Method for combating large-scale access pilot contamination attack based on channel decontamination and communication system

Abstract

The application discloses a kind of based on channel decontamination large-scale access pilot pollution attack method and communication system: based on millimeter wave unlicensed mMTC network construction uplink multiple-input multiple-output signal communication model, three-dimensional transmission model and channel model are constructed in angle-delay-time slot domain;Based on the three-dimensional model constructed, the PCA countermeasure problem is converted into the multi-dimensional sparse signal recovery problem based on multi-dimensional dictionary;By utilizing the sparsity of millimeter wave channel in virtual angle domain and delay domain, and the time correlation of legal LU active state, a sparse adaptive matching pursuit algorithm based on multi-dictionary is designed to recover the CVR matrix;Finally, according to the activity estimation, PCA detection and channel decontamination criterion, the channel reconstruction of all LUs is realized;The application improves the accuracy of user channel estimation and data decoding, improves the performance of anti-pilot pollution attack, and ensures communication security.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of random access and data transmission in large-scale access networks, and particularly relates to a large-scale access pilot contamination attack countermeasure method based on channel decontamination and a communication system. BACKGROUND

[0002] The license-free access mechanism can avoid a large amount of signaling overhead and access delay generated by devices and base stations in a random access process, provide efficient access technology support for a massive machine type communications (mMTC) network, and improve communication efficiency. Under the license-free access mechanism, devices can directly send data to the base station without waiting for authorization from the base station. Since this mechanism lacks identity verification of access devices, and only distinguishes users by identifying received pilot sequences, the mMTC network is vulnerable to pilot contamination attacks (PCAs). PCA refers to an attacker impersonating the identity of a legitimate user (LU) by sending the same pilot sequence as the LU, thereby accessing the base station, and contaminating the channel information of the LU with noise, thereby deteriorating the accuracy of the base station's activity detection, channel estimation, and subsequent data reception of the legitimate user. How to design a license-free mMTC network PCA countermeasure scheme based on channel decontamination to address this problem is a technical problem that needs to be solved.

[0003] In existing research on channel decontamination and PCA countermeasures, most of them are aimed at the PCA countermeasure scenario under the condition of single user and single attacker. For example, Chinese Patent No. CN109905203A provides a cooperative vector secure transmission scheme to resist PCA, which uses a trusted third-party node to cooperate and fuse detection to resist PCA and achieve secure transmission. However, in a license-free mMTC network, there are a large number of legitimate users and potential attackers at the same time, and the base station needs to identify the legitimate users and attackers that are in a signal sending state at this time among the numerous users, i.e., complete user activity detection, and then perform subsequent attack detection. Meanwhile, the implementation of this scheme relies on the cooperation of a trusted third-party node, which has a high requirement for the ability of the third-party node to process a large number of user signals in the mMTC network. This makes the scheme less applicable in the license-free mMTC network scenario. Therefore, it is challenging to design a PCA countermeasure scheme that can simultaneously identify legitimate user activation detection, PCA detection, and channel decontamination in a license-free mMTC network. SUMMARY

[0004] In order to solve the problems in the prior art, the application provides a large-scale access pilot contamination attack countermeasure method based on channel decontamination, which discriminates the legitimacy of an access device by utilizing the joint characteristics of a millimeter wave channel in an angle-delay-time slot domain, and decontaminates and reconstructs the channel of an attacked LU to counter the PCA.

[0005] In order to achieve the above-mentioned purpose, in a first aspect, the application provides a large-scale access pilot contamination attack countermeasure method based on channel decontamination, comprising the following steps:

[0006] A millimeter wave uplink multiple-input multiple-output system model is constructed based on an unlicensed mMTC network, the transmission characteristics of a legitimate user and an attacker in an angle-delay-time slot domain are jointly considered, a three-dimensional sparse CVR matrix H and a three-dimensional received signal Y are constructed, and a three-dimensional transmission model with sparsity in an angle-delay domain and time correlation in a time slot domain is obtained;

[0007] Based on the constructed three-dimensional transmission model, the channel decontamination problem of the pilot contamination attack can be converted into the following multi-dimensional sparse signal recovery problem:

[0008]

[0009] Wherein, t represents a time slot index, T represents the total number of time slots contained in one transmission, O {：，t} =vec(Y {：，:，t} ) represents a vectorized observation signal, Φ represents a measurement tensor, Ψ k represents a dictionary, and represent the atomic index set of the dictionary, Wherein, i1 and i2 represent the column index and row index of the dictionary Ψ1 respectively, j1 and j2 represent the column index and row index of the dictionary Ψ2 respectively, N represents the total number of users in the system, D represents the number of channel delay taps, M represents the number of base station antennas, R delay , R AoA represent the number of delays and the number of angles of the dictionary; H represents an estimated CVR matrix, H {j，t} represents the estimated value of the estimated CVR matrix at t time slot, j index;

[0010] A multi-dimensional dictionary-based sparse degree adaptive matching pursuit algorithm is used to solve the multi-dimensional sparse signal recovery problem, and an estimated result of the CVR matrix H is obtained

[0011] Based on the estimated result of the CVR matrix H The PCA attack is counteracted by combining user activity detection, PCA detection and signal decontamination criteria, and an estimated set of legitimate active users is obtained PCA attacker set Channel matrix of legitimate users after decontamination

[0012] Further, the transmission system includes a base station with M antennas, N single-antenna LUs and multiple PCA attackers; assuming that each LU generates sporadic traffic and is fixed in position, the PCA attackers can randomly launch attacks and are variable in position; the LU index set is represented as The base station antenna index set is represented as Each LU is assigned a pilot sequence s n with a length of C; assuming that multiple PCA attackers pollute the channel information of the LUs by sending the same pilot sequence as the active LUs, the pilot signal received by the base station from all active LUs and PCA attackers at time slot t is represented by the channel model Y t =G t S+G′ t S+W t , where G t and G′ t represent the channel matrices of the LUs and the attackers, respectively, S represents the pilot matrix, and W t represents noise.

[0013] Further, in the channel model Y t =G t S+G′ t S+W t , the channel matrix is represented as where the two-dimensional indicator variables and represent the active state and the attacked state of LUn, respectively, and represent the channel matrix of LUn at time slot t and the channel matrix of the attacker attacking LUn at time slot t, respectively; the channel information is composed of D channel taps, i.e. represents the channel information at tap d, and the geometric channel model is used , where L represents the number of scatterers, represents the channel gain of the lth path of LUn at time slot t, θ n，l represents the physical AoA of the lth path, a(θ n，l ) represents the antenna array response, τ n，l represents the time delay of path l, T s represents the sampling period, and f(τ) represents the sampling pulse; the CVR is introduced to describe the sparse characteristics of the millimeter wave channel, and then is represented as where B represents a virtual AoA dictionary, and f[d] represents a time delay dictionary, denotes the CVR matrix of LUn at time slot t, which has sparse characteristics in angle-delay domain; correspondingly, for an attacker who attacks LUn, the channel information denotes the CVR matrix of LUn at time slot t, which has sparse characteristics in angle-delay domain; correspondingly, for an attacker who attacks LUn, the channel information denotes the CVR matrix of LUn at time slot t, which has sparse characteristics in angle-delay domain; correspondingly, for an attacker who attacks LUn, the channel information denotes the CVR matrix of LUn at time slot t, which has sparse characteristics in angle-delay domain; correspondingly, for an attacker who attacks LUn, the channel information

[0014] Further, a three-dimensional transmission model Y composed of multiple time slots is constructed {：，：，t} = BH {：，：，t} FS+W t , t = 1, …, T, wherein Y {：，：，t} = Y t denotes the received signal at time slot t, denotes the unified CVR matrix at time slot t, wherein denotes the CVR of LUn at time slot t in the attacked, safe, and idle states, respectively, wherein denotes the set of LUs in the active and attacked states at time slot t; the active state of the LUs remains unchanged within the T time slots, and the CVR has time correlation for the user LUn in the safe state.

[0015] Further, a multi-dimensional dictionary-based sparse degree adaptive matching pursuit algorithm is used to solve the multi-dimensional sparse signal recovery problem, including: estimating the CVR matrix under a fixed sparse degree through a matching projection and a residual update step, and updating the sparse degree iteratively to obtain an estimation result closest to the real CVR matrix H wherein the matching projection step is represented as:

[0016]

[0017] wherein O res denotes the residual observation component; the residual update step is:

[0018]

[0019] wherein Γ denotes the estimated matrix support set at present, denotes the deformation matrix of the tensor Φ and the current estimation result denotes the deformation matrix of the tensor Φ and the current estimation result denotes the deformation matrix of the tensor Φ and the current estimation result is the Kronecker product of the dictionary.

[0020] Further, when updating the sparse degree iteratively, the signal recovery is performed under a fixed sparse degree , and the sparse degree is updated based on the output result of the inner layer.

[0021] Further, based on the estimation result of the CVR matrix H​ Implementing joint user activity detection, PCA detection and channel decontamination, including the following steps:

[0022] According to the estimated CVR matrix of each user The user activity detection is realized according to the power of each user The estimated LU activity indication vector is obtained:

[0023]

[0024] Wherein, The estimated active LU set is obtained according to the power threshold

[0025] The PCA detection is performed to screen the effective path set Wherein, The estimated CVR of LUn at time slot t is represented as The estimated LU attack state indication vector is obtained according to the power threshold

[0026]

[0027] The estimated attacked LU set is obtained

[0028] The channel decontamination is performed, and the time correlation of the LU activity state is utilized to decontaminate and reconstruct the CVR matrix of LUn at time slot t to obtain the decontaminated CVR matrix of LUn at time slot t:

[0029]

[0030] Wherein, Based on The decontaminated CVR matrix of all N LUs at time slot t is obtained And the channel information matrix

[0031] In a second aspect, the present application can provide a communication system, including a base station with M antennas, N single-antenna LUs and multiple PCA attackers; each LU generates sporadic traffic and is fixed in position, and the PCA attacker can randomly launch attacks and is variable in position; the LU index set is represented as The base station antenna index set is represented as Each LU is allocated a pilot sequence s n with a length of C; it is assumed that the multiple PCA attackers attack the channel information of the LU by sending the same pilot sequence as the active LU; the pilot signal received by the base station from all active LUs and PCA attackers at time slot t is represented by a channel model Y t = G t S + G' tS+W t denotes, wherein G t denotes, and G' t denote the channel matrix of LU and the attacker respectively, S denotes the pilot matrix, W t denotes noise; based on the above channel decontamination-based large-scale access pilot pollution attack countermeasure method, communication is carried out.

[0032] The application can also provide a computer device, comprising a processor and a memory, the memory being used to store a computer executable program, the processor reading the computer executable program from the memory and executing, and the processor can realize the above-mentioned channel decontamination-based large-scale access pilot pollution attack countermeasure method when executing the computer executable program.

[0033] Meanwhile, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program, and the computer program can realize the above-mentioned channel decontamination-based large-scale access pilot pollution attack countermeasure method when being executed by a processor.

[0034] Compared with the prior art, the application has at least the following beneficial effects: the application establishes a three-dimensional channel model of the angle-delay-time slot domain of the millimeter wave unlicensed mMTC network, utilizes the sparsity of the millimeter wave channel in the virtual angle domain and the delay domain and the time correlation of the LU activity; secondly, based on the established model, a PCA resistance method based on channel decontamination is proposed, which can jointly identify user activity, perform PCA detection and channel decontamination, and ensure the security of the mMTC network communication; can effectively resist large-scale access pilot pollution attacks and ensure the security of the mMTC network communication; through sparse signal processing and multi-dimensional dictionary matching, wireless resources can be more efficiently utilized, the robustness of the system to malicious attacks is enhanced through channel decontamination and PCA resistance, and the stability of the network is improved. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 The application establishes a three-dimensional channel model of the angle-delay-time slot domain of the millimeter wave unlicensed mMTC network.

[0036] Figure 2 The application establishes a three-dimensional channel model of the angle-delay-time slot domain of the millimeter wave unlicensed mMTC network.

[0037] Figure 3 The application establishes a three-dimensional channel model of the angle-delay-time slot domain of the millimeter wave unlicensed mMTC network. DETAILED DESCRIPTION

[0038] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described. Obviously, the described embodiments are some of the embodiments of the present application but not all the embodiments of the present application. Based on the embodiments of the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative effort should fall within the scope of the present application.

[0039] In embodiment 1, the present application provides a large-scale access pilot contamination attack countermeasure method based on channel decontamination, including the following steps: a millimeter wave multiple-input multiple-output system is constructed for a grant-free mMTC network, the system includes a base station with M antennas, N single-antenna LUs, and a plurality of PCA attackers. It is assumed that the LUs generate sporadic traffic and are position fixed, while the positions of the attackers are variable. Let And denote the set of LU indices and the set of base station antenna indices.

[0040] Considering a geometric channel model with L scatterers, the channel between the base station and the LUn at time slot t is denoted as:

[0041]

[0042] where denotes the channel gain of the lth path of the LUn at time slot t, θ n，l denotes the physical angle of arrival (Angle of Arrival, AoA) of the lth path, a(θ n，l ) denotes the antenna array response, τ n，l denotes the time delay of the lth path, T s denotes the sampling period, and f(τ) denotes the sampling pulse. Assuming that the positions of the scatterers remain unchanged, θ n，l also remains fixed. Since millimeter wave communication has higher path loss and blocking sensitivity, the millimeter wave channel exhibits sparsity in both the angle domain and the delay domain. The millimeter wave channel is described by introducing channel virtual representation (CVR) to map the actual physical channel to the virtual AoA domain and the delay domain. The virtual AoA dictionary and the delay dictionary are defined, where R AoA and R delay denote the number of quantized virtual AoA and delay, φ r and v r′ denote the rth discrete virtual AoA and the r' th discrete delay. The channel and its CVR are related as:

[0043]

[0044] Since there are only L paths in the channel model, when R AoA R delay >L, the CVR matrix has the coefficient property, i.e., angle-delay domain sparsity.

[0045] In a mmWave unlicensed mMTC network, each LU is assigned a pilot sequence of length C. Assume that multiple PCA attackers launch attacks by sending the same pilot sequence as the LUs. The signal received at the base station side from all active LUs and PCA attackers at time slot t is denoted as:

[0046]

[0047] where and denote the set of active LUs and attacked LUs at time slot t, denotes the channel of LUn at time slot t, denotes the channel of the attacker attacking LUn' at time slot t, denotes the pilot of LUn at D delays, s n [d] denotes the d-th shift of s n , denotes the noise, G t and G' t denote the channel matrix of all active LUs and all attackers at time slot t. Define the activity indicator variable and the attack status indicator variable where or 0 indicates that LUn is active or inactive at time slot t, or 0 indicates whether LUn is attacked by a PCA attacker at time slot t. Thus, the set of active LUs at time slot t the set of attacked LUs at time slot t the channel matrix of all active LUs at time slot t the channel matrix of all attackers

[0048] To disrupt the communication between the base station and the LUs, a PCA attacker can listen to the channel status of the LUs and launch attacks when the LUs are active. In this way, the base station estimates the channel state information of the active LUs based on the contaminated pilot signals, which severely reduces the accuracy of the channel estimation and the base station can not correctly decode the data sent by the active LUs, resulting in communication failure. Therefore, it is assumed that the attacker attacks the LUs in the active state, i.e., define the active and unattacked LUn The state of LUn is defined as "safe" state, then at each time slot, the state of LUn can be any one of "inactive", "safe" and "attacked". Moreover, as LUn generates sporadic traffic, the set of LUs The size of The channel matrix G t is sparse. t

[0049] Angle-delay-slot three-dimensional mmWave channel model

[0050] Define a unified CVR matrix to represent the CVR of LUn in different states at time slot t. The specific definition is:

[0051]

[0052] where, denotes the CVR matrix of the attacker attacking LUn at time slot t. Considering that each transmission contains consecutive T time slots, the CVR matrix of T time slots is Combining, a unified three-dimensional CVR matrix is obtained, where The three-dimensional matrix H n has sparsity in both the virtual angle domain and the delay domain, as shown in Figure 1 In addition, in the actual mMTC network, the user activity may exhibit a correlation feature, that is, the active user will transmit in the adjacent time slot with high probability. Therefore, the time correlation of user activity is considered by assuming that the activity of each LU remains unchanged in consecutive T time slots, that is, Since the positions of all LUs and scatterers are fixed, the physical AoA and virtual AoA of LUs are also unchanged. In addition, it is observed that the path delay of the mmWave channel changes much slower than the path gain, that is, even if the path gain changes greatly between different time slots, the path delay τ n，l still remains unchanged between consecutive time slots. Therefore, when LUn is in a safe state, that is, as LUn2 in Figure 1 , the CVR matrix of LUn within T time slots has the same support set, that is,

[0053]

[0054] where, denotes ​support set, i.e., the set of non-zero element indices. However, since the attacker initiates PCA randomly at each time slot and the attacker’s location is variable, the attacked LUs do not show time correlation. That is, for the LUs in the attacked state (As Figure 1 in LUn3), varies with t.

[0055] Based on the unified CVR matrix, the received signal Y t can be expressed as:

[0056] Y t = BH t FS+W t

[0057] where, denotes the delay dictionary matrix, combining the received signals Y t of T time slots, we obtain the three-dimensional received signal where Y {：，：，t} =Y t . And the single-time-slot transmission model is converted to the following three-dimensional transmission model:

[0058] Y {：，：，t} = BH {：，：，t} FS+W t , t = 1, …, T

[0059] where H {：，：，t} = H t . Based on the constructed three-dimensional transmission model, the PCA countermeasure problem can be converted to the recovery problem of the sparse CVR matrix . In order to resist PCA, the channel of the attacked LU needs to be decontaminated and reconstructed, i.e., for the attacked Since at this time the attacker’s needs to be eliminated and the real is recovered. Most of the existing schemes only utilize the sparsity of the millimeter wave channel in the virtual angle domain, but there is a high probability that the CVR of the attacker and the attacked LU will overlap in the virtual angle domain, i.e., and cannot be completely separated. This will seriously deteriorate the accuracy of channel reconstruction. Therefore, the present application expands the sparsity of the millimeter wave channel in the angle-delay-time slot domain, as shown in the figure, to enhance the sparse structure of , which will significantly reduce the probability of CVR overlap between the attacker and the LU, and effectively improve the performance of channel reconstruction.

[0060] Construct a multi-dimensional sparse signal recovery problem

[0061] According to the constructed three-dimensional transmission model, the three-dimensional sparse matrix H can be represented by two independent dictionaries, i.e., a virtual AoA dictionary B and a delay dictionary F, and the pilot matrix S serves as a measurement matrix. Therefore, the PCA adversarial problem can be represented as a multidimensional sparse recovery problem based on multidimensional dictionaries. Specifically, define a dictionary set {Ψ k} k＝1，2 where The observation matrix is defined as where O {：，t} = vec(Y {：，：，t} ), and vec(·) represents column vectorization of a matrix. In the multidimensional case, the measurement matrix can be constructed in the form of a tensor, denoted as To obtain Φ, first define the matrix and have Define an index set and represent the row index and column index of the dictionary atoms in {Ψ k} k＝1，2 . Further, the multidimensional sparse signal recovery problem can be constructed as:

[0062]

[0063] When the unified CVR matrix H is recovered by the above formula, the real CVR X of all LUs can be reconstructed by eliminating the CVRX' composed of the channel components of all attackers to realize channel decontamination.

[0064] A multidimensional dictionary based sparsity adaptive matching pursuit algorithm is used to solve the multidimensional sparse signal recovery problem. To solve the constructed multidimensional sparse signal recovery problem, a multidimensional dictionary based sparsity adaptive matching pursuit (MD-SAMP) algorithm is designed. The algorithm is designed based on a multidimensional orthogonal matching pursuit (MOMP) algorithm, which can effectively utilize the multidimensional sparse structure of the CVR matrix H to solve the problem. The main process of the MD-SAMP algorithm consists of two iterative parts: matching projection and residual update. In each iteration, the algorithm calculates the inner product of the dictionary atoms and the residual observation components, i.e., projection, to select the atom with the maximum projection value. Then, the residual projection component is updated based on the selected atom projection.

[0065] Specifically, the matching projection step of the constructed multidimensional problem is:

[0066]

[0067] where O res denotes the residual observation component. This maximization step requires the computation of all possible indices of the dictionary atoms This will cause a quite high computational complexity. To reduce the complexity, the optimization of index j = (j1, j2) can be transformed into an iterative optimization procedure based on the history solution , where j1and j2are optimized separately. In this way, the optimization problem can be transformed into

[0068]

[0069] where k, k' = 1, 2, k ≠ k'. After the completion of the matching pursuit, the selected atom indices j = (j1, j2) are added to the estimated support set Γ of H. Then, the residual observation component can be updated according to

[0070]

[0071] where is the estimation of H, which can be obtained by least square estimation, i.e. denotes the matrix inversion, and denotes the modification of Φ and , respectively, denotes the Kronecker product of the dictionary.

[0072] In the MOMP algorithm, the above iterative procedure is repeated until the size of the estimated support set Γ equals to the sparsity of the matrix H, i.e., the size of the support set supp(H). However, in real scenarios, the size of supp(H), i.e., the number of all active LUs and attacker's channel paths, cannot be perfectly known, which limits the practical application of the MOMP algorithm. To solve this problem, the designed MD-SAMP algorithm performs two nested iterative procedures to solve the sparse signal recovery problem by an adaptive method to determine the sparsity, and the algorithm procedure is shown in Table 1. The inner loop performs the signal recovery under a fixed sparsity , and the outer loop updates the sparsity based on the output of the inner loop.

[0073] Let l denote the iteration index of the inner loop. In the l-th inner loop, first, the support set estimation is performed, see the fifth row of Table 1, by matching pursuit, selecting atom indices from the index set , obtaining the set Γ l , and adding the result of the (l-1)-th iteration Γ l-1 to Γ lCombined, we obtain set Λ. Next, we perform a support set correction step, as shown in row 6 of Table 1, selecting s indices from Λ to maximize their corresponding projection values, thus forming set Λ. Where Ξ(·,s) represents selecting the largest parameter. A set of indices for elements. Based on The signal estimation and residual update steps are shown in rows 7 and 8 of Table 1. Afterwards, the index is updated iteratively, starting the (l+1)th inner loop.

[0074] When termination criteria After satisfying the condition, the sparsity is obtained. The support set below and estimated CVR matrix Next, the sparsity update step is performed in the outer loop, as shown in row 11 of Table 1, and the sparsity is estimated. Signal estimation is performed. Finally, the termination criterion is applied. Once satisfied, this means that the sparsity has been overestimated, where This represents the noise floor power. Therefore, the entire iteration process is now complete, and the reconstructed CVR matrix is ​​obtained.

[0075] Table 1 MD-SAMP Algorithm

[0076]

[0077] Combined user activity detection, PCA detection and channel decontamination

[0078] Output results based on the MD-SAMP algorithm The PCA adversarial problem can be solved by combining user activity detection, PCA detection, and channel destaining. First, user activity detection is performed, based on the reconstructed data from LUN. From the power, we obtain the estimated activity indicator variable:

[0079]

[0080] in, This represents the CVR power threshold for active LUs. Next, the estimated set of active LUs can be obtained. Secondly, PCA detection is performed. Since each LU's real channel contains only L paths, if an attacker attacks LUn in time slot t, the number of paths will be greater than L. Therefore, the effective path set is first filtered out based on path power. in Let represent the power threshold of a single path in the CVR. Then, the estimated attack state indicator variable is determined by the following formula:

[0081]

[0082] where |·| denotes the number of elements in a set, the estimated attacked LU set can be obtained Finally, channel decontamination is performed. Since the CVR of the active LUs has time correlation, while the CVR of the attacker does not satisfy this property, the contaminated coefficients in the CVR of the attacked LUs can be distinguished by comparing the path sets in T time slots Specifically, first, the path intersection of LUn in T time slots is obtained Second, L paths with the largest power are selected from Y n as the legitimate paths of the LUn. For the remaining paths, i.e. they are regarded as contaminated paths caused by the attacker and need to be eliminated. Thus, the channel decontamination of LUn can be expressed as:

[0083]

[0084] where denotes the CVR matrix of LUn after decontamination in time slot t. Based on the decontaminated channel coefficients of all LUs in T time slots can be obtained Further, the estimated channel coefficients of the LUs can be obtained by .

[0085] The system performance of the proposed scheme is evaluated by simulation, and the system parameter settings are shown in Table 2:

[0086] Table 2 Simulation parameters

[0087]

[0088]

[0089] To measure the performance of the designed scheme, comparison experiments are conducted with the following compressive sensing algorithms: orthogonal matching pursuit (OMP) algorithm, block sparsity adaptive subspace pursuit (BSASP) algorithm, MOMP algorithm, and Oracle-LS algorithm. Among them, the OMP and BSASP algorithms are classical greedy pursuit algorithms for single-dimensional sparse signal recovery, which only recover the CVR matrix in the angle domain; the MOMP algorithm and the designed MD-SAMP algorithm both recover the CVR in the angle-delay-time slot domain, but the MOMP algorithm can only be implemented on the premise of known sparsity; the Oracle-LS algorithm provides an upper bound for the performance of other algorithms, which assumes that the true active LU set and the attacked LU set, and the path positions of all users are perfectly known by the base station. In addition, to evaluate the impact of PCA on channel estimation performance and data reception performance, the experimental results of the designed scheme without PCA detection and channel decontamination are given.

[0090] Figure 2 、 Figure 3 The performance simulation results of the normalized mean square error of the CVR matrix recovery of all LUs and the bit error rate of the data transmitted by the active LUs decoded by the base station are shown, which varies with the signal-to-noise ratio. The normalized mean square error is defined as where X and represent the true CVR and the estimated CVR of all LUs in T time slots; the bit error rate is defined as where b n，t and represent the data bits transmitted by the LUn and the bits decoded by the base station. From Figure 2 , Figure 3It can be seen that the performance of all schemes is improved with the increase of the signal-to-noise ratio, and the designed MD-SAMP algorithm is significantly better than the OMP and BSASP algorithms, and is very close to the MOMP algorithm. Since the OMP algorithm and the BSASP algorithm only reconstruct the CVR matrix in the angle domain, under a limited number of discretized virtual AoA, the CVR of the attacker and the attacked LU will be highly overlapped, which makes it difficult to eliminate the contaminated coefficients in the CVR of the attacked LU, thereby seriously deteriorating the performance of the receiver in channel estimation and data decoding. On the other hand, the MOMP and the designed MD-SAMP algorithm reconstruct the CVR matrix in the angle-delay-time slot domain, and in the extended channel domain, the sparsity of the CVR of each transmitter is significantly enhanced, thereby significantly reducing the overlap probability of the CVR of each LU and the attacker, thereby improving the channel reconstruction and data decoding performance. However, the MOMP algorithm assumes that the true sparsity of the CVR is perfectly known by the base station, which cannot be realized in practice. In order to solve this problem, the designed MD-SAMP algorithm realizes signal recovery through sparsity adaptive adjustment, and can achieve a performance close to that of the MOMP algorithm. In addition, by comparing with the purple line which does not consider PCA detection and channel decontamination, it can be seen that PCA can seriously reduce the performance of channel estimation and data decoding. Since the PCA detection and channel decontamination criteria are designed in the application, the proposed scheme can effectively resist PCA and improve the system performance. However, the performance of the Oracle-LS algorithm is better than that of the proposed MD-SAMP algorithm, because the Oracle-LS algorithm assumes that the number and position of the paths in the CVR matrix are known, which is impossible in practice.

[0091] In summary, the application proposes a large-scale access pilot contamination attack countermeasure method based on channel decontamination, which fully utilizes the sparsity and correlation of the millimeter wave channel of the legitimate user and the attacker in the angle-delay-time slot domain, and obtains higher channel recovery and data decoding performance. Simulation results show that compared with existing PCA countermeasures, the PCA countermeasure method based on channel decontamination of the application effectively reduces the mean square error of channel estimation, improves the accuracy of data decoding, and has better system performance.

[0092] In embodiment 2, the application provides a communication system, which includes a base station with M antennas, N single-antenna LUs, and a plurality of PCA attackers; each LU generates sporadic traffic and has a fixed position, and the PCA attacker can randomly launch an attack and has a variable position; the LU index set is represented as The base station antenna index set is represented as Each LU is allocated a pilot sequence s n, length is C; it is assumed that multiple PCA attackers pollute the channel information of the LU by sending the same pilot sequence as the active LU; the pilot signal received by the base station from all active LUs and PCA attackers at time slot t is represented by channel model Y t = G t S+G′ t S+W t , wherein G t and G′ t respectively represent the channel matrix of the LU and the attacker, S represents the pilot matrix, W t represents noise; and communication is performed based on the above-mentioned channel decontamination-based large-scale access pilot pollution attack countermeasure method.

[0093] In addition, the present application can also provide a computer device comprising a processor and a memory, the memory being used to store a computer executable program, the processor reading part or all of the computer executable program from the memory and executing, the processor being capable of realizing the channel decontamination-based large-scale access pilot pollution attack countermeasure method of the present application when executing part or all of the computer executable program.

[0094] In another aspect, the present application provides a computer readable storage medium, the computer readable storage medium storing a computer program, the computer program being capable of realizing the channel decontamination-based large-scale access pilot pollution attack countermeasure method of the present application when executed by a processor.

[0095] The computer device can be a notebook computer, a desktop computer or a workstation.

[0096] The processor of the present application can be a central processing unit (CPU), a digital signal processor (DSP), an application specific integrated circuit (ASIC) or a ready-to-program field programmable gate array (FPGA).

[0097] The memory of the present application can be an internal storage unit of a notebook computer, a desktop computer or a workstation, such as a memory, a hard disk; or an external storage unit, such as a mobile hard disk, a flash card.

[0098] The computer-readable storage medium can include computer storage media and communication media. The computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. The computer-readable storage medium can include read-only memory (ROM), random access memory (RAM), solid state disk (SSD), optical disk, etc. Among them, the random access memory can include resistance random access memory (ReRAM) and dynamic random access memory (DRAM).

Claims

1. A channel decontamination-based method for combating large-scale access pilot contamination attack, characterized in that, The method comprises the following steps: Based on the unlicensed mMTC network, a millimeter wave uplink multiple-input multiple-output system model is constructed, and the transmission characteristics of the legitimate users and attackers in the angle-delay-time slot domain are jointly considered to construct a three-dimensional sparse CVR matrix and a three-dimensional received signal , obtaining a three-dimensional transmission model with sparsity in the angle-delay domain and time correlation in the time slot domain; Based on the constructed three-dimensional transmission model, the channel decontamination problem of the pilot pollution attack can be converted into the following multi-dimensional sparse signal recovery problem: wherein, denotes the time slot index, denotes the total number of time slots contained in one transmission, denotes the vectorized observation signal, denotes the measurement tensor, denotes the dictionary, and denotes the atom index set of the dictionary, , wherein denotes the column index and the row index of the dictionary respectively, denotes the column index and the row index of the dictionary respectively, denotes the total number of users in the system, denotes the number of channel taps, denotes the number of base station antennas, , denotes the number of taps and the number of angles of the dictionary; denotes the estimated CVR matrix, denotes the value of the estimated CVR matrix at time slot, index. The multi-dimensional sparse signal recovery problem is solved by a sparsity adaptive matching pursuit algorithm based on a multi-dimensional dictionary to obtain an estimation result of a CVR matrix ; specifically comprising: estimating the CVR matrix under a fixed sparsity by a matching projection and residual update step, and updating the sparsity adaptively by iteration to obtain an estimation result closest to the true CVR matrix ; wherein the matching projection step is represented as:​​ wherein denotes the residual observation component; the residual update step is wherein denotes the current estimated matrix support set, and denotes the tensor and the current estimate of the deformation matrix, is the Kronecker product of the dictionary; Based on CVR matrix estimation results By combining user activity detection, PCA detection, and Cinda's decontamination criteria, PCA countermeasures are achieved, resulting in an estimated set of legitimate active users. PCA attacker set With the decontamination and reconstruction of the legitimate user channel matrix .

2. The channel decontamination based massive access-pilot contamination attack countermeasure method according to claim 1, characterized in that, The transmission system includes a base station with a root antenna, a single antenna LU and multiple PCA attackers; assuming each LU generates sporadic traffic and is location fixed, PCA attackers can randomly launch attacks and are location variable; the LU index set is denoted as , the base station antenna index set is denoted as ; each LU is assigned a pilot sequence with length ; assuming multiple PCA attackers pollute the channel information of LUs by sending the same pilot sequence as the active LUs; at time slot , the pilot signals received by the base station from all active LUs and PCA attackers are denoted by the channel model , where and denote the channel matrices of LUs and attackers, respectively, denotes the pilot matrix, denotes the noise.

3. The channel decontamination based massive access-pilot contamination attack countermeasure method according to claim 2, characterized in that, The channel model In the channel matrix is denoted as , where the two-dimensional indicator variable represents the activation state and the attacked state of the LU respectively, represents the activation state and the attacked state of the LU respectively, The channel matrix of the LU in the time slot , and the channel matrix of the attacker attacking the LU ; the channel information is composed of channel taps, that is , represents the channel information at the tap , and the geometric channel model is denoted as , where represents the number of scatterers, represents the channel gain of the LU in the time slot the first path, represents the physical AoA of the first path, represents the antenna array response, represents the time delay of the path , represents the sampling period, represents the sampling pulse; the sparse characteristics of the millimeter wave channel are described by introducing the CVR, and then is denoted as , where represents the virtual AoA dictionary, represents the time delay dictionary, represents the CVR matrix of the LU in the time slot , and the CVR matrix has sparse characteristics in the angle-delay domain; accordingly, for the attacker attacking the LU , the channel information is denoted as , where, represents the CVR matrix of the attacker attacking the LU in the time slot .​ 4. The method of claim 2, wherein, The three-dimensional transmission model of multi-time slot is constructed as: wherein represents the received signal of time slot , represents the unified CVR matrix of time slot , wherein represents the LU , CVR in time slot is in the attacked, safe, idle state respectively, wherein represents the LU set in time slot is in the active and attacked state; the active state of the LU is kept unchanged in time slots, and the CVR has time correlation for the user LU in the safe state.

5. The channel decontamination based massive access-pilot contamination attack countermeasure method of claim 1, wherein, By iteratively updating the sparsity while fixing the sparsity The signal recovery is performed on the inner layer output result The sparsity is updated.

6. The channel decontamination based massive access-pilot contamination attack countermeasure method of claim 1, wherein, Based on the estimation results of the CVR matrix Implementing joint user activity detection, PCA detection and signal-to-noise decontamination, including the following steps:​ According to the estimated CVR matrix of each user Power achieves user activity detection, resulting in an estimated LU activity indicator vector: wherein, represents a power threshold, resulting in an estimated active LU set ; Perform PCA testing to filter the set of valid paths. ,in, Represents the estimated LU In the time slot CVR at the location, Representing the power threshold, we obtain the estimated LU attack state indication vector: and obtaining an estimated set of attacked LUs ; Channel decontamination is performed, using the time correlation of the LU activity state, to the LU At Decontamination reconstruction is performed on the CVR matrix of the time slot to obtain the LU At the time slot CVR matrix after decontamination: wherein, , based on obtaining time slots all CVR matrices of the LUs after decontamination and channel information matrices .

7. A communication system, characterized by A method for transmitting a signal from a base station to a mobile station includes a base station having a plurality of antennas, a single antenna LU and a plurality of PCA attackers; each LU generates sporadic traffic and is fixed in location, and the PCA attackers can randomly launch attacks and are variable in location; a LU index set is represented as , and a base station antenna index set is represented as ; each LU is assigned a pilot sequence , having a length of ; it is assumed that the plurality of PCA attackers launch attacks by sending the same pilot sequence as the active LUs, which pollutes the channel information of the LUs; at a time slot , the pilot signals received by the base station from all active LUs and PCA attackers are represented by a channel model , where and represent the channel matrices of the LUs and the attackers, respectively, represents a pilot matrix, and represents noise. The method for communication based on the channel decontamination of the large-scale access pilot pollution attack according to any one of claims 1-5.

8. A computer device, comprising: The method comprises a processor and a memory, the memory is used for storing a computer executable program, the processor reads the computer executable program from the memory and executes, and the processor can realize the method for communication based on the channel decontamination of the large-scale access pilot pollution attack according to any one of claims 1-6 when executing the computer executable program.

9. A computer-readable storage medium, characterized in that, A computer readable storage medium stores a computer program, and the computer program can realize the method for communication based on the channel decontamination of the large-scale access pilot pollution attack according to any one of claims 1-6 when executed by a processor.

Images