CN104540147A - Collaborative compression perception quantifying and forwarding method - Google Patents

Abstract

The invention discloses a collaborative compression perception quantifying and forwarding method. The collaborative compression perception quantifying forwarding method disclosed by the invention comprises the following steps: establishing a collaborative communication system model based on a compression perception theory; simultaneously sending information by N source nodes, wherein a channel matrix between each source node and a relay node is used as a measurement matrix, namely a secret key, so that compression perception conversion of information is completed at the relay node; uniformly quantifying information compressed by the relay node; forwarding quantified information to a target node by the relay node; and obtaining original information and a reconstruction error by the target node according to an original compression perception reconstruction algorithm. According to the method, the uniform quantifying method is applied to quantifying a compression perception observation value; and therefore, the whole process can be optimized, the error rate is reduced, and the fact that original signals can be reconstructed by receivers inerrably can be realized.

Description

A kind of cooperation compressed sensing quantizes retransmission method

Technical field

The present invention relates to radio communication physical layer field, particularly relate to a kind of cooperation compressed sensing and quantize retransmission method.

Background technology

Traditional Nyquist sampling theorem requires that the sampling rate of signal is more than or equal to the twice of signal highest frequency, but the sampling process of signal can produce a large amount of redundant datas, and compressed sensing has broken traditional sampling thheorem, sampling and compression are combined, as long as signal meets certain sparse characteristic or become sparse signal under the effect of certain transform-based, just can Exact recovery.The technology of compressed sensing is applied to wireless sensor network, not only can reduces energy ezpenditure but also the quantity of used channel can be reduced, meet the requirement of wireless sensor network.But the measured value that compressed sensing produces not is binary bit stream, and receiving terminal cannot recover initial data accurately, therefore need to adopt quantification manner that observation signal is transformed to binary bit stream to transmit, thus optimize whole transmitting procedure, reduce the error rate of reconstruct, ensure that recipient can reconstruct primary signal without error code.

Summary of the invention

The invention provides a kind of cooperation compressed sensing and quantize retransmission method, this method, based on the quantization algorithm of compressive sensing theory, is applied in multi-node communication model, reduces the error rate of reconstruct, ensure that recipient can reconstruct primary signal without error code substantially, described below:

A kind of cooperation compressed sensing quantizes retransmission method, and described cooperation compressed sensing quantizes retransmission method and comprises the following steps:

Cooperation communication system model is set up based on compressive sensing theory;

N number of source node sends message simultaneously, using the channel matrix between source node to via node as calculation matrix, namely as key, makes message complete compressed sensing conversion at via node place;

Via node carries out uniform quantization to the message after compression; Message after quantification is transmitted to destination node by via node;

Destination node, according to existing compressed sensing restructing algorithm, obtains origination message and reconstructed error.

Described cooperation communication system model is specially:

N number of source node forwards through M relay node cooperation, communicates with destination node; This process is subject to the eavesdropping of E eavesdropping node, supposes that the signal degree of rarefication that source node sends is K, K < < N, M >=cKlog (N/K), the constant of c >=4.

Wherein, the message that described via node receives is:

$Y_{\mathrm{sr}}=\sqrt{P_s}{H}_{\mathrm{sr}}X+n_{\mathrm{sr}}$

Wherein, P _sfor the transmitted power of each source node, H _srfor source node is to the channel matrix of via node; n _srfor source node is to the noise of the channel of via node; X is the transmission information of source node.

Wherein, described via node carries out uniform quantization be specially the message after compression:

$Q_{\mathrm{sr}}=\underset{q\&Element;Q}{\arg \min }\left|\right|q-R_{\mathrm{sr}}\left|\right|$

Wherein, the computing of independent variable value when expression obtains maximum, Q _srfor quantize after message, q be selected code book concentrate closest to message R _srcode element.

The beneficial effect of technical scheme provided by the invention is: the present invention proposes the quantization algorithm based on compressive sensing theory, be applied to multi-node communication model, utilize the characteristic of access channel, message is sent to via node by multiple source signals simultaneously, the calculation matrix of channel matrix as compressed sensing is compressed, then at via node, quantification is carried out to the signal after compression and forward.The method of uniform quantization is applied in the quantification of compressed sensing measured value and can optimizes whole process, reduce the error rate, ensure that recipient substantially inerrably can reconstruct primary signal.

Accompanying drawing explanation

Fig. 1 is the schematic diagram of cooperation compression perceptual system mode;

Fig. 2 is the flow chart that a kind of compressed sensing that cooperates quantizes retransmission method;

Fig. 3 is primary signal and reconstruction signal comparison diagram;

Embodiment

For making the object, technical solutions and advantages of the present invention clearly, below embodiment of the present invention is described further in detail.

The invention provides a kind of cooperation compressed sensing and quantize retransmission method, see Fig. 1 and Fig. 2, this retransmission method comprises the following steps:

101: set up cooperation communication system model based on compressive sensing theory;

Wherein, compressive sensing theory can be divided into three phases: rarefaction representation stage, projection stage and the reconstruction stage of signal.Any signal can resolve into the linear combination of other forms of signal, conventional as discrete cosine transform in message area, wavelet transformation etc., essence of all these conversion are all find the sparse base of signal, obtain a kind of more succinct representation of signal, i.e. rarefaction representation form.Afterwards by the reduction process of projection matrix settling signal, obtain projection value, projection matrix and sparse basis array form the calculation matrix of compressed sensing.Finally, according to projection value and calculation matrix.By some restructing algorithms, as orthogonal matching pursuit algorithm OMP (Orthogonal Matching Pursuit), BP (Basis Pursuit) algorithm reconstruct primary signal followed the trail of by base.

See Fig. 1, based on above-mentioned compressive sensing theory, the present invention sets up cooperation communication system model and is specially: N number of source node forwards through M relay node cooperation, communicates with destination node D.This process is subject to the eavesdropping of E eavesdropping node, supposes that the signal degree of rarefication that source node sends is K, K < < N, M >=cKlog (N/K), the constant of c >=4.

102:N source node sends message simultaneously, using the channel matrix between source node to via node as calculation matrix, namely as key, makes message complete compressed sensing conversion at via node place, also just completes ciphering process;

Namely, N number of source node sends message simultaneously, utilizes multiple access to access the characteristic of channel, using the channel matrix between source node to via node as calculation matrix, measured value is quantized at via node, makes signal complete compressed sensing transform and quantization at via node place.

The message that via node and eavesdropping node receive is respectively:

$Y_{\mathrm{sr}}=\sqrt{P_s}{H}_{\mathrm{sr}}X+n_{\mathrm{sr}}---\left(1\right)$

$Y_{\mathrm{se}}=\sqrt{P_s}{H}_{\mathrm{se}}X+n_{\mathrm{se}}---\left(2\right)$

Wherein, P _sfor the transmitted power of each source node, H _srfor source node is to the channel matrix of via node; H _sefor source node is to the channel matrix of eavesdropping node; n _srfor source node is to the noise of the channel of via node; n _sefor source node is to the noise of the channel of eavesdropping node.H _srand H _sebe the calculation matrix in compressive sensing theory, X is the message that information source sends, Y _srfor the message that via node receives, Y _sefor the message that eavesdropping node receives.

103: via node carries out uniform quantization to the message after compression;

Wherein, if the code book quantized integrates as Q, uniform quantization is exactly from code book, select the code element closest to measured value, ensures that quantization error is minimum:

$Q_{\mathrm{sr}}=\underset{q\&Element;Q}{\arg \min }\left|\right|q-R_{\mathrm{sr}}\left|\right|---\left(3\right)$

Wherein, the computing of independent variable value when expression obtains maximum, Q _srfor quantize after message, q be selected code book concentrate closest to message R _srcode element.

104: the message after quantification is transmitted to destination node by via node;

Wherein, quantize to forward through relaying, the signal that destination node and eavesdropping node receive is respectively:

Y＝H _rdQ _sr+n _rd(4)

E _re＝H _reQ _sr+n _re(5)

Wherein, Q _srfor the message after quantification, H _rdfor via node is to the channel matrix of destination node; H _refor via node is to the channel matrix of eavesdropping node; n _rdfor via node is to the noise of the channel of destination node; n _refor via node is to the noise of the channel of eavesdropping node.

105: destination node, according to existing compressed sensing restructing algorithm, obtains origination message and reconstructed error.

make ${\left|\right|Y_{\mathrm{rd}}/\sqrt{P_s}-H_{\mathrm{rd}}\mathrm{\&beta;}{H}_{\mathrm{sr}}\hat{x}\left|\right|}_2<\mathrm{\&epsiv;}---\left(6\right)$

Wherein, reconstruction signal, || || ₁represent a norm, || || ₂represent two norms, ε is the upper limit of noise amplitude, and minimize refers to find minimum value.

In order to verify the performance of said method, utilize experimental situation to emulate, as shown in Figure 3, draw thus and quantized to forward by observation signal, receiving terminal can reconstruct primary signal to less mistake to simulation result.

It will be appreciated by those skilled in the art that accompanying drawing is the schematic diagram of a preferred embodiment, the invention described above embodiment sequence number, just to describing, does not represent the quality of embodiment.

The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (4)

1. the compressed sensing that cooperates quantizes a retransmission method, it is characterized in that, described cooperation compressed sensing quantizes retransmission method and comprises the following steps:

Cooperation communication system model is set up based on compressive sensing theory;

N number of source node sends message simultaneously, using the channel matrix between source node to via node as calculation matrix, namely as key, makes message complete compressed sensing conversion at via node place;

Via node carries out uniform quantization to the message after compression; Message after quantification is transmitted to destination node by via node;

Destination node, according to existing compressed sensing restructing algorithm, obtains origination message and reconstructed error.

2. one cooperation compressed sensing according to claim 1 quantizes retransmission method, and it is characterized in that, described cooperation communication system model is specially:

N number of source node forwards through M relay node cooperation, communicates with destination node; This process is subject to the eavesdropping of E eavesdropping node, supposes that the signal degree of rarefication that source node sends is K, K < < N, M >=cKlog (N/K), the constant of c >=4.

3. one cooperation compressed sensing according to claim 1 quantizes retransmission method, and it is characterized in that, the message that described via node receives is:

$Y_{\mathrm{sr}}=\sqrt{P_s}{H}_{\mathrm{sr}}X+n_{\mathrm{sr}}$

Wherein, P _sfor the transmitted power of each source node, H _srfor source node is to the channel matrix of via node; n _srfor source node is to the noise of the channel of via node; X is the transmission information of source node.

4. one cooperation compressed sensing according to claim 1 quantizes retransmission method, and it is characterized in that, described via node carries out uniform quantization to the message after compression and is specially:

$Q_{\mathrm{sr}}=\underset{q\&Element;Q}{\arg \min }\left|\right|q-R_{\mathrm{sr}}\left|\right|$

Wherein, the computing of independent variable value when expression obtains maximum, Q _srfor quantize after message, q be selected code book concentrate closest to message R _srcode element.