or the threshold value of channel quality, l _intrepresent the link that can cause interference for link l;

(4) the transmission benefit of link is obtained

Transmission benefit Utility (l) of link l is defined as

$\mathrm{Utility}\left(l\right)=\frac{\mathrm{Revenue}\left(l\right)}{\mathrm{Cost}\left(l\right)}.$

4. according to claim 3 in multi-hop wireless network based on the method for compressed sensing technical transmission picture signal, it is characterized in that, the transmission of the described calculating link supply factor, is specially:

First calculate the node weights of each via node, be specially:

A the joint behavior value Q of () each via node is initialized as 0, node weight weight values W is initialized as 0, and node is initialized as just infinite to the jumping figure H of receiving terminal node; The Q of receiving terminal node is just infinite, and W is just infinite, and H is 0;

B () travels through all node N except receiving terminal node successively _iif have and node N _ithe node N be directly connected _jmeet

$W\left(N_i\right)<\frac{\min (q(N_i,N_j),Q\left(N_j\right))}{(1+H\left(N_j\right))}$

Then

$W\left(N_i\right)\&LeftArrow;\frac{\min (q(N_i,N_j),Q\left(N_j\right))}{(1+H\left(N_j\right))}$

Q(N _i)←min(q(N _i,N _j),Q(N _j))

H(N _i)←1+H(N _j)

Namely N is worked as _iwith N _jfor transmission down hop, and node N _iwhen obtaining more excellent node weight weight values, then more new node N _ijoint behavior value Q, node weight weight values W and node to the jumping figure H of receiving terminal node; Wherein, q (l)=q (N _i, N _j) represent with N _ifor sending node, N _jfor the description bag number that the link l of receiving node can transmit in a transmission time slot, W (N _i) represent node N _iweighted value, Q (N _i) represent node N _iperformance number, H (N _i) represent node N _ito the jumping figure of receiving terminal node;

C () repeats (b) until joint behavior value Q, the node weight weight values W of all nodes and node no longer change in once traversal to the jumping figure H of receiving terminal node; Obtain the node weight weight values W of each node;

Node weights has weighed the ability of each via node to receiving terminal transmission data;

Be N for receiving node _rlink l, link l transmit supply factor TSF (l) be defined as

TSF(l)=W(N _r)

Wherein, W (N _r) represent node N _rweighted value.

5. according to claim 1 in multi-hop wireless network based on the method for compressed sensing technical transmission picture signal, it is characterized in that, described 3rd step, is specially:

(1), when starting, transmission benefit is that positive link is all divided into the link that may be established, and joins the set U of the link that will be established _estLinkin;

(2) the link l that transmission benefit is maximum is found out _max, then from U _estLinkremoving is to link l _maxproduce other links S of interference _int(l _max); And by link l _maxtransmission benefit zero, i.e. Utility (l _max)=0;

(3) (2) are repeated until U _estLinkin link do not interfere with each other, namely then U _estLinkfor the link set that this timeslot scheduling scheme will be set up.

6. according to claim 1 in multi-hop wireless network based on the method for compressed sensing technical transmission picture signal, it is characterized in that, in described 4th step, receiving terminal is decoded to paid-in description bag according to actual conditions.

7. according to claim 6 in multi-hop wireless network based on the method for compressed sensing technical transmission picture signal, it is characterized in that, in described 4th step, if at time t, it is { k that receiving terminal receives K corresponding sampling number ₁, k ₂..., the description bag of K}, wherein if describe the sampled value comprised in bag be $\{y_{k_1},y_{k_2},...,y_K\},$ Be stacked as by sampling number $y_r={\left[\begin{array}{cccc}{y}_{k_1}& y_{k_2}& \&CenterDot;\&CenterDot;\&CenterDot;& y_K\end{array}\right]}^T,$ Corresponding sampling matrix is ${\mathrm{\&Phi;}}_r={\left[\begin{array}{cccc}{\mathrm{\&phi;}}_{k_1}^T& {\mathrm{\&phi;}}_{k_2}^T& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{\&phi;}}_K^T\end{array}\right]}^T;$ Restore original signal, i.e. solving-optimizing problem

min TV(F)

s.t.y _r＝Φ _rV _F

Wherein, column vector V _fbe that two dimensional image signal F is stacking by row, TV (F) represents V _fthe total variance of corresponding two-dimentional original image F.