CN106161297A - In ofdm system, anti-pilot tone spoofing attack channel based on independent component analysis is estimated and recognition methods - Google Patents

Abstract

The present invention provides anti-pilot tone spoofing attack channel based on independent component analysis in a kind of ofdm system to estimate and recognition methods, the characteristic of random pilot is utilized to attack to weaken pilot tone spoofing attack for Pilot Interference, the legitimate channel estimated according to the autonomous behavior of channel simultaneously and identify under Pilot Interference environment between up legal transmitter and receiver.The present invention has only to spend the time resource of 3 OFDM symbol and a number of band resource, and immune to the Pilot Interference power of active listener-in.Up legal transmitter is upset the premeditated pilot tone of actively listener-in and is attacked by sending uplink random pilot tone, simultaneously, receiver is by utilizing the mutual independence between legitimate channel and attack channel, legitimate channel information is extracted from the contaminated signal received, promote the precision of channel estimation of legitimate channel, thus solve the problem that precision of channel estimation that pilot tone spoofing attack causes is low.

Description

In ofdm system anti-pilot tone spoofing attack channel based on independent component analysis estimate and Recognition methods

Technical field

The invention belongs to wireless communication field, the channel relating to anti-pilot tone spoofing attack in ofdm system is estimated and cognitron System, specially in ofdm system, anti-pilot tone spoofing attack channel based on independent component analysis is estimated and recognition methods.

Background technology

Along with the fast development of wireless communication technology, security risk the most gradually threatens and governs wireless communication technology Application and development.The broadcast characteristic of wireless channel so that unwanted signals source eavesdrop the secret of other nodes and private information or Person endangers the communication link disturbing other normal node.At present, the password encryption deciphering system on upper strata can be to a certain extent Ensureing the safety of wireless communications environment, but be as the development of computer technology, unwanted signals source often has powerful meter Calculate, eavesdropping and attacking ability, even can decode the encryption system on upper strata completely, and then cause the information of validated user to be completely exposed In oneself.Pay close attention to widely to this end, safety of physical layer mechanism causes and study, from initial theory α coefficient performance evaluation to In recent years towards the safety of physical layer Mechanism Design of practical communication system, safety of physical layer technology is evolving always, but, Particularly the most little towards the anti-interference Mechanism Design of practical communication system towards the physical layer research of anti-interference attack.OFDM system System, as the wireless communication system architecture of current main-stream, is widely used in the wireless communication system of each reality.But, due to Agreement and the relatively-stationary characteristic of framework, assailant often can obtain these information, and then system causes serious threat.

The angle estimated from physical layer channel, when a listener-in learns frame synchronization information and the pilot tone letter of legal transceiver After number, can start pilot tone spoofing attack, period, this assailant believes by synchronously sending identical pilot tone with a certain validated user Number, between the pairing of interfering legality transceiver, channel is estimated, thus reduces the security performance of system, promotes security risk.And, lead Frequently, after spoofing attack once successful implementation, legal transceiver is difficult to obtain accurate legal downlink channel state information again.? In ofdm system, channel is estimated to need to rely on the definitiveness pilot signal in different sub carrier, and these pilot signals can be hacked Person is known, and then implements pilot tone spoofing attack.Accordingly, it would be desirable to new channel estimation mechanism and pilot design mechanism are resisted and are cut The impact that channel is estimated by the pilot tone spoofing attack in weak ofdm system.

Summary of the invention

For problems of the prior art, the present invention provides based on independent component analysis anti-in a kind of ofdm system Pilot tone spoofing attack channel is estimated and recognition methods, makes full use of the channel between legal transceiver and actively listener-in's extremely reception The independence of channel between machine, in conjunction with pilot tone feedback mechanism, only need to take the time of 3 OFDM symbol, just can significantly promote The estimated accuracy of legitimate channel, and the impact of this precision the most under fire jamming power.

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

In ofdm system, anti-pilot tone spoofing attack channel based on independent component analysis is estimated and recognition methods, including as follows Step,

Step1, regards the pilot signal in ofdm system as linear system, and channel regards the signal source of input system as, turns Change the reception model of pilot signal；

Step2, random pilot signal launched by the legal transmitter of up-link in an ofdm system, after receiver is based on conversion The pilot signal received is analyzed by the model that receives of pilot signal, detects two kinds of behavioral pattern of actively listener-in；

Step3, according to the behavioral pattern of the assailant detected, receiver takes corresponding signal analysis strategy to carry out letter Road is estimated: if be detected that attack, and uses CS-JADE algorithm to carry out channel estimation, otherwise, uses LS algorithm to carry out channel and estimate Meter；

Step4, after channel is estimated to terminate, the random pilot signal of up legal transmitter feedback channel estimated service life is connect Receipts machine；Receiver if be detected that pilot tone is attacked, then utilizes in the channel that this random pilot signal estimates from Step3 and identifies Channel information between the pairing of legal transceiver；If be detected that occur without attacking, according to LS algorithm, receiver directly knows conjunction Channel information between method transceiver.

Preferably, ofdm system includes a pair legal transceiver and an active listener-in, and two of formation up logical Letter link is respectively 1. up legal transmitted from transmitter to receiver, is defined as legitimate channel；2. actively listener-in, to receiver, defines For attack channel；Receiver has N_TRoot antenna, up legal transmitter and actively listener-in are single antenna, every uplink The every antenna on road takies N number of subcarrier in each OFDM symbol simultaneously；Channel coefficients on each subcarrier is obeyed Average be 0 variance be the independent same distribution of 1, the noise on each subcarrier obey average be 0 variance be the independent same distribution of 1 White Gaussian noise.

Further, Step1 converts the pilot signal reception model obtained as follows,

$\tilde{Y}=\tilde{D}H+\tilde{V};$

Wherein,

$\tilde{Y}=\left[\begin{array}{c}{y}_j^T\\ y_k^T\\ y_t^T\end{array}\right],\tilde{D}=\left\{\begin{array}{cc}{\left[\begin{array}{ccc}{d}_{L,j}^H& d_{L,k}^H& d_{L,t}^H\end{array}\right]}^H& I\\ {\left[\begin{array}{ccc}{d}_{L,j}^H& d_{L,k}^H& d_{L,t}^H\\ d_{E,j}^H& d_{E,k}^H& d_{E,t}^H\end{array}\right]}^H& II\end{array}\right.,\tilde{V}=\left[\begin{array}{ccc}{v}_{L,(1,j)}^T& \mathrm{...}& v_{L,(N_T,j)}^T\\ v_{L,(1,k)}^T& \mathrm{...}& v_{L,(N_T,k)}^T\\ v_{L,(1,t)}^T& \mathrm{...}& v_{L,(N_T,t)}^T\end{array}\right];$

$y_j={\[y_{(i,j)}^T,\mathrm{...},y_{(N_T,j)}^T\]}^T,H=\left\{\begin{array}{cc}\left[\begin{array}{ccc}{h}_{L,(1,j)}^T& \mathrm{...}& h_{L,(N_T,j)}^T\end{array}\right]& I\\ \left[\begin{array}{ccc}{h}_{L,(1,j)}^T& \mathrm{...}& h_{L,(N_T,j)}^T\\ h_{E,(1,j)}^T& \mathrm{...}& h_{E,(N_T,j)}^T\end{array}\right]& II\end{array}\right.;$

J, k, t are three OFDM symbol,The signal received for receiver,For pilot signal,For receiver Receiving terminal noise, H is channel, y_(i,j)The signal received at jth time slot for i-th antenna of receiver；d_L,(i,j)(n), d_E,(i,j)(n) be respectively up legal transceiver user and actively listener-in in jth OFDM symbol, i-th antenna, n-th The uplink pilot signal used in sub-carrier channels and between receiver；For receiver at i-th antenna jth time slot Interior noise；h_L,(i,j), h_E,(i,j)It is respectively descending legitimate receiver and active listener-in at jth time slot to descending legal Penetrate the sub-carrier channels of i-th antenna of machine.

Further, the CS-JADE algorithm described in Step3 comprises the steps,

3.1 any 2 the time slot j stored according to receiver, the data that k receivesUse under receiver line JADE method estimates contaminated channel matrix

3.2 according to Y, calculates weighting matrix W and makes it meet:Estimate weighting matrix S simultaneously^H；

3.3 calculate weighting matrix G makes it meet

Wherein,d_L,j, d_L,kObtained by the pilot signal fed back；

3.4 to estimate the uplink sub-carrier channel vector of up legal transmitter as follows:

${\hat{h}}_L=PGWY={\mathrm{PGS}}_{opt}\hat{H},$

Wherein: P meets

Preferably, Step2 uses the method based on the noise subspace signal to receivingDetect, to judge Whether there is active attack, specific as follows,

First, rightCarry out Eigenvalues Decomposition, obtain three eigenvalue σ₀≤σ₁≤σ₂；

Next, if σ₁-σ₀=0, then do not occur pilot tone to attack, otherwise, there is pilot tone and attack.

Preferably, when in Step4, the random pilot signal of user feedback channel estimated service life is to receiver, according to such as ShiShimonoseki System carries out the feedback of pilot signal,

θ_j=θ_k+θ₀=θ_t+θ₁；

Wherein, θ_j,θ_k,θ_tIt is respectively the phase place of random pilot signal corresponding to three OFDM symbol time, and meets fixing Phase contrast, corresponding phase contrast is respectively θ₀,θ₁；.

Compared with prior art, the present invention has a following useful technique effect:

The present invention utilizes the characteristic of random pilot to attack to weaken pilot tone spoofing attack for Pilot Interference, simultaneously according to channel The autonomous behavior legitimate channel estimating and identify under Pilot Interference environment between up legal transmitter and receiver.This The bright time resource having only to spend 3 OFDM symbol and a number of band resource, and the pilot tone to active listener-in Jamming power immunity.Up legal transmitter is upset the premeditated pilot tone of actively listener-in and is attacked by sending uplink random pilot tone, Meanwhile, receiver is by utilizing the mutual independence between legitimate channel and attack channel, from the contaminated signal received In extract legitimate channel information, promote the precision of channel estimation of legitimate channel, thus solve the letter that pilot tone spoofing attack causes The problem that road estimated accuracy is low.The present invention uses dual stage process help system to improve the estimation to legitimate channel simultaneously；First In the stage, random pilot launched by up legal transmitter；Second stage, the pilot tone that the up legal transmitter feedback first stage uses Signal is to receiver；Receiver, if be detected that pilot tone is attacked, facilitates the channel estimated further by this information from the first stage In identify legitimate channel information；If occurred without attacking, according to traditional LS algorithm, it is legal that uplink receiver can directly be known Channel information.

Further, estimate owing to random pilot interfering legality channel may be sent first stage actively listener-in, logical Cross the signal that receiver receives by method based on noise subspace to detect, to judge whether active attack occurs, as Fruit detects attack, utilizes the CS-JADE algorithm proposed to carry out channel estimation, otherwise, uses traditional LS algorithm to carry out channel Estimate；The uplink channel estimation of validated user is as target, by according with three OFDM to improve under pilot tone attack context for the present invention Statistical Analysis of Signals received on subcarrier in number time, just can recover legal use uplink receiving end is high-precision The uplink channel information at family.Further, the impact of the precision of channel estimation brought jamming power the most under fire of being suggested plans, from And the robustness of significantly boosting algorithm.

Accompanying drawing explanation

Fig. 1 is time-frequency structure schematic diagram.

Fig. 2 is the FB(flow block) of method described in present example.

Fig. 3 is the normalized channel mean square error described in present example and legal pilot SNR curve chart.

Fig. 4 is the normalized channel mean square error described in present example and antenna and subcarrier number curve chart.

Fig. 5 is that under the disturbance power described in present example, normalized channel mean square error is believed with legal pilot tone Make an uproar and compare curve chart.

Detailed description of the invention

Below in conjunction with specific embodiment, the present invention is described in further detail, described in be explanation of the invention and It not to limit.

Method of the present invention, in system model when setting up, it is considered to a pair legal transceiver and one actively eavesdropping Person, two uplink communication links, respectively 1. upstream transmitter → receiver, it is defined as legitimate channel 2. actively listener-in → connect Receipts machine is defined as attack channel.Receiver has N_TRoot antenna, upstream transmitter and actively listener-in are single antenna, such as Fig. 1 Shown in, every antenna of every up-link takies N number of subcarrier in each OFDM symbol simultaneously.On each subcarrier It is 0 that channel coefficients all obeys average, and variance is the independent same distribution of 1, and it is 0 that the noise on each subcarrier obeys average, side Difference is the independent same distribution white Gaussian noise of 1.If upstream transmitter uses definitiveness pilot tone, when pilot pollution attacks generation, Actively listener-in can launch identical pilot signal in the pilot point position identical with up legal transmitter, when up legal When transmitter uses random pilot, actively listener-in may launch random in the pilot point position identical with legal transmitter Pilot signal carries out channel and estimates interference.Wherein, the signal to noise ratio of pilot signal is respectively,

|d_L,(i,j)(n)|²=ρ_L,|d_E,(i,j)(n)|²=ρ_E,0≤n≤N-1；

In order to prevent pilot pollution, upstream transmitter uses random pilot to carry out channel estimation, and now, actively listener-in can Take two kinds of behavioral pattern: 1. have to use the channel of random pilot interfering legality channel to estimate, then, pilot tone spoofing attack The most just it is converted into Pilot Interference to attack；2. keep mourning in silence.So-called actively listener-in is the one legal transceiver of only active interference The listener-in that channel is estimated.

Fig. 2 gives the system framework figure of described method, comprises the steps:

Step 1, for the problem of channel coherency time resource-constrained, it is proposed that new pilot signal reception model；Pass through Convert and obtain new pilot signal reception model, uplink receiver is received in the jth OFDM symbol of i-th antenna Pilot signal:

$y_{(i,j)}=\left\{\begin{array}{cc}{H}_{L,(i,j)}{d}_{L,(i,j)}+v_{L,(i,j)}& I\\ H_{L,(i,j)}{d}_{L,(i,j)}+H_{E,(i,j)}{d}_{E,(i,j)}+v_{L,(i,j)}& II\end{array}\right.;$

H_L,(i,j)=diag{h_L,(i,j)(0),...,h_L,(i,j)(N-1)}

H_E,(i,j)=diag{h_E,(i,j)(0),...,h_E,(i,j)(N-1)}；

d_L,(i,j)=[d_L,(i,j)(0),...,d_L,(i,j)(N-1)]^T

d_E,(i,j)=[d_E,(i,j)(0),...,d_E,(i,j)(N-1)]^T

v_L,(i,j)=[v_L,(i,j)(0),...,v_L,(i,j)(N-1)]^T；

Wherein, I represents that not receiving pilot tone attacks, and II represents and attacked by pilot tone, h_L,(i,j)(n) and d_L,(i,j)N () is respectively In jth OFDM symbol, the n-th uplink sub-carrier channel between up legal transmitter to i-th antenna of receiver and upper Row pilot signal, h_E,(i,j)(n) and d_E,(i,j)N () is respectively in jth OFDM symbol, actively listener-in is to i-th sky of receiver The n-th uplink sub-carrier channel and ascending pilot frequency between line, v_L,(i,j)N () is that receiver is in i-th sky of jth OFDM symbol The noise on the n-th uplink sub-carrier channel on line.

To this end, by making each user identical with all uplink pilot signal on antenna at each subcarrier:

d_L,(i,j)(0)=...=d_L,(i,j)(N-1)=d_L,j

$d_{E,(i,j)}\left(0\right)=\mathrm{...}=d_{E,(i,j)}(N-1)=d_{E,j},\∀i\∈\{1,\mathrm{...},N_T\};$

We obtain three OFDM symbol j, pilot signal reception model new in k, t:

$\tilde{Y}=\tilde{D}H+\tilde{V};$

Wherein:

$\tilde{Y}=\left[\begin{array}{c}{y}_j^T\\ y_k^T\\ y_t^T\end{array}\right],\tilde{D}=\left\{\begin{array}{cc}{\left[\begin{array}{ccc}{d}_{L,j}^H& d_{L,k}^H& d_{L,t}^H\end{array}\right]}^H& I\\ {\left[\begin{array}{ccc}{d}_{L,j}^H& d_{L,k}^H& d_{L,t}^H\\ d_{E,j}^H& d_{E,k}^H& d_{E,t}^H\end{array}\right]}^H& II\end{array}\right.,\tilde{V}=\left[\begin{array}{ccc}{v}_{L,(1,j)}^T& \mathrm{...}& v_{L,(N_T,j)}^T\\ v_{L,(1,k)}^T& \mathrm{...}& v_{L,(N_T,k)}^T\\ v_{L,(1,t)}^T& \mathrm{...}& v_{L,(N_T,t)}^T\end{array}\right];$

$y_j={\[y_{(i,j)}^T,\mathrm{...},y_{(N_T,j)}^T\]}^T,H=\left\{\begin{array}{cc}\left[\begin{array}{ccc}{h}_{L,(1,j)}^T& \mathrm{...}& h_{L,(N_T,j)}^T\end{array}\right]& I\\ \left[\begin{array}{ccc}{h}_{L,(1,j)}^T& \mathrm{...}& h_{L,(N_T,j)}^T\\ h_{E,(1,j)}^T& \mathrm{...}& h_{E,(N_T,j)}^T\end{array}\right]& II\end{array}\right.;$

h_L,(i,j)=[h_L,(i,j)(0),...,h_L,(i,j)(N-1)]^T；h_E,(i,j)=[h_E,(i,j)(0),...,h_E,(i,j)(N- 1)]^T；

Unlike prior art, pilot signalBeing considered as linear hybrid system, H is considered as signal Source.

Based on pilot signal reception model mentioned above, the present invention anti-pilot tone based on independent component analysis spoofing attack Channel is estimated and recognition methods is divided into two stages: the first stage, up legal transmitter is launched random pilot, now actively stolen Hearer has two kinds of selections: 1. send random pilot interfering legality channel and estimate, 2. keep mourning in silence, do not send the legal letter of Pilot Interference Road is estimated；In order to detect the behavior of actively listener-in, uplink receiver uses method based on noise subspace to receiving Signal detects, to judge whether active attack occurs.If be detected that attack, uplink receiver utilizes CS-JADE algorithm Carry out channel estimation, otherwise, use traditional LS algorithm to carry out channel estimation；Second stage, up legal transmitter feedback the The pilot signal that one stage used is to receiver；Receiver if be detected that pilot tone is attacked, convenient by this information further from the The channel that one stage was estimated identifies legitimate channel information；If occurred without attacking, according to traditional LS algorithm, uplink receiving Machine can directly know legitimate channel information.It is implemented as follows:

Step2, in order to detect whether pilot tone attack exists, uses the method based on the noise subspace signal to receivingIt is analyzed: first, rightCarry out Eigenvalues Decomposition, obtain three eigenvalue: σ₀≤σ₁≤σ₂；Next, if σ₁-σ₀ =0, then do not occur pilot tone to attack, otherwise, there is pilot tone and attack.

In order to improve the estimated accuracy of validated user channel, it is proposed that a kind of pilot signal feedback mechanism, this machining function After pilot tone estimation procedure, specific design mechanism is as follows:

θ_j=θ_k+θ₀=θ_t+θ₁；

Wherein, θ_j,θ_k,θ_tIt is respectively random pilot signal corresponding to three OFDM symbol time, but meets fixing phase Potential difference θ₀,θ₁。

Step3, according to above result, attacks if there is pilot tone, then perform CS-JADE algorithm carry out channel estimate and Identify, otherwise, directly perform traditional LS channel and estimate；

The CS-JADE algorithm proposed carries out channel estimation and identification, and this algorithm is by means of the pilot signal of feedback so that connect Receipts machine can online under complete channel estimate and identify, estimate legal sub-carrier channels vectorSpecifically comprise the following steps that

Step3.1: according to any 2 time slot j of receiver storage, the data that k receivesUnder receiver line JADE method is used to estimate contaminated channel matrix

Step3.2: according to Y, calculates weighting matrix W and makes it meet:Estimate weighting matrix simultaneously S^H；

Step3.3: calculate weighting matrix G and make it meet:Wherein, d_L,j, d_L,kCan be obtained by the pilot signal fed back；

Step3.4: estimate the uplink sub-carrier channel vector of up legal transmitter:

${\hat{h}}_L=PGWY={\mathrm{PGS}}_{opt}\hat{H},$

Wherein: P meets:

Step4, after estimating to terminate based on channel, the random pilot signal of up legal transmitter feedback channel estimated service life To receiver, up legal transmitter only need to feed back θ_j,θ_k,θ_tIn some phase information, receiver is according to this phase place Information and known phase difference θ₀,θ₁, all pilot signals can be known.

The simulating, verifying of the present invention is expressed as Fig. 3, Fig. 4 and Fig. 5.Fig. 3 illustrates this invention in pilot tone spoofing attack High-precision can recover legitimate channel under environment, break the precision of channel estimation platform caused that pilot tone is attacked.Fig. 4 represents The present invention estimate the normalized mean squared error of channel along with antenna number and subcarrier number situation of change, along with antenna number Or the rising of subcarrier number, the precision of channel estimation of the present invention gradually tends towards stability.Fig. 5 indicates the present invention and leads in difference Frequently the robustness of the normalized mean squared error performance under the conditions of jamming power, especially, along with active attack person's jamming power Being gradually increased, the precision of channel estimation of the present invention is little affected by the impact of jamming power.

Claims (6)

1. In 1.OFDM system, anti-pilot tone spoofing attack channel based on independent component analysis is estimated and recognition methods, and its feature exists In, comprise the steps,

   Step1, regards the pilot signal in ofdm system as linear system, and channel regards the signal source of input system as, and conversion is led Frequently the reception model of signal；

   Step2, random pilot signal launched by the legal transmitter of up-link in an ofdm system, and receiver is based on pilot tone after conversion The pilot signal received is analyzed by the model that receives of signal, detects two kinds of behavioral pattern of actively listener-in；

   Step3, according to the behavioral pattern of the assailant detected, receiver is taked corresponding signal analysis strategy to carry out channel to estimate Meter: if be detected that attack, uses CS-JADE algorithm to carry out channel estimation, otherwise, uses LS algorithm to carry out channel estimation；

   Step4, after channel is estimated to terminate, the random pilot signal of up legal transmitter feedback channel estimated service life is given and is received Machine；Receiver if be detected that pilot tone is attacked, then utilizes in the channel that this random pilot signal estimates from Step3 and identifies conjunction Channel information between the pairing of method transceiver；If be detected that occur without attacking, according to LS algorithm, it is legal that receiver is directly known Channel information between transceiver.
2. In ofdm system the most according to claim 1 anti-pilot tone spoofing attack channel based on independent component analysis estimate and Recognition methods, it is characterised in that described ofdm system includes a pair legal transceiver and an active listener-in, formation Two uplink communication links are respectively 1. up legal transmitted from transmitter to receiver, are defined as legitimate channel；2. actively listener-in arrives Receiver, is defined as attack channel；Receiver has N_TRoot antenna, up legal transmitter and actively listener-in are Dan Tian Line, every antenna of every up-link takies N number of subcarrier in each OFDM symbol simultaneously；Letter on each subcarrier Road coefficient all obey average be 0 variance be the independent same distribution of 1, the noise on each subcarrier obey average be 0 variance be 1 Independent same distribution white Gaussian noise.
3. In ofdm system the most according to claim 2 anti-pilot tone spoofing attack channel based on independent component analysis estimate and Recognition methods, it is characterised in that convert the pilot signal reception model obtained in Step1 as follows,

   $\tilde{Y}=\tilde{D}H+\tilde{V};$

   Wherein,

   $\tilde{Y}=\left[\begin{array}{c}{y}_j^T\\ y_k^T\\ y_t^T\end{array}\right],\tilde{D}=\left\{\begin{array}{cc}{\left[\begin{array}{ccc}{d}_{L,j}^H& d_{L,k}^H& d_{L,t}^H\end{array}\right]}^H& I\\ {\left[\begin{array}{ccc}{d}_{L,j}^H& d_{L,k}^H& d_{L,t}^H\\ d_{E,j}^H& d_{E,k}^H& d_{E,t}^H\end{array}\right]}^H& II\end{array}\right.,\tilde{V}=\left[\begin{array}{ccc}{v}_{L,(1,j)}^T& \mathrm{...}& v_{L,(N_T,j)}^T\\ v_{L,(1,k)}^T& \mathrm{...}& v_{L,(N_T,k)}^T\\ v_{L,(1,t)}^T& \mathrm{...}& v_{L,(N_T,t)}^T\end{array}\right];$

   $y_j={\[y_{(i,j)}^T,\mathrm{...},y_{(N_T,j)}^T\]}^T,H=\left\{\begin{array}{cc}\left[\begin{array}{ccc}{h}_{L,(1,j)}^T& \mathrm{...}& h_{L,(N_T,j)}^T\end{array}\right]& I\\ \left[\begin{array}{ccc}{h}_{L,(1,j)}^T& \mathrm{...}& h_{L,(N_T,j)}^T\\ h_{E,(1,j)}^T& \mathrm{...}& h_{E,(N_T,j)}^T\end{array}\right]& II\end{array}\right.;$

   J, k, t are three OFDM symbol,The signal received for receiver,For pilot signal,Receiving terminal for receiver Noise, H is channel, y_(i,j)The signal received at jth time slot for i-th antenna of receiver；d_L,(i,j)(n), d_E,(i,j)(n) It is respectively up legal transceiver user and active listener-in in jth OFDM symbol, i-th antenna, the n-th sub-carrier channels The uplink pilot signal above and used between receiver；For receiver noise in i-th antenna jth time slot； h_L,(i,j), h_E,(i,j)It is respectively descending legitimate receiver and active listener-in at the i-th of jth time slot to descending legal transmitter The sub-carrier channels of root antenna.
4. In ofdm system the most according to claim 3 anti-pilot tone spoofing attack channel based on independent component analysis estimate and Recognition methods, it is characterised in that the CS-JADE algorithm described in Step3 comprises the steps,

   3.1 any 2 the time slot j stored according to receiver, the data that k receivesJADE method is used under receiver line Estimate contaminated channel matrix

   3.2 according to Y, calculates weighting matrix W and makes it meet:Estimate weighting matrix S simultaneously^H；

   3.3 calculate weighting matrix G makes it meet

   Wherein,d_L,j, d_L,kObtained by the pilot signal fed back；

   3.4 to estimate the uplink sub-carrier channel vector of up legal transmitter as follows:

   ${\hat{h}}_L=PGWY={\mathrm{PGS}}_{opt}\hat{H},$

   Wherein: P meets
5. In ofdm system the most according to claim 1 anti-pilot tone spoofing attack channel based on independent component analysis estimate and Recognition methods, it is characterised in that use the method based on the noise subspace signal to receiving in Step2Detect, It is to judge whether to occur active attack, specific as follows,

   First, rightCarry out Eigenvalues Decomposition, obtain three eigenvalue σ₀≤σ₁≤σ₂；

   Next, if σ₁-σ₀=0, then do not occur pilot tone to attack, otherwise, there is pilot tone and attack.
6. In ofdm system the most according to claim 1 anti-pilot tone spoofing attack channel based on independent component analysis estimate and Recognition methods, it is characterised in that when the random pilot signal of user feedback channel estimated service life is to receiver in Step4, according to Following relation carries out the feedback of pilot signal,

   $d_{L,j}=e^{{\mathrm{i\θ}}_j},d_{L,k}=e^{{\mathrm{i\θ}}_k},d_{L,t}=e^{{\mathrm{i\θ}}_t},{\mathrm{\θ}}_j={\mathrm{\θ}}_k+{\mathrm{\θ}}_0={\mathrm{\θ}}_t+{\mathrm{\θ}}_1;$

   Wherein, θ_j,θ_k,θ_tIt is respectively the phase place of random pilot signal corresponding to three OFDM symbol time, and meets fixing phase Potential difference, corresponding phase contrast is respectively θ₀,θ₁。