A digital communication data security transmission system

Abstract

The application discloses a kind of digital communication data security transmission systems, belong to data transmission technical field, comprising: acquisition module, for obtaining the data to be transmitted of communication parties and instantaneous channel matrix;Determination module, for ensuring the channel random state of communication parties is discernible as target, determine the multi-order free cumulant about instantaneous channel matrix;Reconstruction module, for exchanging the multi-order free cumulant of communication parties respectively, reconstruct to obtain the channel positive definite matrix consistent with communication parties;Generation module, for generating session key based on channel positive definite matrix;Transmission module, for encrypting transmission to the data to be transmitted by session key.It avoids the security risks caused by traditional static parameters, and can fully utilize the randomness existing in communication environment to enhance the key strength, while improving the consistency of the keys of the two communication parties, thereby effectively improving the confidentiality, integrity and reliability in the process of data transmission.

Description

Technical Field

[0001] This invention relates to the field of data transmission technology, and in particular to a secure digital communication data transmission system. Background Technology

[0002] Digital communication data refers to information data represented, stored, and transmitted in binary digits (0s and 1s), including text, voice, images, video, control commands, sensor-collected information, and various types of business data. This type of data is reliably transmitted between communicating parties via wired or wireless networks using technologies such as digital encoding, modulation and demodulation, and communication protocols, and is the fundamental information carrier in modern communication systems.

[0003] Digital communication data transmission is a fundamental component of modern information network operation, and its transmission quality directly impacts the real-time performance, accuracy, and reliability of information exchange. In scenarios such as industrial control, the Internet of Things (IoT), financial transactions, telemedicine, intelligent transportation, and defense communications, numerous critical business operations rely on the secure transmission of digital communication data. Data loss, tampering, leakage, or delays during transmission can not only lead to system malfunctions but also potentially cause economic losses, privacy breaches, and even security incidents. Therefore, secure and reliable digital communication data transmission technology is of paramount importance for ensuring information integrity, improving communication efficiency, and maintaining cyberspace security.

[0004] However, in the existing process of secure transmission of digital communication data, key generation and distribution usually rely on static parameters. Key patterns are easily analyzed and predicted by attackers, resulting in insufficient security. Furthermore, it is difficult to cope with channel randomness. Due to the lack of dynamic utilization of the instantaneous channel state of both communicating parties, the encryption process may experience synchronization failure or lack of consistency, which seriously affects the confidentiality and integrity of data transmission. Summary of the Invention

[0005] To address the technical problems in existing digital communication data security transmission processes, where key generation and distribution typically rely on static parameters, key patterns are easily analyzed and predicted by attackers, leading to insufficient security and difficulty in handling channel randomness, and the lack of dynamic utilization of the instantaneous channel state of both communicating parties, resulting in synchronization failures or inconsistencies in the encryption process, seriously affecting the confidentiality and integrity of data transmission, this invention provides a digital communication data security transmission system.

[0006] This invention provides a secure digital communication data transmission system, comprising: The acquisition module is used to acquire the data to be transmitted and the instantaneous channel matrix of both communicating parties; The determination module is used to determine the multi-order free accumulator of the instantaneous channel matrix with the goal of ensuring that the random channel states of both communicating parties are identifiable. The reconstruction module is used to exchange the multi-order free cumulative quantities of the two communicating parties and reconstruct a consistent positive definite channel matrix for both parties. The generation module is used to generate session keys based on the positive definite channel matrix; The transmission module is used to encrypt and transmit data to be transmitted using a session key.

[0007] The beneficial effects of the technical solutions provided by the embodiments of the present invention include at least the following: In this embodiment of the invention, the communicating parties first acquire their respective observed instantaneous channel matrices, which characterize the propagation characteristics and random changes of the current communication link at a specific moment. Subsequently, multi-order free cumulants are extracted from the instantaneous channel matrices. These multi-order free cumulants can extract stable and discriminative statistical structural features from complex random matrices, thus accurately characterizing the random state of the channel. After exchanging their respective statistical features, the communicating parties obtain a consistent positive definite channel matrix through a reconstruction process. The positive definite channel matrix is ​​a unified mathematical expression of the shared channel features of both parties, preserving the intrinsic channel information while possessing good stability and uniqueness. Based on this, a session key is generated. Since the session key originates from the shared and real-time changing channel features of both parties, it possesses strong randomness, unpredictability, and dynamic update capability. Finally, the session key is used to encrypt the data to be transmitted, making it difficult to recover even if the data content is intercepted. This not only avoids the security risks associated with traditional static parameters but also fully utilizes the natural randomness inherent in the communication environment to enhance key strength and improve the consistency of keys between the communicating parties, thereby effectively improving the confidentiality, integrity, and reliability of data transmission. Attached Figure Description

[0008] The preferred embodiments will now be described in a clear and easy-to-understand manner, in conjunction with the accompanying drawings, to further explain the above-mentioned characteristics, technical features, advantages, and implementation methods of the present invention.

[0009] Figure 1 This is a schematic diagram of the structure of a digital communication data security transmission system provided by the present invention. Detailed Implementation

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the specific implementation methods of the present invention will be described below with reference to the accompanying drawings. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings and other implementation methods can be obtained based on these drawings without any creative effort.

[0011] To keep the drawings concise, each figure only schematically shows the parts relevant to the invention, and these do not represent the actual structure of the product. Furthermore, to facilitate understanding, in some figures, only one of components with the same structure or function is schematically depicted, or only one is labeled. In this document, "one" not only means "only one," but can also mean "more than one."

[0012] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0013] In this document, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections. They can refer to mechanical connections or electrical connections. They can refer to direct connections or indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0014] Furthermore, in the description of this invention, the terms "first," "second," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance.

[0015] In one embodiment, refer to Figure 1 The diagram shows a structural schematic of a secure digital communication data transmission system provided by the present invention.

[0016] This invention provides a secure digital communication data transmission system, comprising: Acquisition module 1 is used to acquire the data to be transmitted and the instantaneous channel matrix of both communicating parties.

[0017] The instantaneous channel matrix is ​​a mathematical matrix used to describe the propagation characteristics of a wireless communication link at a certain moment. It reflects information such as fading, reflection, scattering, multipath effects, and noise during signal propagation between the transmitter and receiver, and is a digital representation of the current communication environment. By obtaining the data to be transmitted from both communicating parties and the corresponding instantaneous channel matrix, not only can the information content that needs to be protected be clearly identified, but the random characteristics of the current communication environment can also be obtained simultaneously.

[0018] In one possible implementation, the two parties to the communication include a sender and a receiver.

[0019] In one possible implementation, the acquisition module 1 is specifically used for: S101: Instruct both communicating parties to alternately send orthogonal pilot sequences; Specifically, in time-division duplex communication mode, the two communicating parties alternately send orthogonal pilot sequences within the same channel coherent time window.

[0020] S102: Collect the received pilot signals from both communicating parties; S103: Perform least squares estimation on the received pilot signal of either of the communicating parties to obtain the instantaneous channel matrix.

[0021] Understandably, this process can be implemented in time-division duplex mode, where the sender and receiver alternately transmit pre-designed orthogonal pilot sequences while simultaneously acquiring the pilot signals received by the other party. Subsequently, a least-squares estimation method is used to match the received signals with the known pilot sequences, thereby obtaining the instantaneous channel matrix describing the channel gain and phase changes. The orthogonal pilot sequences can excite each channel dimension independently without interference, enabling the receiver to accurately estimate the channel response. This allows for rapid and accurate capture of the channel's instantaneous characteristics, ensuring that both communicating parties obtain consistent channel information.

[0022] More specifically, the process of obtaining the instantaneous channel matrix through least squares estimation is as follows: A linear relationship model is established between the pilot signals acquired at the receiver and the known orthogonal pilot sequences from the transmitter. By minimizing the squared error between the received and predicted signals, the estimated value of the channel matrix is ​​obtained. The least squares criterion is then used to obtain the optimal linear unbiased estimate in the presence of noise, minimizing the estimation error. This process is computationally simple, stable, and reliable, and can quickly and accurately reflect the instantaneous characteristics of the channel, providing precise channel information for key generation or adaptive transmission.

[0023] Module 2 is used to determine the multi-order free accumulator of the instantaneous channel matrix with the goal of ensuring that the random channel states of both communicating parties are identifiable. Among them, multi-order free cumulants are statistical characteristic quantities in free probability theory, used to describe the high-order statistical structure and random distribution characteristics within a random matrix. Multi-order free cumulants can more comprehensively reflect the correlations between different orders within a random matrix, thus extracting intrinsic feature information with strong discriminative power and stability. By calculating multi-order free cumulants on the instantaneous channel matrix, core statistical features characterizing the current random state of the channel can be extracted from complex and constantly changing channel information, reducing the dimensionality of the original data while retaining key structural information.

[0024] Multi-order free accumulator, i.e., the previous k The free cumulant is equivalent to the identity feature points of the current channel. It's not necessary to store all features of the entire channel matrix; only the features needed are required. k The most distinctive and reliable feature can uniquely identify the random state of the current channel and be used to generate a consistent key between the two parties.

[0025] In one possible implementation, the determining module 2 is specifically used for: S201: Based on the instantaneous channel matrix, establish the sample covariance matrix of either of the communicating parties; The formulas for the sample covariance matrices of the two communicating parties are as follows: ; in, Represents the instantaneous channel matrix C The corresponding sample covariance matrix, This indicates the conjugate transpose. Indicates the number of pilot symbols; The number of pilot symbols is a publicly known parameter agreed upon by both parties beforehand. For example, if both parties agree to send 32 pilots each time for estimation, then... L =32, which remains constant throughout. The number of pilot symbols is the total number of pilot symbols continuously transmitted by one party each time the two parties estimate the channel; that is, the total number of training sampling points used for channel estimation.

[0026] S202: Calculate the sample moments of each order of the sample covariance matrix; In step S202, the sample covariance matrix refers to the sample covariance matrix of either of the communicating parties.

[0027] r The formula for calculating the first-order sample moments is: ; in, Represents the sample covariance matrix of r Order sample moments, Represents the sample covariance matrix of r Matrix powers This indicates the number of antennas at the transmitting end, which is the number of antennas at the other end of the communication. Represents the matrix trace operation; S203: Based on the sample moments of each order, and combined with the non-cross-partitioning method, determine the candidate multi-order free cumulants of the corresponding sample covariance matrix; The specific formula for calculating candidate multi-order free cumulants is as follows: ; ; in, Indicates candidate r First-order free cumulative quantity, Represents the set (1,2,..., r All non-crossing partition sets in ) This represents a non-crossing partition state in a set of non-crossing partitions. express One of the groups, express The total number of groups in the middle, This indicates the number of elements in group B. Indicates the multiplication symbol. express r The maximum element partition in non-crossing partitions of order 1. Indicates about and The Möbius function, Represents the factorial symbol; Non-crossing partitions refer to sets (1, 2, ..., ...) r The elements in the range are grouped so that no two groups overlap. (For example, if r=4, dividing them into {1,3} and {2,4} would be overlapping and would not meet the requirement; dividing them into {1} and {2,3,4} would be non-overlapping and would meet the requirement). {1} and {2,3,4} are each a subgroup in this state. r The maximum element partition in a non-crossing partition refers to the partition of the set (1, 2, ..., ...). r (It is not split; the whole thing is treated as a group.)

[0028] It should be noted that in random matrix spectral analysis, sample moments directly measure the global statistical properties of the eigenvalue distribution, but their higher-order moments are extremely sensitive to noise and highly correlated with each other, making them difficult to directly serve as robust channel features. This step introduces a non-cross-partitioning combinatorial structure to transform arbitrary-order sample moments into free cumulants. Non-cross-partitioning requires dividing the index set into several groups such that any two groups do not overlap on the annulus, and their counting follows the Catalan number rule, filtering out cross-correlation terms caused by noise. Inversion using the Möbius function on the non-cross-partitioning lattice is essentially discorrelated and redundancy-removing of the multiple mixtures of moments. The Möbius function assigns alternating positive and negative weights according to the coarsening of the partition, precisely canceling out the contamination of higher-order moments by lower-order moments, and peeling away the independent spectral components hidden in the moment sequence one by one. Since the free cumulants are strictly additable for the spectral superposition of independent signal sources, and exhibit rapid decay for the complex conjugate cross terms of noise, the extracted free cumulants of each order can tightly characterize the essential structure of the signal subspace. At the same time, they greatly suppress random disturbances caused by asymmetric hardware and receiver thermal noise, providing an objective and highly discriminative numerical basis for subsequent key negotiation based on analytical spectrum reconstruction.

[0029] S204: Verify order by order, and output the maximum order that satisfies the inequality of random matrix theory as the cutoff order; The inequality is as follows: ; in, This indicates the pre-calibrated receiver thermal noise power for each of the communicating parties. express The variance; It should be noted that, based on the universal upper bound of the variance of free cumulants estimated in random matrix theory, a statistical reliability test is performed on the candidate cumulants with progressively increasing order. This inequality precisely correlates the limit of the estimated variance with the receiver thermal noise power, pilot sampling size, and combinational structure. When the theoretical upper bound of a certain order cumulant makes its variance comparable to the noise floor, it is determined that the cumulant of that order can no longer be effectively identified from random disturbances. This automatic truncation ensures that all used cumulants are within the statistically significant range, maximizing the preservation of independent feature dimensions of the channel spectrum while eliminating the arbitrariness of manually set thresholds, making the entire feature extraction process a data-driven and reproducible workflow.

[0030] S205: Output the candidate multi-order free accumulators that are less than or equal to the truncation order as multi-order free accumulators.

[0031] Specifically, this process begins by constructing a sample covariance matrix based on the instantaneous channel matrix. The instantaneous channel matrix describes the propagation characteristics and random changes of the current communication link, while the sample covariance matrix characterizes the correlations between various dimensions of the channel, transforming the original channel information into a matrix form suitable for statistical analysis. Subsequently, matrix exponentiation and trace operations are performed on the sample covariance matrix to obtain sample moments of various orders. These sample moments characterize the overall statistical regularity of the matrix eigenvalue distribution, reflecting the internal structural characteristics of the channel from different perspectives. Based on this, the non-cross-partitioning method from free probability theory is used to combine the sample moments of various orders. Non-cross-partitioning is a grouping method that ensures no cross-nesting between groups, effectively separating the coupling effects between different statistical orders. The Möbius function is then used to recursively correct the statistical contributions of each order, resulting in candidate free cumulants. Since free cumulants directly reflect the intrinsic statistical structure within the random matrix, they possess stronger stability and discriminative power compared to the original matrix data, and can be considered as the identity characteristics of the current channel's random state. Furthermore, a variance constraint condition for the free cumulant is established using random matrix theory. The variance measures the degree of fluctuation of statistical features due to noise. By verifying and selecting the highest effective order that satisfies the constraint condition, higher-order information that is heavily affected by noise or lacks sufficient statistical reliability can be automatically eliminated, thus determining the optimal truncation range. Finally, the free cumulant within the truncation range is output as the channel feature, ensuring that the retained statistical features contain sufficiently rich intrinsic channel information while avoiding the introduction of redundant data and noise. In this way, the current channel state can be uniquely characterized using a small number of highly reliable statistical features without needing to store the complete channel matrix. This significantly reduces data processing and storage overhead and improves the consistency and stability of channel feature extraction by both communicating parties.

[0032] Reconstruction module 3 is used to exchange the multi-order free cumulative quantities of the two communicating parties and reconstruct a consistent positive definite channel matrix for both parties. The positive definite channel matrix is ​​a mathematically reconstructed matrix used to represent the channel characteristics shared by both communicating parties. It satisfies the positive definiteness condition, meaning all eigenvalues ​​are positive, ensuring the matrix is ​​invertible and has stable numerical properties, making it a reliable basis for key generation. By exchanging the multi-order free cumulants of both parties and reconstructing the matrix, a consistent positive definite channel matrix can be obtained. This matrix retains the key characteristics of the shared channel and possesses stability and uniqueness.

[0033] In one possible implementation, the reconstruction module 3 is specifically used for: S301: Obtain the truncation order of both communicating parties and output the minimum truncation order as the retained truncation order; S302: Obtain the multi-order free cumulative quantities of both communicating parties corresponding to the retained truncation order, and exchange them; Alternatively, the exchange can be performed via a public, noiseless control channel.

[0034] S303: Calculate the mean of free cumulants of the same order in the retained cutoff order; The specific formula for calculating the mean of free cumulants of the same order is as follows: ; in, Indicates the order of truncation to be retained. v The corresponding mean of free cumulants of the same order, and They represent the sender in the communication between the two parties. v Free accumulator and receiver v Order free cumulative quantity; It should be noted that this step involves arithmetic averaging of the free cumulants of the same order held by each communicating party, generating a shared mean cumulant. When two random matrices are merged in a free and independent sense, the free cumulant of their joint spectrum is equal to the sum of their individual free cumulants. The differences in channel estimation between the two parties mainly stem from asymmetric hardware and local noise. These perturbation components tend to be free and independent from the common channel response in an asymptotic sense. Therefore, the averaging operation is equivalent to preserving the common signal subspace components in the cumulant domain while compressing the independent perturbation components by a factor. This process involves only scalar addition and division, resulting in extremely low computational complexity. Furthermore, it does not require exchanging the original channel matrix; only a small number of cumulant values ​​need to be transferred. This approach balances minimizing communication overhead with accurate extraction of common spectral features, providing an unbiased and low-noise numerical driver for subsequent continuous fractional reconstruction.

[0035] S304: Combining the mean of free cumulants of the same order and the retained truncation order, construct the continued fractional Stilts transform of the consistent average spectral distribution of the two communicating parties.

[0036] in, Describe the complex independent variable of the Stilchez transform. This represents the consistent average spectral distribution of the communicating parties. express The continued fractional Stilchez transform; It should be noted that this step utilizes the explicit bijective relation between the R-transform and continued fractions in free probability theory to transform the mean sequence of free cumulants of the same order into a rational function approximation of the Stieltjes transform of the average spectral distribution. Any spectral distribution characterized by free cumulants can be expressed as a finite-level continued fraction with cumulants as coefficients, the number of truncation levels being naturally determined by the order of truncation. The calculation is recursively performed from the innermost denominator outwards; the entire process involves only algebraic substitution, without iterative optimization or artificial smoothing. The continued fraction structure has a natural numerical damping effect on small perturbations of higher-order cumulants, resulting in a highly smooth reconstructed spectral density while preserving the main characteristics of the signal subspace. Therefore, both communicating parties can independently reproduce completely identical spectral distribution functions using only a small number of scalar cumulants, laying a rigid and unambiguous analytical foundation for subsequent eigenvalue extraction from quantiles, combining determinism, low communication overhead, and numerical robustness.

[0037] S305: Perform the inverse Stilchez transform on the continuously fractional Stilchez transform to obtain a uniform average spectral distribution; S306: Extract multiple eigenvalues ​​from a uniform average spectral distribution and construct a diagonal matrix of the eigenvalues; Specifically, based on the consistent average spectral distribution jointly obtained by both communicating parties, the method first constructs the corresponding cumulative distribution function by integrating the spectral density function, and determines the required number of eigenvalues ​​according to the minimum number of antennas at each end of the communicating parties. Then, a quantile uniform sampling strategy is employed to select multiple preset quantile probabilities on the cumulative distribution function, and the corresponding quantiles are calculated as eigenvalues, thereby obtaining a set of eigenvalues ​​covering the entire spectral distribution interval. Finally, the eigenvalues ​​are arranged in a predetermined order on the main diagonal to construct an eigenvalue diagonal matrix. Utilizing the quantile mapping mechanism of the cumulative distribution function, the continuous spectral distribution is transformed into representative discrete spectral features, ensuring that the extraction results fully reflect the overall spectral structure. This method allows both communicating parties to independently obtain a completely consistent set of eigenvalues ​​without additional negotiation, exhibiting strong determinism, high spectral information utilization, strong resistance to random disturbances, and adaptive matching with antenna dimensions.

[0038] In one possible implementation, the number of eigenvalues ​​is consistent with the minimum number of antennas at the peer end of the communicating parties.

[0039] It is understandable that taking the minimum number of paired antennas as the number of eigenvalues ​​can both fully preserve the distinguishable signal subspace dimension and avoid introducing redundant components without physical correspondence, so that the positive definite matrix can be accurately matched with the real channel degrees of freedom.

[0040] S307: Perform a hash operation on the sequence composed of the mean values ​​of each free accumulator of the same order to obtain a random seed, and generate an orthogonal matrix with the minimum number of antennas at the opposite end based on the random seed; Specifically, based on a pre-shared random seed between the communicating parties, a pseudo-random number generator is first initialized to generate a random matrix with independent elements that follows a standard normal distribution. Then, this random matrix is ​​subjected to QR orthogonal decomposition, and the orthogonal matrix is ​​extracted as the target orthogonal matrix. The principle is that the orthogonal matrix obtained after QR decomposition of a Gaussian random matrix naturally follows a Haar uniform distribution, enabling unbiased random sampling within the orthogonal matrix space. This not only ensures good randomness and statistical uniformity of the generated result, but also, because the random seed uniquely determines the random sequence, allows both communicating parties to independently reconstruct a completely identical orthogonal matrix without transmitting the matrix itself. This achieves repeatability, synchronization, security, and low communication overhead.

[0041] S308: By fusing the diagonal matrix and the orthogonal matrix, the positive definite matrix of the channel is obtained.

[0042] The formula for the positive definite channel matrix is ​​as follows: ; in, Represents the positive definite matrix of the channel. and Let them represent an orthogonal matrix and a diagonal matrix, respectively. express The conjugate transpose of .

[0043] Specifically, this reconstruction module enables both communicating parties to locally resolve and reconstruct a consistent positive definite channel matrix by exchanging only low-dimensional free cumulants rather than the complete channel matrix, avoiding the bandwidth consumption and information leakage risks associated with high-dimensional matrix transmission. The free cumulants possess strict additivity under non-cross-partitioning, ensuring that the mean of the cumulants from both parties accurately characterizes the common spectral structure of the shared channel signal subspace, while automatically filtering out differences caused by asymmetric hardware and high-order cross-contamination from receiver thermal noise. Subsequently, using the explicit bijective relationship between the R-transform and the continued fractional Stilchez transform in the free probability, the average spectral density function is recovered in a closed loop from the mean cumulants. The entire process relies solely on the cumulant sequence itself without any manually adjusted parameters, guaranteeing the objective uniqueness of the spectral reconstruction. A diagonal matrix of eigenvalues ​​is obtained by quantile extraction of the spectral density, and a positive definite matrix is ​​constructed using a hash seed-driven Haar distribution orthogonal matrix. This ensures that the matrices generated by both parties are numerically identical and that the orthogonal basis is statistically unbiased and uniform, eliminating any dependence on specific directions. The resulting positive definite channel matrix is ​​strictly symmetric with all positive eigenvalues, exhibiting invertibility and numerical stability. It fully integrates the shared channel spectrum features of both parties. Furthermore, through statistically discriminative truncation of free cumulants and analytical spectrum reconstruction, it can robustly extract effective signal dimensions even in strong noise environments, providing an unambiguous and highly consistent common reference basis for subsequent key bit extraction based on generalized eigenvalue decomposition. The entire process, from cumulant mean to continued fraction transformation, then to inverse spectral transformation and positive definite matrix synthesis, constitutes a fully analytically driven deterministic pipeline, completely eliminating subjective thresholding and iterative convergence issues, while balancing the physical layer security of key negotiation with the reproducibility of system implementation.

[0044] Generation module 4 is used to generate session keys based on the positive definite channel matrix; The session key is an encryption key temporarily generated and shared by both parties during a communication session, used to encrypt and decrypt data within that session. By generating the session key using shared channel characteristic information contained in the positive definite channel matrix, both parties can obtain a consistent encryption basis without pre-distributing fixed keys, thereby improving the randomness and unpredictability of the key.

[0045] In one possible implementation, generation module 4 is specifically used for: S401: Extract the principal eigenvectors of the positive definite channel matrix; The principal eigenvector is the eigenvector corresponding to the largest eigenvalue after eigenvalue decomposition of the positive definite channel matrix.

[0046] S402: Quantize the main feature vector elements into a binary bit sequence to obtain the initial session key bit stream; Optionally, the continuous elements of the main feature vector can first be mapped to discrete integer values ​​according to a predetermined quantization precision, for example, through interval partitioning or threshold comparison to achieve binarization; then, each discrete value is converted into a fixed-length binary bit string, and concatenated according to the vector order to obtain a complete bit sequence. By utilizing the channel's main feature vector to carry randomness and uniqueness information, and through quantization and binary mapping, the continuous feature space is effectively transformed into a discrete bit space, ensuring that the bit sequences generated independently by both parties are completely consistent. The entropy characteristics of the channel are fully utilized to ensure key security and randomness.

[0047] S403: Input the initial session key bitstream into a hash function driven by free cumulant and spectral features, and output the session key.

[0048] Specifically, this module uses the principal eigenvector of the positive definite channel matrix as a random source. It generates an initial bitstream by quantizing and binary encoding its continuous elements, then uses a cryptographic hash function for privacy amplification, ultimately outputting a fixed-length session key. The principal eigenvector captures the maximum energy direction of the signal subspace, which is uniquely determined by the shared physical channel spectrum structure. This direction is naturally adaptable to receiver noise and hardware asymmetry, ensuring high consistency of quantization results at legitimate communication endpoints. The random fading characteristics of the channel itself inject unpredictable physical entropy into the key, while the hash function further eliminates statistical dependencies from quantization residues, ensuring the output key meets the uniform distribution and independence required for information-theoretic security. The entire process avoids the risk of pre-distributed keys; the key is only extracted from the channel in real-time when needed and becomes invalid after the session ends, significantly enhancing the adaptability and anti-eavesdropping capabilities of the communication security system.

[0049] Specifically, the hash function is a hash function driven by free cumulant and spectral features, thereby realizing the generation of the session key. Specifically: ; ; in, Indicates the bit stream of the initial session key. Session key, Indicates the first The binary value of the initial session key bitstream. , Indicates the total length of the initial session key bitstream. This indicates that the truncation order should be retained. This indicates that the truncation order is retained. r The mean of the free cumulants of the same order, This represents the largest eigenvalue extracted from a uniformly average spectral distribution. This indicates operations involving taking the decimal part. Indicates rounding down. Describes the first sample drawn from a uniform average spectral distribution. j 1 eigenvalue, , This represents the total number of eigenvalues ​​drawn from a uniform average spectral distribution, where... That is, the minimum number of antennas at the opposite end. This represents the sum of the extracted feature values. This represents the XOR operation.

[0050] Specifically, in the generation of the session key, the first channel multiplies the mean of each free cumulant by the largest eigenvalue, takes the root of that order, and then multiplies it by a convergence factor constructed from all initial bits through product embedding. This is then mapped to a fixed-length integer by rounding down and removing the fractional part. The second channel multiplies each normalized eigenvalue by a combination of the harmonic radicals of all cumulant means, also mapped to an integer. The results from both channels are then fused using a bitwise XOR operation. This scheme is based on the additivity of cumulants to independent spectral components in free probability theory, the unique representation of the channel state by the spectral distribution of a random matrix, and the exponential diffusion of single-bit flips by product mapping in ergodic theory. This ensures that the hash output not only possesses the uniformity and avalanche effect of standard cryptographic hashes but is also bound to the real-time statistical structure of the current channel. Since both communicating parties share the same cumulant mean and spectral distribution, while the eavesdropper's location results in a different channel matrix and an inability to obtain consistent statistics, the session key output by this hash function is naturally consistent between the communicating parties. However, it becomes ineffective against attackers because the mapping relationship itself is irreproducible. Meanwhile, the combined participation of multi-order cumulants and full-spectrum eigenvalues ​​enables the key entropy source to cover all identifiable dimensions from the main signal subspace to the noise boundary, greatly enhancing the ability to resist brute-force attacks and known-plaintext attacks. Moreover, no additional key material needs to be transmitted throughout the process, ensuring security while ensuring lightweight communication.

[0051] Transmission module 5 is used to encrypt and transmit the data to be transmitted using a session key.

[0052] In one possible implementation, the transmission module 5 is specifically used for: S501: Initialize the encryption algorithm using the sender's session key; Specifically, the process of initializing the encryption algorithm using the sender's session key involves first inputting the generated session key into the key register or key scheduling module of the symmetric encryption algorithm to complete the internal state initialization of the encryptor. This process uses the session key as the unique input of the algorithm to determine the encryption stream or round key sequence, ensuring that each encryption output is strictly associated with the session key. This guarantees that the sender and receiver can generate a consistent encryption state using the same session key, achieving reproducibility and security of data encryption, while eliminating the need to transmit additional key information and reducing communication overhead.

[0053] In one possible implementation, the encryption algorithm is AES-128, AES-192, or AES-256.

[0054] S502: Encrypt the data to be transmitted using the initial encryption algorithm and output the ciphertext data; It should be noted that during the encryption process, both communicating parties use a negotiated session key and a consistent average spectral distribution. Maximum eigenvalue A complex parameter is determined through a key derivation function. ,in, and The hash value of the session key is obtained by segmenting and mapping. First, the session key is concatenated with a fixed auxiliary string, and a 256-bit hash value is calculated using SHA-256. Then, this hash value is divided into two equal parts, H1 and H2, each 128 bits long, and parsed as big-endian unsigned integers, resulting in u1 and u2 respectively. Finally, u1 and u2 are normalized to the [0,1] interval and linearly mapped to a preset range of positive real numbers, thus obtaining the hash value. Similarly... (Derived from u2 using the same principle). The selection rules ensure complex parameters. To ensure that the spectrum is far from a uniform average distribution. Invertible and with good condition number, where, express N × N The identity matrix. Then, the data to be transmitted is vectorized, dividing the data stream into segments... N Each symbol is grouped into N symbols (N being the minimum number of antennas at the other end), and each group is mapped to a complex vector x. If the last group is insufficient, it is padded with zeros. Then, using... Encrypting the complex vector x yields the ciphertext vector. The sender transmits the ciphertext vector y to the receiver, who then uses its independent... and exactly the same The plaintext can be recovered without loss by solving the system of linear equations. .

[0055] It should be noted that this encryption process deeply integrates physical layer channel characteristics with cryptographic hash dynamics. Based on a pre-negotiated session key and the maximum eigenvalue of the consistent channel spectrum, the communicating parties generate complex parameters far removed from the spectral distribution through hash segmentation mapping, ensuring the invertibility and numerical robustness of the pre-defined matrix. After the plaintext is vectorized, it undergoes encryption transformation using the inverse operator of the pre-defined matrix constructed from the positive definite channel matrix to generate ciphertext. The pre-defined matrix carries complete channel spectrum structure information, and this positive definite matrix is ​​reconstructed by both parties from the free cumulative amount of the shared channel, exhibiting strict reciprocity and consistency. Since eavesdroppers are located in different spaces, their channel characteristics are independent of the communicating parties, making it impossible for them to reproduce the same matrix. Even if they intercept the complex parameters and ciphertext, they cannot solve for the plaintext, thus elevating security from computational complexity to the uniqueness of the physical layer channel. Simultaneously, the complex parameters are dynamically determined by the hash segmentation mapping of the session key, ensuring that each encryption operator is independent, naturally possessing forward security. Furthermore, the matrix dimension is determined only by the number of antennas, resulting in minimal computational overhead, achieving efficient transmission with no key distribution and strong physical layer security.

[0056] S503: Transmit the encrypted data to the receiver in the communication process, completing the encrypted transmission of the data to be transmitted.

[0057] Understandably, this transmission module uses the session key as the unique initialization seed for the symmetric encryption algorithm, driving the encryptor's internal state to generate a keystream strictly bound to the key. This keystream is then combined symbol-by-symbol with the data to be transmitted to form ciphertext, which is sent to the receiver via a public channel. Since both parties independently extract identical session keys from the shared physical channel, the sender's encryption state and the receiver's decryption state are naturally synchronized, eliminating the need for additional key transmission or initialization vectors. The entire process seamlessly integrates key generation and data encryption. The key is only valid within the current session and dynamically updated as the channel changes. Even if the ciphertext is completely intercepted, attackers cannot reconstruct the key due to a lack of simultaneous channel physical information. Thus, without sacrificing effective throughput, a balance between data confidentiality and transmission efficiency is achieved.

[0058] In one possible implementation, it also includes: Update module 6 is used to update the session key at preset intervals.

[0059] It should be noted that those skilled in the art can set the preset duration according to actual needs, and this invention does not limit this.

[0060] Understandably, periodically forcibly refreshing the session key can effectively limit the exposure time window of a single key. Even if the current key is leaked, attackers will not be able to decrypt subsequent communication data, thereby continuously enhancing the forward security of the system.

[0061] Furthermore, upon receiving the ciphertext, the receiver uses its independently generated session key, identical to the sender's, derived from the shared channel positive definite matrix, to initialize the decryption module according to the same symmetric algorithm specification as the encryption end, generating a keystream synchronized with the encryption end. The ciphertext data is then input into the decryptor, and the keystream mask is stripped away through symbol-by-symbol inverse operations to recover the original plaintext data bits transmitted by the sender. Because the session key originates from the reciprocal characteristics of the physical channels between the two parties and is not transmitted over the air interface, even if an eavesdropper intercepts the entire ciphertext, they cannot decrypt it without identical channel information. This ensures data confidentiality while achieving end-to-end secure transmission without the need for additional key distribution.

[0062] In practical applications, this scheme achieves a closed-loop data transmission security by directly converting the physical randomness of the wireless channel into a dynamic session key. The entire process begins with both communicating parties alternately sending orthogonal pilots and independently estimating the instantaneous channel matrix during the channel coherence time. The channel matrix is ​​then transformed into a sample covariance matrix. Through non-cross-partitioning and Möbius inversion, free cumulants of various orders are extracted from the sample moments. The essence of this operation is to utilize the additivity of cumulants in free probability theory to remove cross-coupling terms caused by noise and hardware asymmetry, tightly capturing high-order statistical features belonging only to the shared channel signal subspace. Using the upper bound of variance given by random matrix theory, the cumulants are tested for significance at each order, automatically truncating to the highest statistically identifiable order, completely eliminating manual parameter tuning. After exchanging their respective low-dimensional cumulants, the average of these cumulants is taken, and the fusion of common spectral features and perturbation suppression are completed in the cumulant domain. Then, utilizing the explicit bijective relationship between the R-transform and the continued fractional Stilchez transform in free probability, a consistent average spectral density function is uniquely reconstructed from the mean cumulant. This is then combined using quantile extraction and hash seed-driven orthogonal matrix synthesis to obtain a channel positive definite matrix identical to both parties. This matrix serves as the digital fingerprint of the shared channel, and its eigenvectors carry irreproducible physical entropy. After polar quantization and privacy amplification, a session key is generated. The key is only valid within the current session. The sender uses its initial symmetric encryption algorithm to stream-encrypt the data, and the receiver uses the independently generated identical key to decrypt synchronously, without transmitting any key material. Even if the ciphertext is completely intercepted, the attacker cannot derive the key due to the lack of physical information about the same channel at the same time. Thus, a forward-secure one-key-one-transmission security method based on objective physical layer characteristics is achieved without being subject to any artificial thresholds or sacrificing effective throughput.

[0063] Compared with the prior art, the present invention has at least the following beneficial technical effects: In this embodiment of the invention, the communicating parties first acquire their respective observed instantaneous channel matrices, which characterize the propagation characteristics and random changes of the current communication link at a specific moment. Subsequently, multi-order free cumulants are extracted from the instantaneous channel matrices. These multi-order free cumulants can extract stable and discriminative statistical structural features from complex random matrices, thus accurately characterizing the random state of the channel. After exchanging their respective statistical features, the communicating parties obtain a consistent positive definite channel matrix through a reconstruction process. The positive definite channel matrix is ​​a unified mathematical expression of the shared channel features of both parties, preserving the intrinsic channel information while possessing good stability and uniqueness. Based on this, a session key is generated. Since the session key originates from the shared and real-time changing channel features of both parties, it possesses strong randomness, unpredictability, and dynamic update capability. Finally, the session key is used to encrypt the data to be transmitted, making it difficult to recover even if the data content is intercepted. This not only avoids the security risks associated with traditional static parameters but also fully utilizes the natural randomness inherent in the communication environment to enhance key strength and improve the consistency of keys between the communicating parties, thereby effectively improving the confidentiality, integrity, and reliability of data transmission.

[0064] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0065] The above embodiments merely illustrate several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A digital communication data secure transmission system, characterized in that, include: The acquisition module is used to acquire the data to be transmitted and the instantaneous channel matrix of both communicating parties; The determination module is used to determine the multi-order free accumulator of the instantaneous channel matrix with the goal of ensuring that the random channel state of the two communicating parties is identifiable; The reconstruction module is used to exchange the multi-order free cumulative quantities of the two communicating parties and reconstruct a consistent positive definite channel matrix for the two communicating parties. The generation module is used to generate a session key based on the positive definite channel matrix; The transmission module is used to encrypt and transmit the data to be transmitted using the session key.

2. The digital communication data secure transmission system according to claim 1, characterized in that, The two parties in the communication include the sender and the receiver.

3. The digital communication data secure transmission system according to claim 1, characterized in that, The acquisition module is specifically used for: S101: Instruct the two communicating parties to alternately send orthogonal pilot sequences; S102: Collect the received pilot signals from both communicating parties; S103: Perform least squares estimation on the received pilot signal of either of the two communicating parties to obtain the instantaneous channel matrix.

4. The digital communication data secure transmission system according to claim 1, characterized in that, The determining module is specifically used for: S201: Based on the instantaneous channel matrix, establish the sample covariance matrix of either of the communicating parties; S202: Calculate the sample moments of each order of the sample covariance matrix; S203: Based on the sample moments of each order, and in combination with the non-cross-partitioning method, determine the candidate multi-order free cumulants of the corresponding sample covariance matrix; S204: Verify order by order, and output the maximum order that satisfies the inequality of random matrix theory as the cutoff order; S205: Output the candidate multi-order free accumulator that is less than or equal to the truncation order as the multi-order free accumulator.

5. The digital communication data secure transmission system according to claim 4, characterized in that, The reconstruction module is specifically used for: S301: Obtain the truncation order of the two communicating parties, and output the minimum truncation order as the retained truncation order; S302: Obtain the multi-order free cumulative quantities of the two communicating parties corresponding to the reserved truncation order, and exchange them; S303: Calculate the average free cumulants of the same order in the retained cutoff order; S304: Combining the mean of the free cumulants of the same order and the retained truncation order, construct the continuous fractional Stilts transform of the consistent average spectral distribution of the two communicating parties; S305: Perform an inverse Stilchez transform on the continuously fractional Stilchez transform to obtain the uniform average spectral distribution; S306: Extract multiple eigenvalues ​​from the uniform average spectral distribution and construct a diagonal matrix of the eigenvalues; S307: Perform a hash operation on the sequence composed of the mean values ​​of each of the free accumulators of the same order to obtain a random seed, and generate an orthogonal matrix with the minimum number of paired antennas based on the random seed; S308: The diagonal matrix and the orthogonal matrix are fused to obtain the positive definite channel matrix.

6. The digital communication data secure transmission system according to claim 5, characterized in that, The number of eigenvalues ​​is consistent with the minimum number of antennas at the peer end in the communication.

7. The digital communication data secure transmission system according to claim 1, characterized in that, The generation module is specifically used for: S401: Extract the principal eigenvector of the positive definite channel matrix; S402: Quantize the main feature vector elements into a binary bit sequence to obtain the initial session key bit stream; S403: Input the initial session key bitstream into a hash function driven by free cumulant and spectral features, and output the session key.

8. The digital communication data secure transmission system according to claim 1, characterized in that, The transmission module is specifically used for: S501: Initialize the encryption algorithm using the sender's session key; S502: Encrypt the data to be transmitted using the initialized encryption algorithm and output the ciphertext data; S503: The encrypted data is transmitted to the receiver in the two communicating parties, thus completing the encrypted transmission of the data to be transmitted.

9. The digital communication data secure transmission system according to claim 8, characterized in that, The encryption algorithm is AES-128, AES-192, or AES-256.

10. The digital communication data secure transmission system according to claim 1, characterized in that, Also includes: The update module is used to update the session key at preset intervals.

Images